The definitive guide to 24-bit FLAC

The definitive guide to 24-bit FLAC

High resolution downloads are changing the way we enjoy music, opening up possibilities undreamed of back in digital’s early days…

CD was revolutionary when it first appeared, just over 30 years ago: none of the surface noise of well-played LPs, no annoying ticks and pops caused by invisible scratches, no need to handle the discs with kid gloves – though the early demonstrations of discs smeared with jam still playing were a little over the top! – and instant access to any track on an album at the push of a button. CD running orders could be programmed, favourite tracks repeated and the duff ones skipped, and in time added features such as CD Text readouts of album, track title and artist would become possible.

For all that, CD was always a compromise, using technology that was only just cutting edge at the time. 30 years on, we can do so much better.

Digital audio encodes the analogue musical waveform, using Pulse Code Modulation. With PCM, the two basic factors that determine the sound quality are bit depth (which determines the quality with which the analogue signal is digitised), and the sampling frequency (which is the number of times per second that the analogue signal is digitised). CD is a 16-bit, 44.1kHz system, what’s needed to improve conformity between the analogue original and the digitised version is a combination of more bits in each samples, and more samples for each second of music: the more bits you use, the more natural the music sounds, and the faster the samples, the wider the frequency range.

CD’s 16-bit/44.1kHz digital system may have been the apparent state of the art in 1982, but it didn’t take long for it to get past its sell-by date. Indeed, some say it was only ever almost good enough for high-quality sound reproduction: Malcolm Hawksford, Professor of Psychoacoustics at Essex University, feels that, ‘It was near the limit, but in my view probably a bit marginal. Ideally, at least 20-bit resolution at 60 kHz sampling frequency would have been better.’

 

FROZEN CDS

However, 16-bit/44.1kHz is what we got back in 1982, and 16-bit/44.1kHz is what CD still uses more than three decades on: we’ve been trapped in a technology ‘almost good enough’ 30 years ago for all that time, just like we would be with everything else we own if all development had been stopped back then.

Yes, there have been attempts to replace the CD with something better – in the early 1990s not one but two ‘high-resolution’ audio formats appeared, and battle was joined between DVD-Audio and the Sony/Philips Super Audio CD (or SACD), the former offering 24-bit audio quality for greater resolution, improved dynamics and greatly reduced distortion, while the latter’s Direct Stream Digital system used a sampling rate of 2.8224Mhz, 64x that of CD and single-bit sampling.

The new formats also offered the opportunity for multichannel surround sound and greater security and copy protection – one good thing for consumers, it was suggested, and one for the record industry. However, there was a problem or two…

DVD-A discs needed a special player – not just a standard DVD-Video machine – and these were a while coming, while the music retail industry wasn’t keen on stocking releases in multiple formats, and dealing with a hassle of buyers picking up a DVD-A disc by mistake and finding it wouldn’t play on their CD player.

SACD had an answer to that one: its multilayer discs allowed back-compatibility with CD hardware, which just saw them as standard CDs. Only when used in an SACD player was the higher-resolution layer ‘unlocked’. That meant record stores could sell just one disc for SACD and CD buyers, but then the pitching of SACD as a premium medium would have meant CD prices rising, at a time when deep discounting on music was really beginning to kick in.

 

What has, unfortunately, made a big impression on the market is compressed audio formats, notably MP3: originally used to allow sensible amounts of music to be stored on early personal music players with very limited storage capacity, MP3 can reduce the size of a music file by around 80% by removing most of the sound you can’t hear – quieter sounds masked by louder ones, for example. Unfortunately along the way it also loses quite a lot of the sound quality you can hear, and while the losses may not be so apparent on lower-quality equipment or when played in noisy environments – on headphones in the street or on public transport, for example, or on standard-fit audio systems in many cars –, when you step up to better audio systems it’s all too easy to hear the reduction in quality.

The good news is that while MP3 does’t sound too great, it did open up new channels through which music could be obtained – via online downloads.With those channels opened up, Free Lossless Audio Codec – or FLAC – now makes it possible to download music at CD quality or better, or store your CDs on computer or external storage, while saving up to 50% of the file size without losing any quality.

The clue’s in the term Lossless: FLAC packs music without losing any of the content or quality, and when ‘unzipped’ correctly the file is exactly the same as the original file. And while we’ve seen a number of rival lossless systems, from Apple Lossless (ALAC) to Windows Media Audio Lossless (WMA Lossless), the Free Lossless Audio Codec (FLAC) is the system of choice for many music-loving hi-fi buffs, simply because it’s not tied to any one manufacturer’s systems.

 And it can do more than cut the size of CD-quality files: FLAC can also come in 24-bit form and beyond, as Albert Yong of Bowers & Wilkins explains: ‘The system is so flexible that it can take anything from 4 to 32 bits and sample rates up to 655350Hz in 1 Hz steps.’ he says.

And FLAC, first introduced in 2001, is opening up a whole new world of high-end, high-resolution audio for enthusiasts: 24-bit digital has a theoretical resolution of 144dB (compared to around 96dB in CD’s 16-bit) which is more than enough, given that 130dB is the threshold of pain for the human ear. That means you can get ever closer to the master-tape used in the studio, as well as gaining all the information made possible by the greater data-rates of these high-resolution files: Albert Yong says ‘The difference is in the detail. Sounds generally are more open, and there is an extra level of airiness in the music. Voice and instruments sound closer to live, and more dynamic as well.’

While you can use computer software to rip your CDs into FLAC, there’s also a growing number of sites offering downloads at CD quality and beyond including Society of Sound. We’ve structured our site to make downloading FLAC a breeze. Once you’re signed up as a member you get free albums every month to download; it’s just a matter of downloading a zip file, unzipping it into whichever folder you want on your computer or storage device, and then playing your music using your choice of player software.

BUILDING YOUR OWN LIBRARY

If you’re looking to build up a FLAC library of your existing CD collection, then a good starting point for ripping is Exact Audio Copy or dBpowerAMP. These are freely downloadable applications for Windows, commonly thought to be the best sounding and most bug-free. Mac Users can try Max, MacFlac or X Lossless Decoder.

To play the music, there’s a range of software players available, though as you’re using FLAC you’ll find neither iTunes nor Windows Media Player will handle the files directly, so you may have to use a modification to those programs or convert the files to play them. Fluke, for example, is a freeware program that works as ‘plug in’ for iTunes (Mac platform), but it can be buggy.

There are various programs available to convert files to PCM or even Apple Lossless, such as Max for MacOSX, but the best option is probably to use a different player solution, which will play all files in their native format.

On Windows systems, Winamp can play a wide variety of lie formats, and even burn 16-bit files to CD while other good player packages include Foobar2000, Media Monkey, Vox and Nightingale, the last of these being a ‘forked’ version of the now defunct Songbird package, having an interface very similar to iTunes and capable of playing almost any file you throw at it.

Nightingale is also available for Mac OSX, and like other players will hunt down all the music on your computer and add it to its own library if you require. These programs also allow editing of the ‘metadata’ of your music files – the information used to identify the music – should you find a rip or download doesn’t quite have the information you want about artist, album, track and so on.

 Amarra Mini by Sonic Studio is a more sophisticated system that piggybacks on to iTunes, replacing much of its audio code with a superior sounding and more flexible system. Sound is excellent but you pay for it; the basic Junior version which gives FLAC playback up to 24/96 starts at $99.

 Audirvana Plus (www.audirvana.com) is another fine application that, like Amarra, shuts down certain parts of your computer to minimise drain on CPU resources, to (so the theory goes) give better sound. There’s a free trial download but you’ll have to pay for the full version. It now offers optional integration with iTunes and is a strong, stable performer. For Mac users, its combination of value and performance is currently hard to beat.

 All these programs work as the playback system, but unless you’re happy to listen via your computer’s speakers, or a pair of headphones plugged into it, you’ll need a means of connecting computer to your audio system in order to hear your new high-resolution music collection.

       The simplest way to do this is to take an analogue line output from your computer soundcard. This is likely to have a 3.5mm stereo mini-jack plug, so you’ll need a cable with a 3.5mm stereo plug on one end and a pair of RCA phonos on the other, to connect into a spare line input on your amplifier to system (although systems and amplifiers increasingly have a 3.5mm input on the from for just this purpose. This method will give a reasonable sound, particularly if you have a good quality soundcard, but it’s not ideal computers are electrically ‘noisy’ environments, and it’s best if at all possible to pipe out the digital signal away from the PC or Mac.

       A range of USB DACs are now available to plug straight into your computer and provide an analogue output. Some provide a line output to connect to a hi-fi system, while others allow headphones to be connected for a truly portable high-resolution system, and there’s now a wide range of digital to analogue converters and even amplifiers with USB inputs. Prices start at about £100, while it’s perfectly possible to spend many times that amount on a DAC or headphone amplifier designed for high-resolution music.

Most of these DACs are powered from the computer, and use an asynchronous USB connection, so the high quality clock onboard will control the computer’s digital datastream. This gives tighter, smoother sound compared to standard asynchronous USB, as a result of lower jitter.

       There’s also a small selection of personal players able to play high-resolution FLAC files (among other formats), should you want to take your music library with you on your travels. The most affordable of these starts from under £200.

       Wireless and network streaming open up much greater flexibility. and also allows the streaming of music without the need to have your computer on all the time. You can copy your entire library to a Network-Attached Storage (or NAS) device, which is basically a hard-disk store designed to connect to your wired or wireless network storage, and which can be accessed by multiple players (either standalone systems or network music players designed to connect to a hi-fi system) on your home network.

Several players can be used at once, each playing different music if required: if you want a whole-house music solution, networking is definitely the way to go. Prices for NAS devices start from under £200 for a simple unit, and players at under £400. And the choice out there keeps on growing, as FLAC music catches on.

However, whichever solution you choose, the sound you achieve will only ever be as good as the quality of the files you’re playing, and that’s why we at Bowers & Wilkins, with almost 50 years’ experience of making top-quality premium loudspeakers, take FLAC very seriously. Not only is it convenient, with its shorter download times and almost bombproof content tagging, it’s also upgradable as high-resolution formats evolve: one label – Norway’s 2L – is already offering 24-bit/352.8kHz FLAC files taken from its original DXD masters.

Bowers & Wilkins’s Albert Yong says that ‘FLAC is as good as it can get for now. I think what it brings is not limited to quality of audio, but the added convenience as well’. And whereas once the choice of high-quality music in high-resolution FLAC was limited, initiatives such as the Bowers & Wilkins Society of Sound, are devoted to making more music available, and more people aware of what high-resolution music can bring to their listening experience.

High-quality, high-resolution music, recorded and delivered with fanatical attention to detail and quality – that’s what Society of Sound is all about, and why we’re so keen for you to share our enjoyment of the music we offer.

118 Comments

  • Chris Connaker says:

    Nice article Susanna. It’s wonderful to see B&W talking about 24 bit FLAC.

  • Markus says:

    In the article you were saying:
    “it’s always preferable to isolate the electrically noisy computer from the hi-fi by using an optical digital link.”

    Does this mean you recommend connecting for example a Sonos ZP90 Networkplayer to a Naim SuperNait rather via the optical link rather the digital cable?

  • Mike says:

    Can someone please tell me if the Zeppelin can accept 24-bit using the optical inlet?

    I’m using my computer with the M-Audio transit and can send flac files to the Zeppelin. I’d like to buy some 24-bit recordings and want to know if Zeppelin’s internal DAC can process them.

    I can’t seem to get B&W customer support to respond.

  • Shaun: B&W says:

    Hello Mike

    Thanks for your interest in 24-bit. As you are probably aware our Society of Sound offers a wide selection of albums in 24-bit.

    As for your Zeppelin question, unfortunately the Zeppelin doesn’t support 24-bit.

    Regards

    Shaun

  • Sam (UnRated NYC) says:

    Hi there. I just wanted to tell you that this article is truly incredible. I’m the editor-in-chief of unrated nyc and I plan on sharing this piece with all my readers. Great work!

    _Sam

  • Andrew Levine says:

    I was irritated when I failed to find any information on FLAC’s developer, then discovered his name is “Josh Coalson” (not Coulson). Other than that I found your article quite interesting. I am not sure I agree with the universal need for >16 bit audio though.

    As a sonophile audio professional I always record at and work with 24 bit audio, frequently at higher sample rates than 44.1, but when I carefully dither my final mixes down to 16 bit (using iZotope’s MBIT+ algorithm or one of the fine Airwindows dithers) I have always been very happy with the results.

    Certainly it’s preferable to leave the material in the original resolution, especially when dealing with the extreme dynamic range and subtleties of classical and avant garde music performed in “real” venues, but for most consumers of the music I produce the dithered version will be way superior than many other off-the-shelf recordings.

    It might boil down to the care you take in training your ears, selecting your gear, preparing for and then recording the acoustic event in the first place, as well as the purity of the processing chain. And let’s not forget that it all begins with good music, performed by inspired musicians. Many thanks to them!

    Regards,

    blumlein records – Andrew Levine

  • Don Baragar says:

    Great article but it does not talk about the effect of amps and speakers on the result. If the frequency response of speaker id 25 hz to 20khz will you notice a difference in 16/44 and 24/96 sound?

  • Andrew Levine says:

    Do we need 24 bit audio for sonic nirvana?

    http://radio.blumlein.net/2009/12/do-we-need-24-bit-audio-for-sonic.html

  • SusannaGrant says:

    Hi Andrew,
    Many thanks for your comments – I have amended Josh’s name and apologise for that.
    I thought your last point is particularly salient and is one of the cornerstones of Society of Sound.
    Kind regards,
    Susanna

  • Vereecken says:

    I would highly welcome 96 kHz / 24 bit downloads on the B&W music site.

  • digifish says:

    This is worth a read. I have been producing music for 15 years and concurrently worked in human factors research in the medical device industry, so I am data driven…

    http://mixonline.com/recording/mixing/audio_emperors_new_sampling/

    16 bit 44.1 kHz is indistinguishable from higher sample rates and bit depths. That you prefer SACD or some other high-end format is likely that it was produced better in the first place.

    digifish

  • Paul Dzwonowski says:

    Years ago I was doing the usual browsing at a music shop and stumbled on at the time – for me – an unusual cardboard mini lp CD import from Japan – Charlie Rouse – remastered in 24 bit. At the time I had no idea what 24 bit was – I was just entering the digital world from analog 12” LPS. The CD, however, had a mystique about it compared to all the other “ordinary” plastic jewel cased CDs in the store, and I decided to purchase it and give it a listen. Ever since then I have been an absolute believer in 24 bit recordings and have been collecting the majority of music in 24 bit; never did cross over to SACD format. A few years back I started noticing alot of remastered music coming from Japan in 20 bit format after 24 bit had started. The 20 bit recordings sounded awesome but it puzzled me. The whole process and the art of remastering recordings has since really puzzled me; however aside from the inherit individual preferences in remastering music, why would one remaster a recording in 20 bit instead of 24 bit? Aside from 20 bit vs 24 bit I’d sure like to see a book published to both celebrate and discuss the whole art of remastering recordings!

  • John Clayton says:

    How do I get from MPEG-4 Audio File Format (.m4a)to Flac so I can play the B&W tracks on my winamp

  • digifish says:

    I’m back, Ethan Winer of RealTraps (makers of sound control solutions for studios) just uploaded a fantastic video relevant to this topic from the 2009 Audio Engineering Society.

    Audio Myths Workshop.

    Ethan & co-presenters cover such issues as placebo effects in audio, loudness vs quality, ‘scam’ equipment, dithering, expensive vs cheap A/D D/A converters and more.

    Start at 41m:15s to cover D/A A/D converters and then bit-depth sections. If you have time
    I would encourage you all to watch the whole thing.

    http://www.youtube.com/watch?v=BYTlN6wjcvQ

  • Brad Jensen says:

    This is exactly what I have been waiting for!!

    FLAC is the best method for storing and listening to music, in my opinion. It’s a free codec, it compresses very well, it decodes well, and it’s lossless. It even holds multi-channel sound.

    If a better version of FLAC ever comes out, it’s easy to just convert those old lossless files to the newer version, without losing any quality!

    I have all of my CDs ripped to FLAC files and mirrored on another hard drive as a backup. :)

    Now, if only we can get mastering engineers / record labels to quit the loudness war and to start master old albums to the best quality they can be, and have those albums released like these 24 bit FLAC files..

  • Paul Stanwix says:

    Dear Audiophiles.

    Music is for listening to, wherever you are. You don’t need to worry about wether you are not hearing all the information. Just listen and enjoy. Even the cheapest CD player from your nearest supermarket does the job, just fine. I started recording my favourite music on a Phillips portable cassette recorder and progressed to a dansette record player when I was a kid. I wasn’t worried about what I wasn’t hearing, I just enjoyed , what was for me, a life changing experience.

    Sorry about that rant! But can any one help me with this problem? I have downloaded the 24 bit version of Peter gabriel’s CD. I want to place it in my i tunes libary. I have downloaded Max as I have a mac, running 10.5.8. What do I have to change the flac file into, to then drag and drop it into my i tunes libary.

    Thanks.

    Paul.

  • mackie says:

    Good article. I enjoyed reading it:)

    One note: MP3 is not open source format as Flac is. It’s proprietary and you have to pay for it if you want to use it. As a user you probably don’t pay for it directly ‘couse the company which software/hardware you use payed for it and charged you when you bought their stuff.

    This is also why most Linux distributions did not play MP3 files out of the box. Right now it may changed. Dunno ‘couse don’t use it any more.

  • thewelltemperedcomputer.com says:

    whereas FLAC is of course a free, open source system like MP3
    As far as I know, MP3 is not open source, you have to pay a license to Thomson

  • Mick says:

    I wish someone would take the trouble to give this article, with its misplaced links and trailing sentences, a decent edit so it’s actually readable. It would then reflect much better on its host B&W.

    Thanks for the SoS though. I’m enjoying my trial membership.

  • Joe Lubow says:

    Paul Stanwix says:
    “I have downloaded Max as I have a mac, running 10.5.8. What do I have to change the flac file into, to then drag and drop it into my i tunes libary.”

    Paul, Max can convert FLAC 24-bit files directly to Apple Lossless and send them straight to iTunes. In Max Preferences go to the “Formats” pane. Choose “Apple MPEG-4 Audio.” In the “Output” pane choose “Same as Source file” or choose some other place. In the iTunes pane check “use iTunes compatability mode” and “Add output fies to iTunes library.”

    Max will preserve the bit rate and sampling rate, and the Apple Lossless files will be identical in quality to the downloaded flac files. Since iTunes copies the files into your iTunes library, once the tracks show up in iTunes you can delete the Apple lossless files from the output folder, (I have Max set to put them in the same folder as the FLAC files). You could also delete the original flac files, since you could always convert the Apple Lossless files back to flac if you wanted to. I’m superstitous though, and even though they are not only theoretically and practically (to my ears through a Linn system) the same, I back up the flac files to an external drive before deleting them. Since Time Machine backs up iTunes, this is unnecessary, but it just feels weird to me to delete the original files.

    The only catch is that older iPods won’t play 24-bit files. I have an iPhone 3GS, and it plays 24-bit files beautifully, especially hooked up to a good rig, but it won’t play files with a sampling rate over 48k. Scratch My Back, like most of the Society of Sound recordings, is 48k, so you should have no problem; but for recordings with higher sampling rates, you’ll need to downsample. Max can’t do that but XLD is freeware that can. So if you ever find yourself with files sampled faster than 48k and want a version that will play on your iPhone, use XLD to downsample. But save the original higher sample’d file to play through iTunes.

    Depending on why you want them in iTunes, you could also forego iTunes and play the FLAC files directly in Cog or Songbird. But if you want to transfer them into an iPod or iPhone, you’ll need to use iTunes.

    I hope that helped…

  • David says:

    I have downloaded many of the SoS albums and mostly they sound great. However the first track on the Daby Touré & Skip McDonald album (Diaguanu) sounds awful to me. Is there a problem with it? Is the problem my end? The rest of the tracks seem to be OK.

    Thanks

  • David says:

    Sorry – I should have said, it is the 24-bit FLAC version that I have.

  • Matt says:

    Besides the Transporter from Slim (now Logitech) Olive seems to develop a 24-bit music server (www.olive.us) …

  • Josh says:

    Flac shmac … it is still a PCM format. I have been quite impressed with some 96 kHz/24 bit recordings, but the most impressive recordings are pure DSD recordings on SACD. The article also seems to place an undue emphasis on bit rates –> it seems to me that the real gains in removing digital artifacts comes from high sampling rates … not from high bit rates. There is also vastly more hi-rez content out there on SACD (over 6000 titles now) than in FLAC (which is currently in the low hundreds).

  • Paul C says:

    Thank you for this excellent article B&W, and the useful advice it contains.

    I have to say that to my ears, the difference between 16bit/44 and 24bit/88 is enourmous. I am sitting here listening to a previously downloaded 24bit wma sampler including music which I would never normally listen to and it is like disovering music for the first time.

    This is the first time that the sound coming out of my speakers from a digital source is genuinely enjoyable.

  • Paul L says:

    I hope some people find this link interesting

    http://www.soundfountain.com/amb/cd25years.html

    24/88, 96 sounds to me more natural especially on live music recorded in places with certain acoustical character. Hi-End should about preserving the perfomance at the time and space I took as faithfully as possible, but there is always a limit to technology we use, …what is real, anyway.

  • Graham W says:

    Does anyone know if the MM-1 speakers have a 24 bit DAC. The only information I have found suggests that the DAC is 16 bit only.

  • Andrew Henry says:

    Aren’t HiRes fans in a bit of a conundrum at the moment? If we didn’t have cheap computer based music servers to connect to our hifi systems, we’d all have SACD or DVD-Audio players, but these days, downloadable content and music servers are all the rage, and what’s better than ripping your entire music collection to a computer so you don’t have to buy an expensive CD changer or keep swapping discs??

    But you can’t rip SACD!! Is it worth the purcahse if you want a computer based music library? You can rip DVD-Audio… just… with a moderate splattering of hassle (even multi channel). But it’s certainly not as “accessible” to rip HiRes music discs as it is to rip standard CD’s. Isn’t the “best” option, the most “future-proof” one, to download HiRes FLACs where available?

    I have a Pioneer DVD player that’s nearly 10 yrs old that can play DVD-A and SACD, and I’ve just started buying HiRes discs (1-2 DVD-A’s and a few more SACD’s, but I know I can never have these in my digital music collection, so it puts me off buying more physical discs.

    Anyone else have this problem?

  • SusannaGrant says:

    Hi Graham,
    The MM-1s do have a 16-bit DAC.
    Susanna

  • SusannaGrant says:

    Hi Graham,
    You’re correct, the MM-1s do have a 16-bit DAC.
    Susanna

  • Jan says:

    @ Andrew Henry

    Hello Andy, I am convinced that there should be no problem to record your hi-res albs to your computer collection. Your CD/SACD certainly has a digital output has it? If so you can easily connect the player with pc using some audio card such as M-AUDIO to grab your hi-res signal.

    Jan

  • Seb says:

    @Jan : Sony forbids SACD players from offering any hi-res digital output (IIRC, they’re limited to 16/44.1).

  • Matt says:

    Is it worth recording vinyl music to a 24bit digital flac file as oppose to 16bit? If it is can you explain why?

    The reason i ask is because i thought vinyl has a depth of around 12bit??

    Look forward to your comments :-)

  • Kyle W says:

    Why do the MM-1s have a 16-bit DAC? Maybe I should return mine! I assume that the internal DAC is 16-bit and that for everything except the couple 24-bit titles I have there is no difference. However, I got the MM-1s for their quality and also as a better DAC for my P5s! I guess I should have done some more research. So if I’m listening to the 2009 Beatles 24 bit FLACs and comparing them to the 16 bits does that mean that they will be the same or just almost the same? I assume the downconversion won’t end up with the same 16bits.

    I have been most impressed with the quality of B&W products. I trust them as a brand and company that anything I buy from them will be of the highest quality. More than that, very competitive on price. Comparing a B&W product to the others on the market for instance the Zeppelin and the Bose sound docks makes the Bose look like a complete rip off.

  • Andysor says:

    The way 24 bit vs. 16 bit is presented here is simply not true or, dare I say, honest.

    The sample point analogy shown in the graphic is completely misleading. Mathematically it is PROVABLE that ANY analogue waveform below the Nyquist limit for the frequency can be reproduced PERFECTLY! I’ll say that again: PERFECTLY!

    Bit depth only becomes relevant when discussing signal to noise ratio, also known as the noise floor. Through the process of dithering all quantisation noise, the inaccuracies introduced as a result of the sampling, are ELIMINATED and merely raise the noise floor.

    In other words the ONLY benefit of 24 bit is the increased dynamic range. 16 bit has a dynamic range after dithering of at least 96dB which can be increased using “noise shaped” dither. But this is irrelevant as no master recording in existence has a dynamic range of more than 70dB! Even in the most quiet of settings background noise is at least 30dB meaning that even with the dynamic range of 16 bit you would expose yourself to 126dB of SPL! That will make you deaf quite quickly.

    The only reason to use 24 bit for playback (this does not apply to recording) is if you need to subject yourself to more than 126dB of SPL.

    For a more in-depth explanation go to http://www.head-fi.org/forum/thread/415361/24bit-vs-16bit-the-myth-exploded

  • Richard says:

    Great article, thanks!
    I have a question regarding ripping CDs to FLAC files for computer storage/playback. Since a CD is 16-bit, I believe you should rip at that conversion (eg 16/42). If you try to convert to 24/96, etc, isn’t the conversion utility just using interpolation to add missing data? You can put data in what’s missing, you may be able to gain something but it probably won’t be that accurate. Please advise, and thanks :-)

  • Richard says:

    Whoop,s I think I should correct that, it’s “CDs to AIFF files” I believe.

  • Per Schunnesson says:

    Very good article,
    However, I want to add Musicbee to the list of media players as an cexcellent choice for Windows users. Musicbee offers, in my opinion, great sound quality and lots of good features.

  • Andy McGhee says:

    Sorry Andysor but that is entirely incorrect. Nyqusit-Shannon states that all frequency compnents can be recovered but not perfectly as perfectly suggests 0 distortion. If you take your statement to it’s logical conclusion then a 1 bit DAC (and I dont mean 1 bit as in delta-sigma) could perfectly reproduce any signals below half the sampling frequency is simply incorrect. Quantisation error cannot be ignored.

    The other big benefit to increasing both sampling rate and and bit depth is that filtering sampling noise from the output of the DAC becomes easier, and consequently produces less distortion in the filters.

    Hope that helps.

  • Andysor says:

    @Andy McGhee:
    As I thought I made clear in my post the “perfectly” refers to reproduction above the noise floor.

    Quantization error CAN be ignored if noise shaped dither allows us to hear the sound reproduced beyond our ears’ dynamic range while keeping dither noise below the noise floor.

    The main error in this article relates to the “resolution” argument. As “resolution” relates to frequency reproduction 44.1/16 provides all the resolution we need up to 20kHz, and in terms of dynamic range it provides all that our undamaged ears can tolerate.

    A low-pass filter is required for ANY DAC to avoid the errors introduced by exceeding the frequency limit, and the 20-22kHz range is more than enough to accomplish this transparently with today’s electronics. However, because 44.1/16 DACs are more mature in this implementation it is possible that there may be audible differences when simply switching between 2 streams on a dual mode DAC.

  • BrianBzed says:

    Resources like the one you mentioned here will be very useful to me! I will post a link to this page on my blog. I am sure my visitors will find that very useful.

  • iQuackington says:

    “FLAC is of course a free, open source system like MP3”

    There is nothing free nor open source about MP3! There are in fact a huge number of patent and licensing issues with it:

    http://en.wikipedia.org/wiki/MP3#Licensing_and_patent_issues

    Other than this, a very interesting article!

  • flo says:

    FLAC 24 is the way to go but unfortunately I’d love to know why most of the recordings via this site sound so flat!
    The only exception is “strangely” the Peter Gabriel Album… Which really sounds great.

    Don’t get me wrong, I like the concept of investing in artists but on the B&W site I would have expected top quality for all recordings.

    I have downloaded a few tracks from other sites (I won’t mention names) at full 24 bits 96k or 88k and they all sound stunning!

    All the LSO recordings are so flat, there is no life in them. I own Berlioz’s symphonie fantastique on CD from the LSO live collection and it sounds far better than the 24 bit recordings available on the site. Why? The venue is the same and I’m guessing the same team was involved in the recording.

    I ran the tracks through a spectrum analyser to see if I could spot something dodgy had occured during the convertion to FLAC and I found that all LSO tracks included a 16khz or 21khz spurious signal within the recording. What is this all about? My CD doesn’t show anything similar.

    I haven’t downloaded all the tracks but so far I’m nearing MP3 quality on most of them…

    On another point, why are the 24 bit flac files sampled at 48k? One more time the LSO files are SACD recording at 88k which isn’t a multiple of 48. There is bound to be some information loss during the convertion!

    I’m fortunate to own a very revealing system (I spent tens of thousands of pounds on it) and if something has been badly produced, it will show very quickly. I convert the files back to WAV and burn them on DVDs to play them from a PS audio transport, so no pollution from the PC… The PS audio will play anything up to 24bit 192k.

    So unless things improve, audiophiles please look elsewhere. It is sad because there really is some good music available.

  • Kevin Murphy says:

    Someone help me here. What is the point of all this refinement if it can only be played back on a computer system or hand-held device? Even with the MM1s the result is far from hi-fi. Similar to downloading Avatar to your Blackberry if you ask me. I am a professional classical musician and I agree the LSO sounds like crap on the Sound Society recording. Awful! I have only seen a handful of proper goodies meant to stream audio from a computer file to a proper hi-fi (yes, mine is a good one) but they were so expensive as to not really make sense. Again, what is the point? Someone please put me straight.

  • Johnny Rosenberg says:

    As usual there was a lot talk about how to play FLAC in MacOS X and Windows, but no one mentioned anything about GNU/Linux. Maybe it’s not very necessary though, since all distributions that I know of has full FLAC support out of the box.

  • Krzysztof K says:

    Compression is not an idea it is just technical work around for saving space on disk. Hearing audio from PC is not a solution for audiophile community becouse of technical disadvantages, preferences, custom or individual reasons. In addition typical of shelf audio componente like amplifiers base on obsolette electronic circuits. It cause sound degradation much more audible than noice and dynamic loose in old CD format – only 96dB (Best CD players has no more than 103dB) .

  • Miguel says:

    I have a question?
    Why don’t you offer 24-bit ALAC. ALAC supports 24-bit recordings, and I’ve made recordings all the way up to 96 Khz. An iPod classic won’t play music over 48 Khz, but you’ve got to admit that 24-bit 48 Khz is pretty good for an iPod.

    Whenever I find 24-bit FLACs I have to convert them with xACT (http://groups.yahoo.com/group/xACT-users ) to ALAC to play them on my iPod and iTunes. I’ve never had a problem or a loss (sonically speaking) by converting them using this program, but it would certainly make my life easier if you just offered 24-bit ALAC.

  • Sam Faris says:

    Can I use my Phillips Streamium to listen to 24-bit FLAC?

  • Philippe says:

    I agree with most of the comments saying that quality of sound is not what we can expect from B&w.
    Except Peter GAbriel.
    Have much better result with a top cd player!
    In addition sourd Lebel is often too l’os
    I suspect that I s hall not pay for another year

  • realalan says:

    Hifi fans will never stop buying cd (SACD, HDCD, AMCD) or vinyl over downloads- it’s the art work, the ‘hunt’ for an obscure album at a record shop. Digital is for iPods, and who need to change or search for a ‘favourite’ song. When wa the last time you’d listen to an album from beginning to end???

  • Stuart says:

    Digital timeline misses out HDCD format from Pacific Microsonics.

  • Cian says:

    It seems like an awful lot of trouble with buying the DAC and external soundcard for the sake of a rare 24 bit album. I have never really noticed any problems with 16-bit WMA-formatted recordings myself, even 320kbps MP3 usually sounds perfectly fine.

  • Pug says:

    i do a lot of recording in 24 and 32 bit as i have a DAW using Cubase and Ableton.
    the difference i find is the sampling frequency jumping from 44kHz to 96KHz. i normally record in 32bit. its actually not true that studios are recording in 16 bit. protools, logic, cubase etc; these DAWs have been at these higher resolutions for years. mixing engineers have all the files waiting and ready for us out there. on their computers. all they needed was the format standard and its here now with FLAC. the hifi market is now gearing the hardware towards it. there will son be a flood of post 2000 released material when DAWs became very common. the key to FLACs success is faster braodband availability and the publics willingness to buy online.

    rubbish about no difference in quality,,,, 24 bit 96kHz has depth in the soundstage like CD would never achieve for all types of music. its amazingly wonderful to hear with correct audio hardware. belive me. i hate having to reduce my recordings down to miserable 16bit444kHz its like flattening the music with a hammer.

  • Ziya says:

    for those of you who “do not hear any difference” between lossy and lossless sound file formats I recommend that they listen to “no love” of eminem featuring lil wayne. where is wayne (at the background) in lossy formats; no where, completely lost! personally I am waiting for the days of 24bit FLAC widely used for recordings to come….

  • Nick says:

    Regarding comments that this is only suitable for a computer / hand held devices. Musical fidelity m1 clic supports playing 24bit 192khz flac.. Simply put the files on a USB hdd/stick and plug it in you can use the same method with the logitech squeezebox touch both not requiring a pc any where near your hifi equipment. Unlike all the other products in the squeezebox range the touch has USB and sdcard support and can be browsed directly on the device.

  • Jan says:

    When you say “MUSIC” you do not say WHAT music! If it is an italian youngster squeaking w emaciated falsetto some purported “canzone napu…” then 16 bit is ample enen including shrieks from the audience. If it is Great Mass in c of WAM or Wagner or Mahler … Then i suspect evev 24bit is NOT enough. Anyway we have on record one of the top engineers of Deu. Gramoph. Having said that “to catch it ALL we need 28 or 30 bits”. So!
    I can certainly hear the difference very clearly esp. In violin tutti but also in highest registers of the piano which become the same as glass bells or the triangle. That,s why you shoul keep 96 along with 24 and not dwngrade it to 24bit/44kHz. Have you noticed that the Cello always sounds v. Good and the violin rarely ? 44kHz is ample for one and not enoughnfor high notes of the other.
    By was of an example of great success I quote the V violin concerto in A by Grumiaux rescussitated by Philips o. 24/96 CD – I wish any violin recorded today on 16/44 would come close to that old recording.
    Finally I am sadened to see somuch of MP3 and so little of FLaC.
    Yrs. Jan

  • sacd-man says:

    Why should I buy LSO here, when I can have the SACD and get the closest possible sound to the real thing?

  • Jaap says:

    Can someone explain me how to play a file downloaded from the Society Of Sound club in my iTunes (ON WINDOWS XP)? I want to play the flac files on my itunes on the ipod to the the Zepplin Air.
    Thanks

  • Susanna Grant says:

    Hi,

    FLAC files are not compatible with iTunes however the 16-bit ALAC (Apple lossless) files are fully compatible.
    If you are looking to play the 24-bit FLAC files you need to use a different player that is compatible, something like Songbird (getsongbird.com) would work.

    I hope this helps,
    Susanna Grant
    Society of Sound

  • Naveen says:

    Hi Susanna,

    I have recently purchased B&W MM-1 speakers and joined the Society of Sound. Yesterday I downloaded 24-bit FLAC files and played them in Songbird. To me I didn’t notice any difference between 24-bit FLAC and 16-bit ALC sound. I then converted 24-bit FLAC files to 24-bit ALC and played them in iTunes. And again I didn’t notice any difference between 16-bit ALC and 24-bit ALC. I connected my MM-1 through an USB port. Please let me know why I’m not experiencing the pleasure promised in 24-bit music files.

    Best regards
    Naveen

  • Bowers and Wilkins says:

    Dear Naveen,

    You won’t hear a difference as the MM-1’s do not support 24bit 96kHz files. For more information on how to best play your 24-bit files we suggest having a look at this post: http://blog.bowers-wilkins.com/music/discussing-formats/how-do-you-play-24-bit-flac-files/

    I hope this helps
    Kind regards,

    Frazer
    New Media Product Support

  • Dawn Wyngaard says:

    I want to play these on my hi-fi system using the DVD-Audio option. What software do I use to write 24 bit FLAC files to a DVD-Audio? And where do I find it?

  • Stephan Richter says:

    Hi Dawn,
    I’m using Cirlinca software (http://cirlinca.com/) for years now. Although ;-) / because you have to pay for it, it’s a perfect tool, easy to use and with sufficient functionality.
    Please note that flac files will be converted (automatically and w/o quality loss) to wav before being burned to the disc,
    Have fun –
    Stephan

  • Alex says:

    Indeed I am not thrilled with LSO recordings, but I cat tell for example that Syriana has an excellent soundstage!
    Full digital path down to speakers is the cheapest and extremely effective solution, Squeezbox Touch after November’s upgrade with WD HD, sound extremely realistic from its digital output to Tact 2150 down to my 802s. No attrack correction can give you mathematically correct music as from a HDD.
    Flac HD is what I was expecting a whole HIFi life, and after MP3 disaster to music, I try to buy to increase availability even if friends have the file. Fidelio came to HD Tracks, other good ones will come, other Society of Sound will appear, and we wil be at last happy that disc space made for films will be used by us for HD music!
    Realistic soundstage can be heard if you do not convert again the pure digital source you now have!
    Great ears of the world believe me you have fully realistic source at your cable’s end…

  • Mike says:

    Quote: ‘To understand this more clearly, it’s helpful to imagine taking a digital photo of a squiggly line. The more megapixels the camera that you use has, the higher resolution the snapshot will be and the smoother the squiggle will look. This is the visual equivalent of bit depth.’

    No, sorry it’s not. The number of megapixel (spatial resolution) a difital camera has is directly analagous to sample rate for a digitised audio signal (temporal resolution). Bit depth is directly analagous to bit depth, which is is actually the individual sample resolution. Your engineers must understand that!

    Mike

  • Spuudy says:

    Reply to Dawn Wyngaard re. burning 24-bit flac DVD-A:

    Check out the 24 bit audio FAQ at:: http://24bit.turtleside.com/

    It lists around a half-dozen options for creatinging DVD-A and a few for DVD-V discs, including some free programs, as well as answers to some other basic 24-bit questions. I’ve used Lplex with great success for creating DVD-V’s with stunning results, and these can be played on any DVD player…

    Spuddy

  • Paul C says:

    In the article above I read, “If you’re looking to build up a FLAC library of your existing CD collection, then a good starting point for ripping is……”, and wondered if someone could explain to me why a FLAC copy of a 16bit/44.1hertz CD should sound better than the original CD? Am I missing the point, or is it more about making copies for pc/mac playback?

  • bobbmd says:

    if i buy a halide design hd dac or a cambridge audio dacmagic plus dac and play my stored music (in flac,having ripped my dvdaudio collection to flac via dvdaudioextractor) via mediamonkey or jriver is it really going to sound far superior to listening with just a digital spdif out tomy avreceiver as i do now(my current pc is an ancient hpmediaplayer xp that really has a digital out!) i plan to have made a customized pc music storer with a gigabit ethernet sound card with spdif out and a usb of course and hdmi out does that make the external dacs redundant ot am i going to get goose bumps listening to the deads american beauty,workingmans dead(dvdaudio recordings) the doors the band and dylan and the stones insacd or is it going to be just great but not as good as listening to the originals on a sacd/dvdaudio changer i ues now thanks your thought/comments will be appreciated bobmd

  • David says:

    I use mostly 16/44 wav files and some 24/96 and even 24/192 all wav. To me a 16/44 wav sounds better by far than a 24/96 flac. The 24/192 wav files often seem like too much detail and become tiring.
    Using a top quality DAP playing wav files at 24/96 feeding a decent tube amp and full range single driver horns I am in audio heaven. Shame there’re too many crap recordings revealed by this set up.

  • Sergey says:

    Very nice article! However, can anyone explain to me, whether FLAC files are decoded on the fly while playback (e.g. in Winamp) or the sound which ultimately reaches the speakers is still compressed? I’m asking this because when I play a FLAC file in Winamp or at my Squeezebox Duet, the info page shows a bit rate in a range of 1000-1500kbit/s, while when playing the same track but decoded before (i.e. in wav), the bitrate is considerably higher (i.e. more than 2000kbit/s). Can’t really understand this, since the FLAC is defined as a lossless format. Does it mean that if I want stream source sound to speakers, I should do it via wav format, while using FLAC for archiving purposes only? Thanks.

  • Florianinside says:

    Hi (unnamed?) Author of this brilliant article,
    what is with streaming 24bit-alac(-converted-flacs) to an zeppelin air? Does this make any sense at all?

  • Bowers & Wilkins says:

    Dear Florianinside
    The Zeppelin Air can support up to 24-bit, 96kHz.
    Kind regards
    Bowers & Wilkins

  • Hildy says:

    Other than being locked into an OS are there any pitfalls to converting to ALAC and streaming via @TV for a optical connection to one’s audio system? I made the FLAC/ALAC choice in 2004 thinking that Apple provided the best solution for me at the time. It is my understanding that if I ever decide to return to FLAC there should be no loss due to file conversion, is that incorrect?

    For me the move to lossless was only really constrained by disc space, but disc space has since caught up. My experience is that the biggest barrier is companies like Amazon/iTunes refusing to sell digital music in lossless format. As such I’ve limited my purchases to CD until things change. My other concern is the complete stall by portable music players at 160GB (iPod) as they transfer from spinning disc memory to flash memory. My 3-5 star list in iTunes is over 300 GB and I still can’t take the whole thing with me even though memory has grown substantially.

    Once we convince the sales channel to enable their customers we’ll be off to the races. It just seems to me that companies like Apple have an incremental upgrade business model where they upgrade formats incrementally charging customers to upgrade their library with every step.

    Not to mention the push for cloud based music libraries. If cloud based systems take enough market share there will be limited push to increase file size as that will equate to a need for increased server capacity and bandwidth.

    I’d love to see the industry start treating customers like they aren’t mindless drones willing to pay for what ever they decide to sell us.

    Hildy

  • Chris says:

    “Theoretically, 24-bit digital has a resolution of 144dB, which is more than enough (130dB is the threshold of pain for the human ear)…”

    You’re confusing signal to noise ratio with volume. 24 bits provides a signal to noise ratio of 144dB, which has nothing to do with ear pain, which is a function of volume.

  • Marek z Gliwic says:

    1. Ripping tracks from standards CDs and then converting them to FLAC is pointless! It is like simple up-sampling 16>24 when really acquired data are 16 bit with all limits of 16 bits format.
    So we need 24 (or better) originally recorded data/music, then it may be played using FLAC players.
    2. Of course ripping CD tracks is OK when one uses something like Exact Audio Copy – it is much better, usues full 16 bit resolution, make all rippingmore comprehensive then simple OS based rippers
    3. One of best players is AIMP, forget any Microsoft or Apple solutions.
    4. For playing originally 24 bit recorded music use external DAC’s, or optical digital output of your PC.
    5. All discussion about human 20Hz – 20kHz hearing range is just stupid. It is Dark Ages method,- human audio perception tested for single sine wave! Each millisecond of music contains frequency spectrum much wider than 100kHz. Use high sampling rate recording system, drop a glass on stone floor and see Fourier spectrum.
    Then listen to this sound using 44.1kHz sampling rate and compare it when 192kHz is used. Analogue system had frequency range much wider than 100kHz. It is widely known for audio specialist, nobody knows why still not widely applied. Sorry, I am bored with all of these…. Sorry for mistakes :)

  • Nicholas says:

    Wonderful article. Puts things in perspective so well. 24bit 96k has always been my favourite audio playback But as stated above its just so limited because of the expense. I hope it will one day become the bench mark for all CDs weeding out 16bit 44.1k. Thanks again B&W!!

  • Kyle says:

    First, lets stop muddying the water: MP3 for all intents and purposes is free and open. There is nothing even close to as ubiquitous and readily available. Every music player plays it, virtually every audio software package with any credibility will encode it. So let’s stop pretending otherwise. Second, while FLAC may be far superior to MP3, it needs a separate player and encoder that frankly, is a pain in the ass to get. I still, even after reading this article, cannot find a link to any resource that will provide me with a FLAC encoder for my music. For these reasons, FLAC has failed and will continue to. When it has grown up and is included in the major music players like iTunes, winamp and WMP, it will own MP3. But until then, it is just another useless hobby codec.

  • Mark says:

    This piece http://www.kenrockwell.com/audio/why-cds-sound-great.htm flat out contradicts some of the claims made for 24-bit and SACD. Not sure who’s right!

  • Tyler says:

    Sorry Kyle, I have to disagree with you here. FLAC may not be as widely adopted as MP3 yet, but wait another 5 or 6 years and I think you’ll change your mind. MP3 became popular because it was the first format that allowed people to listen to and transport music without a physical package. Back in the year 2003, MP3’s allowed me to download and stream music that I previously had no access to.

    Broadband is just now reaching speeds that let me download and listen to 40MB FLAC files rather than 4MB MP3 files, and storage is cheap enough to let me take them with me. iTunes doesn’t support FLAC simply because Apple is financially invested in their own proprietary lossless codec (ALAC). Basically all other quality media players support FLAC files. If Apple does begin to support FLAC, it will only be after people start moving away from their software because of it.

    It took me all of 30 seconds on google to find a fully functional FLAC encoder for Mac, Windows, and Linux. Seems like you failed here, not FLAC.

  • Chas says:

    I sincerely hope the engineers at B&W have a better understanding of physics than the folks that wrote this article.

    See here for an actual scientific discussion of HD audio: http://people.xiph.org/~xiphmont/demo/neil-young.html

    Spoiler alert: it is a mathematical fact that there are no benefits to 24 bit (or higher) depth or 192khz (or higher) sampling rates, and ironically in the case of higher sampling rates there are potential major drawbacks. Any differences you hear are either placebo effects or the result of better mastering often done for HD downloads.

  • Pile of nonsense says:

    “Its 16/44.1 specification boasts a dynamic range of around 96dB, which is ample for most music work.”

    Actually, with proper dithering, 16-bit audio has a dynamic range of 120 dB, which is plenty for *all* music work. 24-bit is only relevant to recording, not to playback.

    http://people.xiph.org/~xiphmont/demo/neil-young.html

  • Mon says:

    Why is it that you are offering the LSO albums in 24bit 48kHz quality while they can be bought big by Qobuz in 24bit 96kHz or even 192KHz quality ? The Qobuz downloads sound more dynamic and open.

  • Patrick Butler says:

    You should have a listen for yourself. Personally, I’ve found that the differences are very system dependent. One with a high noise floor tends to mask differences that are quite apparent in another system with a lower noise floor.

    As for the listening test cited in your link, if we really want to be intellectually honest about abx testing, then we can only conclude that the results of a particular test apply only to that test.

  • Juan says:

    I downloaded FLAC24 files from the Society of Sound, as they were the best-quality formats, but I mistakenly used them to burn CD’s, as my burning software allowed, believing 24 bits might be common to both channels or something similar.

    Now I know I can’t burn 24-bit FLAC files to CD and my burning software probably translates them.

    To my not-so-trained ear the CD’s sound crystal clear (SOS’s master quality is easily noticed) but I wonder wether burning 16-bit FLAC without any conversion would be better (how good is my burning software when reducing music resolution?).

    Any suggestions about this? May be no ear will ever notice the difference, as long as my burner isn’t awfully buggy?

    Cheers!

  • Juan says:

    I was thinking about the problem of lack of resolution for low-amplitude audio signals and found that using floating-point numbers instead of integral ones could perfectly do the trick and searched for floating-point audio codecs.

    As expected I found tons of information about floating-point audio codecs. These not only would have good resolution for any amplitude, but would also make (almost) irrelevant scaling audio coding so that it doesn’t overflow. Moreover, using denormalized numbers would avoid the gap around zero, but not using them and adequately scaling so that noise is always inside that gap and signal outside it, could even allow to trivially trim all noise out from the audio, while simplifying floating-point representation and handling.

    Sorry for so much technical information, but the point is, has the Society of Sound thought about using floating point for their digital recordings whenver a stable-enough (and high-quality) codec is available? What do they think about this?

  • BloodyGus says:

    Several websites offer you sample tracks of 16- and 24-bit FLAC tracks for you to test your playback facilities. I’ve listened to both and can’t say I notice much difference. Maybe it’s because my laptop is a fairly ordinary Sony Vaio, recently end-of-line. and you need a 64-bit computer to play 24-bit.FLAC files successfully. Otherwise it would seem to be a waste of money paying the extra for 24-bit when 16-bit is just as good. I hope somebody with more technical knowledge will advise me.

  • Mike says:

    Many years amassing ‘bootleg’ recordings recorded in FLAC.

    More recently the advent of being able to access sampled LP collections using very good TT’s and cartridges and converting to 24-bit FLAC files.

    Tto give myself the best opportunity to listen to these source files, led me to purchasing a Dune HD Base 3D media player.

    24-bit FLAC in (HDD, network, USB) outputted via RCA analogue into a 30W valve amp and then into home made open baffle speakers (vintage 18 inch drivers & all manner of other goodies swapped in and out to compliment).

    Regardless of where anyone else sits on the idea of what I am meant to be able to hear or whether there should be any diffreence immediately obvious to me.

    I say that in my experience, I am able to close my eyes and reach out and touch the musicians who are there in my lounge with me!

    24-bit recordings have transformed my appreciation of listening to music.

    For me they are so much more than cd ever was.

    Whether anyone agrees with me or not, means nothing.

    For me on my journey, it is, for now, the duck’s nuts!

    :)

  • Confused on FLAC says:

    I’m a liitle confused. The Society of Sound article and passage by Paul Rigby makes sense, I have many B&W products and trust in the engineering and quality and by extrapolation what I read within the Society. I am troubled by the statements and link that Pile of Nosense has posted above and have found his/her expositions elsewhere on the net.

    Not being an expert here, I must defer to Bowers and Wilkins as I deem that the engineers and scientists within the company are the experts. However the musings of Pile of Nonesens do seem to make sense. Can we get a B&W expert testimony on FLAC and the associated sample rates dithering etc.

    Is 24-Bit FLAC only relevant to recording and not playback? Do the higher sample rates really make a difference?

  • Bowers & Wilkins says:

    We are on it and will get back to you.

  • Bill says:

    “Confused on FLAC” why don’t you listen to your ears? Ignore what others say, if you prefer 24bit recordings over 16bit recordings then whatever anyone else says is totally irrelevant.
    On my computer I have Creative Gigaworks T40 speakers with no special soundcard and I am able to tell the difference between 16 & 24bit recordings and it’s even easier on my hifi.
    The figures quoted for dynamic range on CD are theoretical and rely on all 16 bits being used on a recording at some point. But if bits are not being used???? This is a common problem in digital systems and your camera will also suffer from not obtaining the max. theoretical dynamic ranges.
    I think many early CD masterings of analog material were produced by engineers frightened of running out of bits at the top end of 64k and so recorded at much too low a level. That and some of the analog sources that were used, apparently the first ‘Dark Side of the Moon’ CD used a Compact Cassette master?????!!!!
    So use you ears!
    btw Shostakovich 7 is one of the greatest pieces, most beautiful pieces of music ever created. Do listen to the Gergiev version on the Society of Sound, it is stunning.

  • Andre le says:

    “If you’re looking to build up a FLAC library of your existing CD collection, then a good starting point for ripping is” (from above).
    Why would I want to do this? Wouldn’t this produce an exact copy of the audio limitations of my CDs and no real difference/improvement? ie “garbage in, garbage out”
    I’m struggling here. Help :)

  • Confused on FLAC says:

    Indeed I have noticed a difference in FLAC over CD. I’ve even done blind listening tests with my family(kind of fun). In almost every case the listener picks FLAC over CD. What is interesting though is many times Vinyl is selected over both. Vinyl can be very confusing as well- quality pressings, how the music is mastered, and from what source etc, all come to play. I would love to see a definitive guide to the best way to select all music especially as it pertains to Vinyl.

  • Peter Z. says:

    People must realize that there are a great number of variables when it comes to listening to recorded music. Aside from the vinyl cd difference, cds can and do sound different. what do i mean? Well lets use Pink Floyd’s Animals for example. There are different cd pressings for many countries, Yes they are all cd and in theory should sound the same but in fact they do not. Remastered? Well thats another factor too. You cannot say that flac sounds better than cd or vice versa because of the variables. I have noticed differences between the same flac file (ripped from my own cds using exact audio copy) when played on different computers. Again it should sound the same but it just does not. i compensate for the differences in sound to a degree by using high quality headphones with their own respective sound signatures. I also use various sound enhancement software at times. Yes I know there are purists out there that would cringe at the thought of this but hey it works for me. I have not noticed any difference in sound when listening to a flac file that has been compressed at different levels. Yes i have listened to 24 bit files that sound great but i have also listened to 320k mp3s that sounded dynamite. Now i am not saying that this happens often but it still happens. Having said all this, in closing i would like to say just enjoy the music!

  • Oscar says:

    “Bowers & Wilkins says:
    We are on it and will get back to you.”

    Is there any official B&W response to previously mentioned analyses which point out that high-resolution audio files are just marginally superior to 16/44.1 CD quality for playback?

    Ethan Winer | Audio Myths Workshop · http://www.youtube.com/watch?v=BYTlN6wjcvQ
    Monty | 24/192 Music Downloads …and why they make no sense · http://people.xiph.org/~xiphmont/demo/neil-young.html
    Ken Rockwell | Why Compact Discs Sound Great · http://www.kenrockwell.com/audio/why-cds-sound-great.htm

    I think we all are eagerly waiting to read an in-depth explanation from an expert at B&W…

  • Michael Fricker says:

    I too am interested in B&W’s views on the merits of 24/48 over 16/44.1 (CD quality).

    I love the quality of SoS recordings and IMO are the best I have ever heard, however, my understanding gained from papers by experts in the field such as Dan Lavry and Bob Katz (google them) is that the extra dynamic range of 24 bits is preferred as it allows greater contingency for error and exceeding limits during the recording and mastering processes but with correct dithering and noise shaping quantisation noise can be pushed out of perceptually important frequency bands to perceptually less important frequency bands (>20 kHz) so that on playback 16 bit will sound exactly the same and is all that is required. All the extra bits do is increase the dynamic range from 96 db to 144 db, however, even the most dynamic classical recordings do not exceed 60db.

    On the sample rate side of things 20–20,000 Hz covers the range of human hearing. In accordance with the Nyquist theorem audio waveforms need to be sampled at twice that, hence we have rates such as 44.1 kHz (CD), 48 kHz (professional audio), 88.2 kHz, or 96 kHz. Sampling rates higher than about 50 kHz cannot supply more usable information for human listeners. Early professional audio equipment manufacturers chose sampling rates in the region of 50 kHz for this reason.

    There has been an industry trend towards sampling rates well beyond the basic requirements such as 96 kHz and even 192 kHz. This is in contrast with laboratory experiments, which have failed to show that ultrasonic frequencies are audible to human observers; however, in some cases ultrasonic sounds do interact with and modulate the audible part of the frequency spectrum (intermodulation distortion). It is noteworthy that intermodulation distortion is not present in the live audio and so it represents an artificial coloration to the live sound. One advantage of higher sampling rates is that they can relax the low-pass filter design requirements for ADCs and DACs, but with modern oversampling sigma-delta converters this advantage is less important.

    The Audio Engineering Society recommends 48 kHz sample rate for most applications but gives recognition to 44.1 kHz for Compact Disc and other consumer uses, 32 kHz for transmission-related application, and 96 kHz for higher bandwidth or relaxed anti-aliasing filtering.

    My understanding is that consensus within the professional music industry is to use 24 bits / 88.2khz for recording/mastering and then dither with noise sampling down to 16 bits / 44.1 for playback.

    The bottom line IMO is that CD quality is actually sufficient for playback and the differences people hear (including myself) is due to a whole host of variables including:

    – different mastering of the same album for different formats
    – equipment limitations such as DACs been better at processing different sample rates
    – poor dithering techniques
    – placebo effect resulting in people justifying the extra expense or being afraid of not hearing what they “should”

    From what I have gleamed it is the skill and expertise of the recording engineer and producer, who together determine and control the positioning of microphones and the mixing and mastering techniques, as well as the quality of the equipment that will have a much greater effect on the “sound” than 24 bit vs 16 bit and 96khz vs 44.1khz.

    My concern is that the public are being asked to pay extra for the original 24/48 or 24/96 recording and less for CD quality when in fact the recording is originally done at these higher bit and sample rates and then dithered down – i.e. CD quality requires more work, not less, and should be less expensive. Go figure.

  • SonicTruth says:

    Just some advice when comparing delivery formats(low-bitrate lossy, high-bitrate lossy, lossless(CD), high-res, etc):

    The differences between those digital delivery formats is infinitesimal compared to differences in mixing and/or mastering before exporting to those formats. In other words, there have been many cases, I.E.: http://www.audiostream.com/content/high-resolution-downloads-nevermind where a new 24bit/96k sampling version of a classic album has been released, where the end-listener or reviewer has been less than impressed.

    You are all probably familiar with the term “loudness war”, where the apparent loudness of the audio on a CD is boosted by chopping off 4-8dB of the dynamic transients, additional compression might be applied, and gain is applied to bring the resulting product to peak within 0.5 of full scale(0dBfs).

    Sure, it’s LOUDER, it’s in your face, but after a while, something makes you turn it down, or press >> to skip to the next track. And you don’t know why.

    Well now you do.

    Buyer beware: Some – not all – high-res downloads and SACDs have been brick-walled to make them louder, at the expense of musicality and dynamic punch.

    In summary, a high-res reissue of a legacy artist’s catalog(pre-1990) should remain sonically transparent to the original, assuming original master tapes were sourced to create that high res.

  • Mike CJ says:

    I think the “placebo” effect is a placebo effect itself. It is like saying a 32 bit color image is not better than a 16 bit color image. Only, it is much harder to manipulate the viewer into thinking they didn’t see what they saw.

    I have the 24/44k version of the Beatles tunes as well as the 16/44 k version. When I have all my tracks playing randomly, including those by other artists, I can ALWAYS tell without looking when the 24 bit version of a Beatles tune comes up. And, cross checking on the playing device, I’m never wrong. So, to me, there is a difference and not in my head.

    Same with other recordings. Can’t think of the name at the moment, but the 24/192 blu ray disc from Norway (I believe) sounds light years better in tone, presence, depth and realism than any CD recording I’ve ever heard, and it isn’t close.

  • portotalley says:

    According to wikipedia, iTunes can currently read, write and convert between MP3, AIFF, WAV, MPEG-4, AAC and Apple Lossless (.m4a). That is we still cannot import and play FLAC in iTunes.

    In order to play FLAC in iTunes, we had better convert FLAC to iTunes more supported audio format with a third party software like dbpoweramp, fluck, idealshare videogo.

    I use iDealshare VideoGo which has both Mac and Windows to convert FLAC to Apple Lossless, MP3, WAV, AAC, AIFF. And it works very well.

    Just google search How to Use iDealshare VideoGo to Convert FLAC to iTunes, you will find a detailed guide.

  • John H says:

    I have audio equipment that can play 24 bit FLAC files at up to 48kHz sample rate with an audio bit rate of up to around 2Mbps. I also have a digital pre amp that up samples a PCM stream to twice the sample rate. I subscribed to Society of Sound because the web site said that their 24 bit FLAC files were at a maximum of 48kHz, but I now find that many of the downloads are at 96 kHz sample rate. In order to play these I convert them to 48kHz sample rate using dPowerAmp, but I still have a problem with the audio bit rate, which is often much higher than 2 Mbps, even if I convert the sample rate down to 44.1KHz. This means that I cannot play the download! Can anyone tell me how to reduce the audio bit rate on a FLAC 24 file to under 2Mbps please? I notice that the FLAC 24 download of Laura Mluva is at about 1.4Mbps, and this plays very clearly on my equipment. When I can play FLAC24 files they reproduce beautifully, but I haven’t quite sorted yet how to convert Society of Sound FLAC 24 files so that I can play them. I welcome any advise. Too high audio bit rates are my problem.

  • SonicTruth says:

    Something to consider: The original Redbook CD spec, based on the Nyquist Theorem, adequately covers the full range of human hearing.

    In other words, decisions made at the mastering stage will impact what you hear significantly more than going from 16bit 44.1kHz sampling rate digital audio to high res – 24bit 96k or higher.

    There are unfortunately far too many examples out there of high-res reissues out there that have been excessively compressed and/or brickwall limited compared to their initial release on regular CD.

    Be forewarned.

  • marco says:

    but for plays the Tubular Bells Studio-Quality 5.1 FLAC, which app players use with the Mac?

  • Bill says:

    Michael Fricker quoting the Nyquist theorem and stating that it protects us from sampling errors is quite wrong.
    First the Nyquist theorem is a theorem and discusses how a sine wave an be reconstructed from digital samples NOT a music waveform but a sine wave.
    Secondly the “samples” mentioned above are taken as measurements at points along this sine wave they are NOT, I repeat NOT samples in the true sense. They are in effect Real number values calculated from the sine wave function and so would be 100% accurate (unlimited dec places). When we talk about samples in digital music reproduction we are talking about something that is measured and then approximated – ie a floating point converted to a 2 byte (16bit) or 3 byte (24bit) integer. So you cannot apply Nyquist in an adhoc fashion.

    And that is without mentioning Aliasing errors!

    Someone mentioned floating point codecs above. The first thing about representing numbers in floating point is that they use many more bits and bytes than do integers, as I said above a 16bit CD will use 2 byte integers, whereas an implementation of Floating Point numbers might require 8 (min required for an effective FLOAT representation) or more bytes per number. Secondly there is the inescapable fact that presently PCM is integer based.

    Floating point are an added complication that we don’t need in general. Yes much audio software does use floating point, eg Audacity, but that is because the processing of the waveform becomes less distorted because you are not converting to an integer at each operation – just at the end. Rather like not working on a JPEG in a graphics package and using TIFF and then converting to JPEG at the end. But whenever we do calculation on a computer and storing the result then we ALWAYS are taking an approximation. Computers cannot truly store all Real numbers, remember the square root of 2 or pi have an infinite number of decimal places. Computers don’t like such things as infinite.

    If we want greater resolution then just use integers with more bytes CD (2bytes)–>24but(3 bytes)–>more & more bytes. A Real number using the same number of bytes as an integer can NOT provide greater resolution than the integer form. The reason us programmers tend to use REALS is that it allows us to scale numbers, eg we cannot store the number 0.24 as an integer. Well we can, we could round it down to 0 – but that wont’ help a lot. What we could say is that the number is 24 (an integer) divided by 100 (an integer) – or in fact 10 to the power 2. That is basically what a floating point number is. In a computer everything comes down to integers in the end!

  • Bill says:

    John H – try Audacity.

    Sonictruth – your first statement does NOT prove your second statement. Although I agree, if the mastering is poor then it don’t matter how many bits are used. But if the mastering is good then 24bits will sound better than 16bits. As I said above do NOT rely on the Nyquist theorem in a practical situation.

  • Bill says:

    Actually thinking about your problem John H, 16/44.1 is about 1.4 Mbits/s, so I expect that even 24/44.1 would be greater than your limit of 2 Mbits/s.

    Once or twice SoS have put out 16bit files as 24bit files by mistake and have very quickly rectified this mistake when informed. I wonder if the ones that work for you are actually 16bit.

    Obviously you need a new piece of kit but in the meantime download the 24bit and 16bit versions and play the 16bit until you get your new hardware. Sorry mate!

  • Fred says:

    “First the Nyquist theorem is a theorem and discusses how a sine wave an be reconstructed from digital samples NOT a music waveform but a sine wave.”

    This is wrong. Look it up.

  • Bill says:

    Sorry Fred it is NOT wrong. The Nyquist theorem reconstructs a sine wave from values actually on the sine wave. It’s a bit like Pythagoras, you can do the maths to solve a right-angled triangle but in practice you will never see a right angle triangle – that is where one angle is 90.00000…(infinite number of zeros) degrees.
    It’s the same for the Nyquist theorem, I don’t need to look it up bit I am afraid you do.

    Try https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem

    To discuss sampling you then need to apply statistics but the actual theory only holds for functions not practical data. How could it possibly hold when we don’t know (or specify) the bit depth?

  • Bill says:

    Actually Fred it’s a shame we can’t discuss this face-to-face because the Nyquist Theorem is not that difficult to understand and I have a beautiful computer animation that I use as a teaching aid in a second year maths course. I can then quite easily take them through the derivation of the theorem.

  • jordans on fire says:

    Thanks, I’ve been hunting for info about this subject matter for ages and yours is the best I’ve discovered so far.
    jordans on fire http://www.autokm0.com/?shop=jordans-on-fire&id=4898

  • Robert Mostyn says:

    I have downloaded the Peter Gabriel concert files, changed my OS X system settings to play high resolution audio, now what do I do? I tried importing into iTunes and nothing happened. I have the files but cannot do anything with them!

  • Liz says:

    Hi Robert

    I have sent you an email regarding this.

    Kind regards

    Liz

  • Jorge says:

    I can not reproduce it on my iMac :/

  • Mr.M says:

    The danger with this sort of post is that whilst it aims to be informative for the lay person and attempt to rationalise the benefits of “more bits” they tend to end up in technical wikiloop discussions about mathematics when in actual fact the best judge is your ears and your wallet.
    Short of selling a kidney and spending £30k on some exotic setup, you can gain some listening pleasure from rather modest investments in kit to improve the path between FLAC and earlobes.
    I have an above average setup both for home listening and on the go music, I’d say the kit that made the greatest overall improvement (without getting into things like PRAT or “dynamics”) was the DAC.
    I currently use an iFi iDSD Nano most of the time both home and on the go (it has it’s own battery and works with phones put simply) and this does a good job getting the music to the speakers or headphones without messing about with the bits in the middle.
    There is a world of difference between modern higher bit rate mastering such as many of the releases on SoS and a lot of CD’s and older recordings which were never mastered properly to start with and attempting to improve them by artificial or mathematical means is an exercise in frustration best avoided.
    You can have a well mastered CD that sounds good and the same CD (different version/mastering steps) that sounds subpar, having CD’s as files in one place or on multiple hard drives is the biggest gain overall, that simplified library plus a decent DAC and you’re all set. It will usually come down to pocket depth and budget but I’d argue that you can achieve great results for modest outlay, perhaps a few hundred £ would give you something you can genuinely feel a difference in enjoyment from. Keep it simple, shop to a sensible budget, and most importantly enjoy the best bit more, the music.

  • Mike says:

    HI Everyone,

    I’m using a mid 2011 mac mini, itunes and bitperfect into a new zeppelin air via usb. However, bitperfect and my mac is telling me that the zeppelin air is only 16 bit at the top end. Is the USB the problem. Should I be using optical audio instead? Any advice?

  • Matias Elena says:

    Hi Mike,
    the same problem here…USB in in the zeppelin air is 16 bit 44 Khz. Use the optical in and you can use high res audio. Remember to change the midi setting or use a program that deal with it. I use Audirvana plus.
    Hope to help you. Kind regards.

    Matías.

  • Graham Macdonald says:

    So I downloaded the Tubular Bells 5.1 track but have not found a way to play it in anything other than 2 channel even using VLC and hooked to my receiver via HDMI – any ideas?

  • Salvo says:

    I truly appreciate this “guide”! Thank you!
    Complementarily to the topic of digital audio tools I’ve found interesting the following article about effective benefits of hi-res digital encodings: https://xiph.org/~xiphmont/demo/neil-young.html
    Several arguments are reported there regarding audibility of extremely high-quality samplings (e.g. 24bit/192kHz). So we can also make an idea of which kind of lossless file is worth using and collecting.

  • Julio Aisemberg says:

    I am an entusiastic classical music listener. I am a Mac man, and listen to all my music in a iPod Classic. i would like to improve the quality of sound, but I do not find any up to date guide to FLAC downloading. I hope i can start soon with the best sound. Any help, please.

  • Daniel Burge says:

    My personal experience is that the music that comes from a cd played on a Naim CD555 has absolutely nothing in common with the phrase ‘almost good enough’.

Add a comment

We welcome debate within Society of Sound, but please keep it friendly, respectful and relevant. We have a few house rules which we ask you to abide by to keep the debate intelligent. Read more.
Product enquiry or support issue? Please click here.

Related posts

Bowers & Wilkins Speakers thumbnail

How Many Speakers?

After listening to the podcast discussion about surround sound in the Lab it got us wondering just how many speakers is enough for … Read more

Bowers & Wilkins Speakers thumbnail

Computer Games: The Future of Surround Sound?

Computer games were once just for kids; many of those kids are in their 30s and 40s now, and many of them are still playing games. … Read more

Bowers & Wilkins Speakers thumbnail

Apple Lossless Confusion

There’s a lot of confusion about audio encoding, and how to get the best out of your digital music. Obviously we at B&W are … Read more