May 28, 2016
 by Paul McGowan

In yesterday's post I had mentioned that I get questions on the two compression types, lossy and lossless. Some have questioned the term compression and wonder, if it's compressed how can it be lossless? Is it a marketing scam or is it real? Probably worth spending a few minutes on this subject.

Lossless is indeed lossless. This means an identical copy can be extracted once uncompressed.

If you've ever gotten a ZIP file you're already familiar with compressing data. You would never expect to unpack a ZIP file and have your picture, text or document anything less than perfect. Audio compression of the lossless nature is the same.

The best known lossless compression schemes, FLAC (Free Lossless Audio Codec) and ALAC (Apple Lossless Audio Codec), are both able to compress a music file into about half the space of the original. Uncompress it and the bits are identical. There are other lesser known schemes just as accurate too, like MLP (Meridian Lossless Packing), WavPack, TAK, Monkey's, WMA, OptimFROG, TTA, among others. The list is long.

There is a lot of talk amongst Audiophiles that despite the fact lossless files are unquestionably bit perfect, they don't sound the same. How could that be? Well, for one thing, the amount of processing resources required to unpack a lossless file are far greater than those playing the original WAV file. If for no other reason, the extra number crunching impacts noise, jitter and increases power supply demands. So while the bits are identical, the playback demands are not.

If you download a compressed lossless file, uncompress it, and place it on your hard drive, that file will now sound identical to the same file on the same hard drive that had never been compressed. In other words, compressed or uncompressed the bits are the same.

Playback of compressed bits has different demands than playback of uncompressed bits. Hence, sound quality varies depending on the hardware reproducing it.

But that should be no surprise.

Subscribe to Paul's Posts

33 comments on “Lossless”

  1. Is there a theory why lossless flac and wav sound different? Can it really be the tagging included in flac?

    It's strange that in audio streaming it's still not possible to get original wav quality with tagged lossless formats.

    1. I tried to explain that. It's because of the extra processing of the FLAC file. If you convert the FLAC file to WAV first, like the computer has to do in order to play it at all, then play the WAV file, you won't hear any difference between the original WAV file and the uncompressed FLAC file.

      1. thanks, so Bridge II has to convert flac to wav internally, but not aiff to wav?

        if I convert all my flac files to aiff to be used by Bridge II, I'm done and on the level of wav?

          1. I just tried out aiff against flac and other than my below mentioned rememberance from comparing uncompressed flac with wav (mainly better bass from wav), now I heard mainyl more open, airy top end from aiff than from flac. I guess I'll convert all my files. I may have to use Jriver for tagging instead of MP3convert, which seems not to support aiff unfortunately. During conversion with dbpoweramp from flac to aiff, fortunately all tags remain.

            1. Miracle, miracle, miracle!
              You get different output from the same data.
              I wonder why.
              The decompressed flac file uses the same data pipes as the wav file.
              There is NO difference in data at the output!

              1. Just listen to it yourself!

                I recognized this between uncompressed flac and wav as well as aif, I asked Paul, why this can be and he answered as you can see above, that the renderer is not only busy uncompressing compressed flac in case it is used, but also busy converting uncompressed flac to PCM, which is not necessary with wav and aif. All these processes seem to affect sound quality. I can’t argue about the theory, as I asked for it myself, I can just discuss about what I hear.

                If you didn't try it already, try it and listen, if you tried it and can't hear anything, be lucky, you don't have to bother about a conversion, I’m just converting, it’s obvious.

                1. Well I've listened to the different file-types.
                  I heard no differences. Maybe my ears are not as good as yours.
                  To be sure that the result does not appear because of my worse ears I did even more.
                  I made spectral analysis of the playback.
                  There was no difference!
                  I also compared the data-stream when playing the flac file and the wav file. Result: No difference.
                  I admit that humans like me are not necessarily able to do multitasking - neither cooperative multitasking nor pre-emtive multitasking.
                  Your computer IS able to do so. It can do a number of processes in parallel.
                  But - maybe I'm wrong and you're right because I made some big mistakes.

                  1. I can’t tell if such analysis can prove an audible result or not, but maybe due to the conversion process, some part in the chain, independent of the signal purity, performs weaker and causes the audible difference.
                    Is far as I know there are many other audible aspects in Hifi, which are not measurable, so maybe one more here…but for sure Paul could deliver better arguments.

                  2. It's probably not you or your ears, bernd. More likely it's your system. Not all systems are setup for resolving small differences. I remember one of the better sounding systems I had, using great speakers and electronics, could not resolve small differences.

                    Turned out it was the speakers which made everything enjoyable, but obfuscated differences like these.

                    There's no question if you were to listen on mine you'd hear it immediately.

              2. This is the age old argument - why, if the data is the same (and it is) do they sound different (they do).

                As I have tried to show, a simple experiment answers the question. The differences only manifest themselves when you play a FLAC file vs. a WAV file in real time.

                Convert the FLAC file to WAV beforehand, and the differences between the (now) two WAV files are indistinguishable.

                1. Paul,
                  I do not have a multi-room system but several systems in different rooms to do the test.

                  From my collection of electronics: I have tube amps (Marantz model 7 / model 9, ARC SP11MkII / Classic 120) transistor amplifiers (Cello Palette / Encore, Pass Labs Aleph P / Aleph 0) and there are some others.
                  DACs in use: PS Audio Nu Wave, Benchmark DAC2, Exogal Comet +.
                  Loudspeakers: ESS AMT1B, Manger Z1 and custom built monitors by Bill Kieltyka.
                  I could not hear a difference with any combination. (BTW - the Manger Z1 has scale putting properties when it comes to resolution, maybe only topped at the absolute upper frequencies by the Heil Air-Motion-Transformer.)
                  I tried a last test with in ear monitors (Shure SE846) directly at the headphones-output of my ASUS ROG GX700VO laptop (dual boot with windows and linux) which utilizes an internal ESS Sabre DAC - again no difference.
                  I've done these tests with three friends of mine. Nobody could hear any difference.

                  Could it also be that your source (lappy, Macbook or so) makes the problems?

                  1. Just because you mentioned the "air motion transformer. My speakers use such a driver. Maybe this helps, but I think it's not a precondition.

                    1. @jazznut
                      My ESS AMT1B loudspeakers utilize the original air motion transformers by Dr Oscar Heil.

  2. I would leave my downloaded files in WAV form if I knew how to attach cover art in iTunes, so I convert them to ALAC. Is there another library program, or is that the nature of WAV files? Then again, I use an Airport Express in my wireless setup (44.1kHz/16-bit), which is quite good enough for my almost 77 yr-old ears (some tinnitus, too - see below). Does it matter?

    I have some older CD's that are particularly jitter-prone. I used to take one of them (Handel:Water Music, Royal Fireworks, BPO, Kubelik, DG) to audition equipment just to hear how the equipment handled the jitter. Now with iTunes as the source, the jitter hash is practically gone from these recordings. Wonderful!

    1. I can't really hear a difference in my FLAC files and the original except sometimes the FLAC files seem better than the original. I know that seems impossible.
      Manniesm, JRiver will let you attach your cover art with WAV files.

      1. in my experience if you compare uncompressed flac to wav, wav has better bass (don't remember in which terms but I commented somewhere in the forum before). Compressed flac on the other hand is also noticeably worse in treble.

      1. Thanks to all that responded to my WAV-tagging question. I will certainly compare going from WAV to AIFF vs WAV to ALAC on my next download, but I suspect that at 44.1kHZ/16-bit, I won't be able to tell much difference.

        As to the Airport Express, Stereophile rates it Class D:

        "While the Airport Express works only with iTunes v4.6 or later (running on both PCs and Macs), is limited to 16-bit data, and functions only at a 44.1kHz sample rate, the combination of iTunes and the Airport Express offered an easy way to pipe CD-quality music around the entire home. "The beauty of this unassuming component," said JA, "is its S/PDIF data output, which allows the Airport Express to assume a respectable role in a true high-end audio system." However, its lack of an internal clock can lead to the first couple of seconds of songs being missed with DACs that are slow to lift their mutes. (Vol.28 No.5 WWW)."

    2. The airport express only receives upto 44.1 and some DACs sound better if the audio is up sampled before the DAC. Also, Apple somehow messed up the sound of these devices with a firmware update many years ago so their sound has a harsh sound on my system that can no longer be cleaned up by a device placed on its toslink output.

    1. Should have bought a Mac. 🙂

      Sorry, couldn't resist.

      You can use a program on your Windows machine like JRiver that allows WAV files to be tagged - or just use FLAC. FLAC files are excellent.

  3. JRiver can be set up to decompress lossless files while loading them into DRAM and then playing the uncompressed file from DRAM. Thus, there is extra CPU processing occurring only during the first few seconds of playback. This seems like it would make playing back lossless files equivalent to playing back the uncompressed versions (due to the effects of extra CPU processing) but I must admit that I haven't done any comparisons.

    1. I have tested my main computer using these files and find that my machine only uses about 4% of my CPU processing power when playing either WAV, DSD, FLAC or others. I thought ahead when buying my machine to what uses it was going to have for music and other playback. It was pricey at the time but has lasted now 5+ years with minimal issues (of course something will blow today now that I have said this).

      I recall another post some time back where processing power was talked about with ripping, that many felt that rips of any kind did not sound the same if the computer had to spend a lot of processing time on the ripping and saving process.

      Windows 7 PC with 64GB of RAM, dual quad-core Xeon processors running at 3Ghz, with 4 x 500GB 10,000 RPM drives. All commercial grade server quality components, and (ouch) weighs in at almost 75 pounds but once it got placed in the basement it just stays out of site controlling all my home devices.

      On another note, I was disappointed more people did not comment on or take the listening test and report results from yesterday's post. I wonder if it is because most people could not tell the difference and didn't want to publicly state so??

      1. About half a year ago I read a study about this test which revealed some serious flaws. Only one example: levels were not 100% equal. A pity I did not bookmark the site. And no idea if they fixed their files in the meantime to make them comparable.

  4. Thx everybody for answering my question.
    I have a PC and Mac(mini), but never tried AIFF on the PC, so I wasn't sure.
    Although I'm not a big Apple fan, I've used the Mac in the past for music files. AIFF, but also WAV, transferred
    from the PC, including cover art. And that works for me.
    And cover art is not so important for me, I'm mainly interested in the music. For me 99,99% of the fun.
    Said that, most of the time I just play "oldfashioned" cd's with the PWT/PWD.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram