WAV vs FLAC revisited

[Vincent Kars]

At the risk of stating the obvious, you do NOT want to sync the clocks in the capture to player. The whole point of this exercise is to see if timing of samples is changed. If DAC timing is locked to external source, then that can't happen (at least not fully).

Amir

I don’t understand.
A possible explanation for WAV / FLAC differences is that you feed the same bits to the DAC with differences in timing (jitter).
If you do an analog loopback (I assume the loopback on the Lynx is analog not digital) than one still capture differences in input jitter at the DAC.
What am I missing?

 

[Vincent Kars]

Mitchco
Just to make sure: the loopback on the Lynx is analog.
Correct?

 

[Bruce B]

the loopback on a Lynx AES16 is digital out/in

 

[Vincent Kars]

Thanks
That is indeed a good test for bit perfect but not for jitter issues.

 

[Mitchco]

Thanks
That is indeed a good test for bit perfect but not for jitter issues.

Hi Vincent, Bruce is right, that is diigital loopback. But that is why I also performed the other test at the analog outputs as that does cover off jitter and any other distortions/artifacts, including the analog output stage. Again, the difference file, at the analog outs, was -90 dB = inaudible, certainly relative to the program level.

Btw, I did update the post to show that Audio DiffMaker works as adverstised as I also compared FLAC to MP3 and we know what that sounds like. You can "hear" the results here: http://www.computeraudiophile.com/blogs/FLAC-vs-WAV-Part-2-Final-Results#comment-131768

In conclusion, just like my test comparing two bit-perfect music players: http://www.computeraudiophile.com/blogs/JRiver-vs-JPLAY-Test-Results there is no difference between FLAC or WAV or "lossless" files. If there is, then something is not configured right or outright broken in the signal chain. You can use Audio DiffMaker to verify or refute that on your own setup.

Cheers, Mitch

 

[rbbert]

Although I consider myself to be more objectivist than most audiophiles, I can't see how -90dB (in a null test) translates to "inaudible". That's clearly more and a more quantifiable difference than many things generally accepted as audible, like differences between CD players, cables, power cords, amplifiers, etc. If that is the measuring limit of your equipment, then there may or may not be a true difference, because it's my impression that much audio testing equipment measures down to -120 dB and more.

 

[Mitchco]

Although I consider myself to be more objectivist than most audiophiles, I can't see how -90dB (in a null test) translates to "inaudible". That's clearly more and a more quantifiable difference than many things generally accepted as audible, like differences between CD players, cables, power cords, amplifiers, etc. If that is the measuring limit of your equipment, then there may or may not be a true difference, because it's my impression that much audio testing equipment measures down to -120 dB and more.

Hi rbbert. I explained in the post that I was comparing in the analog domain and measuring at the analog line output of my Lynx L22 sound card. The 2nd recording/measuring laptop had a -86dB signal to noise ratio using the onboard sound chip (i.e. line amp and ADC). That's the limitation.

Did you download and listen to the -90dB difference file at the end of the post? It is interesting to hear what -90dB sounds like as I hear the noise of the recording laptop as the comparison totally nulls out. But I have to turn up my headphones to almost maximum level to hear anything at all. So, relative to near 0 dB program material level, -90dB is inaudible. That's my point.

However, I did perform the same test in the digital domain and removed the dependency of the 2nd recording laptop and analog domains: http://www.computeraudiophile.com/blogs/FLAC-vs-WAV-Part-2-Final-Results#comment-130591

Result = -161dB I think you will agree is inaudible.

 

[rbbert]

But of course the issue (if there is one) with on-the-fly FLAC playback is with timing issues at the D>A stage, so indeed it is the analog signals which need to be compared. I'm not saying there is a difference, just that your test isn't a good demonstration of lack of subtle differences.

 

[Bulldogger]

Neither of those two conditions would explain the uniform character difference I hear between files played real time as wav and flac. There is clearly a lack of detail when flac is played in real time, which I believe is the flac real time decoding engine. .

--Bill

I read there was no difference and ripped all of my disc to flac. I never bothered to test. Yesterday, a friend played several different files of the same CD, one wav the other flac. To my surprise, I could hear the difference for most of the recordings. On one, I could not for whatever reason. The flac had both a loss of detail and "air." I should have asked my friend before I started to rip. It would have save me some time.

 

[Bulldogger]

Another recent article:

http://www.audiostream.com/content/cut-flac

Hmmm, just read that article. That's my impressions exactly.

 

[Bruce B]

I read there was no difference and ripped all of my disc to flac. I never bothered to test. Yesterday, a friend played several different files of the same CD, one wav the other flac. To my surprise, I could hear the difference for most of the recordings. On one, I could not for whatever reason. The flac had both a loss of detail and "air." I should have asked my friend before I started to rip. It would have save me some time.

There could be alot of variables at work here. What did he use to make the FLAC files? What compression ratio did he use? Did you use checksum or Diffmaker? Maybe your software handles the files differently.

 

[garylkoh]

I read there was no difference and ripped all of my disc to flac. I never bothered to test. Yesterday, a friend played several different files of the same CD, one wav the other flac. To my surprise, I could hear the difference for most of the recordings. On one, I could not for whatever reason. The flac had both a loss of detail and "air." I should have asked my friend before I started to rip. It would have save me some time.

FLAC is a lossless compression. This means that even if you ripped to flac, you can "unpack" the flac files to wav, and there would be absolutely no difference if you had ripped directly to wav. You will get the same file (except for some header differences).

 

[Bruce B]

FLAC is a lossless compression. This means that even if you ripped to flac, you can "unpack" the flac files to wav, and there would be absolutely no difference if you had ripped directly to wav. You will get the same file (except for some header differences).

+1.. no difference!

 

[rbbert]

FLAC is a lossless compression. This means that even if you ripped to flac, you can "unpack" the flac files to wav, and there would be absolutely no difference if you had ripped directly to wav. You will get the same file (except for some header differences).

A point that often seems lost; no need to re-rip even if you prefer the sound of WAV over FLAC in your system.

 

[bblue]

I read there was no difference and ripped all of my disc to flac. I never bothered to test. Yesterday, a friend played several different files of the same CD, one wav the other flac. To my surprise, I could hear the difference for most of the recordings. On one, I could not for whatever reason. The flac had both a loss of detail and "air." I should have asked my friend before I started to rip. It would have save me some time.

Admittedly, it is a subtle difference and takes a very revealing playback system to show the differences, which are pretty much as you describe.

As I mentioned originally, it's NOT a problem with FLAC at all, as a FLAC file will uncompress to its exact original WAV. It's only when the FLAC is played back in real time with a plugin in most players, Foobar, Media Center 17, many others. And it's not something you'd be able to hear on an iPad or iPhone as mentioned in some articles.

When I first heard people talking about this I was positive there could not possibly be a difference (playing a FLAC vs WAV in real time). But when I actually tried it and really listened, it was quite apparent.

Although it's a lot less convenient, I use JPlay Mini, which decompresses the FLAC first, and then loads the tunes into a system buffer. On this player playing FLAC sounds just like the WAV equivalent as heard in other players (or itself). JPlay can also be used as a player type in JRiver's fine Media Center 17 software.

If you're just casually listening as I am right now, the differences aren't bothersome, but they are definitely there. I store everything in FLAC and uncompress it either with JPlay or manually before serious audition.

--Bill

 

[Bulldogger]

When I first heard people talking about this I was positive there could not possibly be a difference (playing a FLAC vs WAV in real time). But when I actually tried it and really listened, it was quite apparent.

Although it's a lot less convenient, I use JPlay Mini, which decompresses the FLAC first, and then loads the tunes into a system buffer. On this player playing FLAC sounds just like the WAV equivalent as heard in other players (or itself). JPlay can also be used as a player type in JRiver's fine Media Center 17 software.

If you're just casually listening as I am right now, the differences aren't bothersome, but they are definitely there. I store everything in FLAC and uncompress it either with JPlay or manually before serious audition.

--Bill

Ahhhh! Thanks. I do use JRiver and JPlay. My friend system has quite a bit of resolution but he is not using JRiver. I will have to test in my own system, 20k speakers, etc.yadayada. No, not a casual listener. I think I will get him to try JRiver and see what happens. Good advice.

 

[Vincent Kars]

uncompress it either with JPlay or manually

Have a look at FlacWavLoader: http://www.mikegalusha.com/

 

[bblue]

+1.. no difference!

There's no debating that -- when FLAC is unpacked to WAV during a normal file operation. The sound issue only appears when you use a real-time unpacking/decoding plug-in to allow the player to play the FLAC file 'natively'.

I'm sure this has something to do with clocking/jitter when the real-time plug-in is in use. Some say it is a matter of CPU utilization for the real-time process. I'm not so sure, but it could be. To me it sounds like a problem with output timing from the plug-in somehow causing jitter. If that was really the case, I would expect any subsequent DA with clock recovery for jitter reduction would correct the issue. I use a Crane Song Avocet monitor controller, which -- I'm Told -- has an excellent recovery algorithm. Yet, the difference between real-time FLAC decode and WAV play is quite audible, regardless of internal or external clocking, or different drivers.

Bruce, I would expect your system to easily reveal this difference. You're using an Avocet controller aren't you? I thought I saw it in one of your pictures.

--Bill

 

[Bruce B]

Bruce, I would expect your system to easily reveal this difference. You're using an Avocet controller aren't you? I thought I saw it in one of your pictures.

--Bill

Yes, we use the Avocet and a Switchman 3. Of couse you can hear the difference between wav and FLAC with different software, but it's not because you're hearing the difference in the format.

 

[bblue]

Yes, we use the Avocet and a Switchman 3. Of couse you can hear the difference between wav and FLAC with different software, but it's not because you're hearing the difference in the format.

Well right, but what we've been talking about is the same software playing a wav format against a flac made from that wav -- in real time. In every program I've tried including Media Center 17 the flac has less detail and air. The only exception to this is the Jplay mini player which does not use the real-time decoder engine for flac playback, but rather completely decodes it before playing. Jplay running under Media Center 17 also works well.

Different subject, have you found it necessary to make any electrical (audio related) changes to the Avocet?

--Bill