Why do my CD's still sound better than flac files?

With your encouragement I did some quick research on the subject. 20 minutes with Google leads me to believe that the Windows Mixer may only resample data going to the soundcard (potentially needed due to clock rates on the soundcard, more on this in a bit). Since I'm not using the soundcard in this stream and only transmitting data through AirPlay to the Airport Extreme there is a reasonable chance that the Windows Mixer is not in the processing path and wouldn't impact it... but who really knows.

Now the interesting part in my opinion is why the Windows Mixer resamples the data stream in the first place. It is because most soundcards don't have a native 44.1 sampling/clock rate, so Windows has to send it data that it can natively use. What I quickly read is that most cards use either a 48 or 96 sampling rate. Now, what has me thinking is that the Airport Express supports 48Khz sample rates so it's quite possible that it does it's own internal resampling. Which if true means that it may not be bit identical. Which also means using my set up it could be impossible to get bit identical streams due to the resampling.

I share this here since is it possible that most devices that support multiple sampling frequencies do in fact do some kind of internal resampling to match clock rates? If they did, would that change the output in any appreciable way?

My 20 mintues with a search engine, only left me with more questions. But I'm interested enough to try the test again both with my Mac and iPhone through the Airport Express.

 

@Grego: You are on the right way. This is the reason why I have wrote on site 2 that when you get foobar and you change the output to the WASAPI driver then the complete Windows Mixer is bypassed!

But this would explain why on your case it sounds different!

 

No, it's because it has to mix sounds from different applications and it can only do that if all the audio streams have the same sampling rate. Hence Windows mixer.

WASAPI and ASIO use the sound card in "exclusive mode" and you'll notice you get no sounds from other applications when using it in your media player. As you say though, it's not relevant in your case because the PC is only being used as a streaming server - the sound card isn't being used.

 

This is not true. When his DAC can only 48KHz, then "anybody" have to resample the 44,1 KHz FLAC files!

 

What do you mean by "his DAC"? If you mean the Squeezebox then, yes, if it only works at 48 kHz then it'll need to be resampled at some point. Not sure when cos I don't know how Squeezebox works exactly but the PC's sound card is still not being used, so WASAPI/ASIO is not relevant.

 

I was talking about two distinct situations and may not have made it clear in my post. I don't believe the Windows Mixer is in the data path as we've all discussed (although I don't know what Windows actually does). I think Chris is referencing my Airport Express (Apple wireless device that can act as a router or music data streamer). It supports sample rates above 44.1Khz, so it's possible that the Airport Express is doing some resampling. If that is true, then there is nothing I can do in software to fix that. As DragonQ suggests (as I also did above) it's possible other devices could also do some resampling. If that is all true than it could explain why some people here a difference, since there is resampling in the digital domain which could change the final 1s and 0s.

 

@Grego's own words:

So, it is not bit perfect!!!!!!!!!!!!!! In his case I think the Airport Express resamples. You need a DAC who is able to support native 44,1KHz!

 

BINGO!

 

Yes, I slightly agree, but, we all have imaginations, perceptions etc, since when is having one a bad thing, or offensive?

Arent we all humans after all?


I thought the point was, that the OP said he wasnt even "sure" he heard a difference, so my first thought still would be that hes simply "fooling himself", and thats not offensive, but simply a normal human reaction in many circumstances involving a "possible" change in any stimuli.

Not to bring this up again, but humans instinctively "look" for changes or problems, even where there are none. Its simply part of being a human.

Given 2 identical things or situations, we instinctively try to find a difference. Thats not a bad thing, but it can let us believe things that are not so often.

The power of perception and advertising are both strong..!

 

Interesting thread. Anyone here seen the latest (Feb 2012) issue of TAS? It has part three of a four part series on digital audio and this issue's article deals precisely with this topic. Bottom line from the article is that FLAC files and uncompressed files not only sound different, but vary depending on the software used to create the FLAC file and to convert from FLAC to WAV. I'm not saying that because it's in TAS it's gotta be true, but they went through a great deal of trouble to look at this issue objectively and came to conclusions that are fundamentally in line with what Barry says he and numerous colleagues say they hear. I also think the guys at TAS who did the work were a bit surprised by the differences they heard with the different programs they used.

For those still interested in this issue, they might want to check out the article.

Cheers,
Mike

 

@MikeP: Have you the link for that article?

NEVER EVER!!!!!!!!!! One simple tryout: Convert to flac then retour to wav and compare the two wav files with md5 - you will get the same checksum!

 

It's in the print version of the magazine. I have no idea if it's available from an on-line source. But looking at the second part of your post, why would you want to read it? You seem to pretty much know what everyone else can/is hearing.

Hey, bits is bits, right? ;-)

Cheers,
Mike

 

Maybe they have some explanation about this voodoo FLAC to WAV conversion?

Aha, everyone is hearing that except myself.

 

Just audiophiles being audiophiles. "Buy our Monster cables! Buy our CD demagnetizer! Buy our green sharpies! Buy our 5 TB hard drives so you can store all your rips in uncompressed PCM!".
Complete bogus. FLAC and PCM produce the same md5 checksum, they are identical, period.

 

Actually, Chris, I have no idea what you, Barry, the editors at TAS or anyone else is hearing. I also can't explain why they are hearing differences if, in fact, they are. They are just reporting their observations.

I do biomedical research for a living. This has, of course, no bearing on what I can or can't hear when I listen to music. But if there is one thing I have learned in many years of doing science, it is that you do the experiment and make your observations. Then you try to figure out how to explain the results. Most people set up an experiment with an idea of what they expect to see when it's done. Often, they see something different from what they expected. This is usually because, no matter how much you know going in, you just don't know everything. So, you scratch your head, think up new assays you can run to figure out what is going on, and plow ahead. It's why I love to do science. It's also how we learn things.

Just because Barry, I or the editors of TAS can't explain to you what I/they hear doesn't mean it's not happening. We just don't have all the answers.

Cheers,
Mike

 

Hi Kevin,

I'm sorry but in my view, you are not merely skeptical.
We will have to agree to disagree as I could not disagree more. Not that human perception is fallible - there we agree. It is just a matter of degree where we are polar opposites.

And I see as insulting and quite offensive (and oh so tiresome), you don't find so. Perhaps if you were to receive same, your perceptions would differ. I just don't know if you'd trust them.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Barry, with all due respect, I think Kevin's remarks were actually pretty reasonable and even-tempered. I don't see them as insulting or offensive at all.

A different opinion is just that. We all hear what we hear; we don't necessarily hear what each other hears. Nothing wrong with that. People are different, ears are different, brains are different, systems are different.

I try to have tolerance (up to a point) towards people who think different than I do, and I try to couch my disagreements this way: "I don't dispute you hear something different, but I've never heard it myself, in my experience." Surely that's not too controversial or obnoxious.

Still, I think the relationship to a million-word novel is valid: if the sound files are different, how is it that I could copy and convert a text file a hundred times, and still wind up with every punctuation mark and space in the same place? I don't see audio, video, text, pictures, or anything else as being fundamentally different in terms of data integrity. It either works, or there's something wrong. I don't see subtle problems when I run into video projects that have digital problems -- they're generally very obvious, measurable, and not bit-accurate.

 

Hi chriss71,

I read the article and the authors are not speaking about comparing checksums. They are talking about sitting down and listening.

There is a difference between analyzing checksums (or performing null tests) and sitting down and actually listening. It is the difference between theory and practice. (Remember what Yogi Berra said on this subject.)

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

I would argue that if you're using the D/A converter in your CD player, vs. the D/A in your computer, it's not a fair comparison. You wind up comparing D/A converters -- and it's been proven that those can sound (and measure) different.

 

Hi Marc,

I'm all for differences of opinion. Sometimes a different opinion makes me re-evaluate my own.

The example you gave is how I would put it too: it was in terms of your experience, not the other person's.

There is a world of difference between me saying what I hear, as in "I hear it very differently" or "I hear it this way" and me saying what you hear. There can be no logical basis for the latter as none of us has any access to the other's perceptions. To offer an "evaluation" of another's perception, as in suggesting the person is imagining what they hear, is, to my mind, an insult. As well as revealing a lack of confidence on the part of one who would suggest such about another's perceptions.

It might not consciously be intended as an insult but that is nonetheless what it is.

The key difference between text and audio data is the text isn't time dependent where the audio is not only time dependent but demands timing precision. This is but one area where I suspect the differences may lie.
But before someone runs off with the word "jitter", I'll say I'm confident there are areas of digital playback issues that have not yet been identified, much less quantified.

The fact remains, I personally wouldn't touch flac with a ten foot speaker cable. ;-}

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

You may find it insulting, but it doesn't make the fact that people's perceptions are often colored any less true. The human brain is easily tricked, whether people want to admit it or not.

Not at all. It reveals a healthy skepticism.

If anything, it seems the lack of confidence may come from those who get insulted at the suggestion that they are imagining something.

 

Someday when we have 100 TB drives we can all expand our flacs to wav or aif and thisdiscussion will be over. Until then, my ears are just fine hearing flacs. Honestly I haven't compared my CDs to flacs mainly because flacs sound so good to me and the convenience of having a file based music library on a 1.5TB drive accessible via my iphone outweighs any sonic discrepancies that may or may not be evident in my modest listening setup.

 

Seems reasonable/logical, its a diff spec called redbook for a reason I suspect, sure they are bits but there must be more going on here than just storing ASCII data.
Just one bit pattern represents FREQ's at a moment in time, now thats plural cause their could be more than one, but we must also distinguish the relative volume between them for starters.

 

Sorry Mr. Bator,

Healthy skepticism is an expression of one's own perspective.
Unless you have access to another's experience, to comment on another's experience is b.s. when it isn't cowardice too.

BD
Www.soundkeeperrecordings.com
Www.barrydiamentaudio.com

 

That's cold comfort to many of those who burnt their rare and collectible CDs to FLAC files and then proceeded to sell off those CDs in the forums here and elsewhere ... and now it's gnawing away at the back or their minds they may have to repurchase them all over again.

The smell of remorse seems to be wafting through this thread.