Why do WAV and FLAC files Sound Different

[davide256]

The decoder running slow might result in anything from a single sample to most of the buffer being repeated. One or two bad samples may or may not be audible (a large error in a single sample causes a clear tick), but any glitch of more than a millisecond is immediately obvious. It is extremely unlikely that the decoder would be just barely keeping pace and falling behind by a few samples now and then. Realistically, it is either fast enough and stays well ahead of the playback (even though the exact distance between decode and playback position will vary back and forth a bit), or it is too slow, in which case you'll hear the music progress but with many repeated fragments (stuttering).

 

For the purpose of FLAC decoding, modern computers are very, very fast. My old i7 from 2009 decodes a 28-minute 96/24 stereo FLAC file in about 6 seconds. That's about 280 times faster than it needs to be. Now even with a fast computer, without a true real-time OS, it is possible for the decoder to get arbitrarily held up, leading to a buffer underrun. If this does happen, it is again unlikely that the resulting glitch with be limited to a few samples, especially repeatedly. A persistent problem of this nature results in mostly correct playback with occasional glitches. In severe cases, it might glitch several times per second but rarely more. This is because when the decoder does get to run, it is fast enough to quickly fill the entire available buffer space. Should your system be thus afflicted, the cure usually consists of increasing the audio buffer size until the problem goes away. A total buffer of 100 ms is usually enough.

 

 

 

Neither of those effects have been shown to exist under controlled conditions.

hmm, I know my source which is my own testing and observation. What unbiased and current source are you basing your "controlled conditions" negation on?

[esldude]

I believe it hinges around the way flac files are compressed. With wav, each and every sample of each channel has equal weighing. The flac encoder may compress each frame of data to varying degrees depending on how compressible the data is. There is also a process of Interchannel Decorrelation, where one of 4 methods are used on a frame-by-frame basis to compress the L/R channel data.

 

Essentially flac is creating slight timing variations between frames (or L/R channels) that dont exist with wav format (even if it may only be less than 1us).

 

No, it is not doing what you describe. How are you supposing these 1 usec variations get into the output? And yes I actually have tested that to see if it is happening, and it isn't. Of course my source computer was an Asus netbook with only 2 gig of memory. It had no trouble playing FLAC just the same as it played WAV files.

[mansr]

I believe it hinges around the way flac files are compressed. With wav, each and every sample of each channel has equal weighing. The flac encoder may compress each frame of data to varying degrees depending on how compressible the data is. There is also a process of Interchannel Decorrelation, where one of 4 methods are used on a frame-by-frame basis to compress the L/R channel data.

 

Essentially flac is creating slight timing variations between frames (or L/R channels) that dont exist with wav format (even if it may only be less than 1us).

Bollocks. That's all I have to say.

[davide256]

No, it is not doing what you describe. How are you supposing these 1 usec variations get into the output? And yes I actually have tested that to see if it is happening, and it isn't. Of course my source computer was an Asus netbook with only 2 gig of memory. It had no trouble playing FLAC just the same as it played WAV files.

 

How did you test please? What level of resolution did you have to determine data sample variation in timing and what was your degree of error?

[dean70]

No, it is not doing what you describe. How are you supposing these 1 usec variations get into the output? And yes I actually have tested that to see if it is happening, and it isn't. Of course my source computer was an Asus netbook with only 2 gig of memory. It had no trouble playing FLAC just the same as it played WAV files.

 

I didn't say it would get into the output, but what about ground plane noise? If it were present, there would be a different profile for flac vs wav.

[esldude]

How did you test please? What level of resolution did you have to determine data sample variation in timing and what was your degree of error?

 

I ran a pair of high frequency tones one octave apart. Do that with the same setup otherwise using both FLAC and WAV files. Null them against each other. If the dominant difference is a timing shift the higher frequency tone will null out 6 db less than the lower tone. You can also calculate what level of timing shift would result in what level of null if you find that is the result. The amount of error is dependent upon the timing accuracy of the gear and noise levels. So what I did wouldn't be good enough for picosecond levels. It is however good enough to see the difference in 1 meter of interconnect vs 2 meters. Which is roughly 3 nanoseconds.

 

There are a list of other things wrong with your supposed explanation about the timing of the channels being different with FLAC. Quite simply what you are imagining is not how it works as the digital info is transmitted out whether from FLAC or WAV files.

[jabbr]

[emoji23][emoji23] Since whatever my files start out as get sliced and diced and converted to DSD256-512 before being sent out across the network ultimately to the DAC, I am hardly concerned with whether they start out as FLAC vs AIFF vs WAV and since my storage devices are compressed it also doesn't really matter -- but I can't hear a difference.

If you can hear a difference you really should reconfigure your system [emoji6] seriously

[sandyk]

[emoji23][emoji23] Since whatever my files start out as get sliced and diced and converted to DSD256-512 before being sent out across the network ultimately to the DAC, I am hardly concerned with whether they start out as FLAC vs AIFF vs WAV and since my storage devices are compressed it also doesn't really matter -- but I can't hear a difference.

If you can hear a difference you really should reconfigure your system [emoji6] seriously

Hi Jon

Not everybody wants, or needs, a network to pipe music right through the house, including the dunny ! (grin)

Neither does everybody feel the need for a DSD DAC , particularly due to the dearth of suitable new material that hasn't had the life squeezed out of it by heavy handed mastering.

Kind Regards

Alex

[audiventory]

I believe it hinges around the way flac files are compressed. With wav, each and every sample of each channel has equal weighing. The flac encoder may compress each frame of data to varying degrees depending on how compressible the data is. There is also a process of Interchannel Decorrelation, where one of 4 methods are used on a frame-by-frame basis to compress the L/R channel data.

 

Essentially flac is creating slight timing variations between frames (or L/R channels) that dont exist with wav format (even if it may only be less than 1us).

 

Dean70,

 

Does you mean: different size of frames impact to smoothness of playback?

 

But there buffer between source file (for both FLAC and WAV) and DAC.

 

If buffer is not emptied then stream audio date to DAC is identical for any source file type (WAV, FLAC, MP3, DSD, ...).

[dean70]

Dean70,

 

Does you mean: different size of frames impact to smoothness of playback?

 

But there buffer between source file (for both FLAC and WAV) and DAC.

 

If buffer is not emptied then stream audio date to DAC is identical for any source file type (WAV, FLAC, MP3, DSD, ...).

 

No, but the variable processing involved vs steady state. Yes there is obviously a buffer for the uncompressed data, but ground plane noise? If present, would it be be a different profile for flac vs wav?

[audiventory]

No, but the variable processing involved vs steady state. Yes there is obviously a buffer for the uncompressed data, but ground plane noise? If present, would it be be a different profile for flac vs wav?

 

There 3 buffers even:

- in DAC,

- in operation system (in sound device's driver),

- in software may be released third buffer.

 

Buffer contains uncompressed data.

 

Noise same for both: FLAC and WAV.

 

Computing FLAC decomression comparable with calculations of physical reading file from HDD, with GUI processing, etc.

 

Inline DSD conversion and resampling (on level that applied to hi-end) consume significantly more calculation resources.

 

So added FLAC processing drowns in big calculation heap by read HDD, drawing GUI, other tasks, etc.

 

If we consider ground noise as product of additional calculation load, then there should not be big difference between with FLAC decompressing or without it.

 

It is reason, why need analyse spectrum (better time-spectrum) diagram altering degree with FLAC decompression.

 

 

 

If present, would it be be a different profile for flac vs wav?

 

I don’t know that inside code of a audio players.

 

But, logically, a programmer should choose easiest way:

 

WAV: read from HDD -> do something -> put in buffer of operation system

 

FLAC: read from HDD -> decode -> do something -> put in buffer of operation system

 

May be some players have other scheme. But it is exception rather, I suppose.

[4est]

Come on now Alex, you know that he/we upsample other non DSD material, and he is just proving the point that his files get processed quite a bit before they are sent out.

 

As to the OPs question, I noticed distinct and repeatable differences when I played Redbook native files using a 2009 Mac Mini. Since using a 2012 Mac Mini and a Windows 7/8/10 I cannot say that I have noticed the difference, but I am doing as jabbr. First pcm with XXHighend, and now dsd with HQPlayer.

 

Hi Jon

Not everybody wants, or needs, a network to pipe music right through the house, including the dunny ! (grin)

Neither does everybody feel the need for a DSD DAC , particularly due to the dearth of suitable new material that hasn't had the life squeezed out of it by heavy handed mastering.

Kind Regards

Alex

[sandyk]

Forrest

Just out of curiosity, did you try those files that Dennis posted where 3 of us thought we heard differences with an AQ cable when connected as recommended, AND when reversed ?

 

Alex

[4est]

I have zero interest in playing a game or proving a point where I know the creator has an agenda one way or another. To make matters worse, I tore down the system in my signature and need to set up my Magnepan/JC1s.

Forrest

Just out of curiosity, did you try those files that Dennis posted where 3 of us thought we heard differences with an AQ cable when connected as recommended, AND when reversed ?

 

Alex

[sandyk]

I have zero interest in playing a game or proving a point where I know the creator has an agenda one way or another.

 

I only tried it out of curiosity, and I certainly didn't expect to hear a difference with a directional cable.

I doubt that Dennis expected anybody to hear a difference either, especially Plissken given his vendetta with AQ .

 

Alex

[4est]

I can't believe I am saying this, but you have much more patience for that ilk than I.

I only tried it out of curiosity, and I certainly didn't expect to hear a difference with a directional cable.

I doubt that Dennis expected anybody to hear a difference either, especially Plissken given his vendetta with AQ .

 

Alex

[davide256]

I ran a pair of high frequency tones one octave apart. Do that with the same setup otherwise using both FLAC and WAV files. Null them against each other. If the dominant difference is a timing shift the higher frequency tone will null out 6 db less than the lower tone. You can also calculate what level of timing shift would result in what level of null if you find that is the result. The amount of error is dependent upon the timing accuracy of the gear and noise levels. So what I did wouldn't be good enough for picosecond levels. It is however good enough to see the difference in 1 meter of interconnect vs 2 meters. Which is roughly 3 nanoseconds.

 

There are a list of other things wrong with your supposed explanation about the timing of the channels being different with FLAC. Quite simply what you are imagining is not how it works as the digital info is transmitted out whether from FLAC or WAV files.

 

Got it, you are basically using the two signals to cancel out and if a significant difference exists one or both signals were altered. What was the processing power of the machine used? Fully agree that on a powerful computer there will be no difference. The question I have is whether on a "lower spec" machine one will see a difference, say something like a raspberry pi or the like.

[jabbr]

Hi Jon

Not everybody wants, or needs, a network to pipe music right through the house, including the dunny ! (grin)

Neither does everybody feel the need for a DSD DAC , particularly due to the dearth of suitable new material that hasn't had the life squeezed out of it by heavy handed mastering.

Kind Regards

Alex

 

Alex,

 

I have zero doubt that you or someone else could construct a system that sounds different for FLAC and WAV. I prefer the system that sounds the same. I go as far as to say that it is a design defect in the system if these formats sound different.

 

Jon

[r_w]

I suppose I can agree that a low spec or poorly configured computer/player/software might not do the concurrent FLAC unpacking and playing very well and possibly noisily.

 

Other than that, on say an Aurender or any well configured, purposeful player there shouldn't be a noticeable difference. I haven't even bothered to try, FLAC sounds great to me.

 

Actually... thiking about it... does anyone know how Aurender do it? As in, when copying the whole file to internal SSD/playback buffer do they unpack FLAC at that time/right after copying, or is it done on the fly/while streaming?

[sandyk]

Alex,

 

I have zero doubt that you or someone else could construct a system that sounds different for FLAC and WAV. I prefer the system that sounds the same. I go as far as to say that it is a design defect in the system if these formats sound different.

Jon

 

Jon

Please check your PMs

G'night

Alex

[mansr]

I suppose I can agree that a low spec or poorly configured computer/player/software might not do the concurrent FLAC unpacking and playing very well and possibly noisily.

 

Decoding FLAC is ridiculously cheap. Any additional noise it creates may well be compensated by reduced disk reads. I haven't done any measurements, and I doubt I'd be able to detect the difference, but the processing involved is of such magnitude that this seems reasonable. On a computer system running any kind of OS, there are more intensive things going on in the background all the time.

[audiventory]

Decoding FLAC is ridiculously cheap. Any additional noise it creates may well be compensated by reduced disk reads.

 

Agree. And disk-read reducing need check for noise effectivity too.

 

Better way for reducing the noise begin from ferrite rings/filters, power supply filters (as low as high voltage), grounding, shielding, balanced and/or fiber connections and other same things. It is more effective than software decisions.

 

Of course, there need pro measurement equipment (hardware spectrum analyzer) because make it properly on hearing almost impossibly.

 

After hardware noise minimizing we may consider software issues.

[davide256]

In my experience the only time I could hear a difference between FLAC and WAV was using an Oppo 103... the processor and memory on it doesn't hold a candle to a PC. Once I made the transition to more powerful renderers I moved from an all wav library to all flac as the tag functionality is poor/missing for wav

[ron spencer]

I don't see any foobar ABX tests posted here. Does anyone have any they could share showing a difference?

[ron spencer]

I suppose I can agree that a low spec or poorly configured computer/player/software might not do the concurrent FLAC unpacking and playing very well and possibly noisily.

 

If you believe this then you could never have opened a zip file...again is there peer reviewed academic research on this?