Jump to content
IGNORED

FLAC to WAV conversion file size?


Recommended Posts

Using (four) different softwares (on the Mac) to decode/convert FLAC into WAV files (no change of bit and sample rate!), I noticed differences in file size (as well as subtle but consistent sonic differences)? Checksum of files isn't identical either. I'd figured conversion softwares are all doing the same thing when it comes to lossless conversions, and am puzzled to find they're not. Anyone have an explanation? Thanks in advance!

 

Greetings from Switzerland, David.

Link to comment
to decode/convert FLAC into WAV files

 

For clarity, are you going WAV to FLAC back to WAV and then comparing the before and after WAV files? I've never actually tried this, looks like an easy experiment.

Analog: Koetsu Rosewood > VPI Aries 3 w/SDS > EAR 834P > EAR 834L: Audiodesk cleaner

Digital Fun: DAS > CAPS v3 w/LPS (JRMC) SOtM USB > Lynx Hilo > EAR 834L

Digital Serious: DAS > CAPS v3 w/LPS (HQPlayer) Ethernet > SMS-100 NAA > Lampi DSD L4 G5 > EAR 834L

Digital Disc: Oppo BDP 95 > EAR 834L

Output: EAR 834L > Xilica XP4080 DSP > Odessey Stratos Mono Extreme > Legacy Aeris

Phones: EAR 834L > Little Dot Mk ii > Senheiser HD 800

Link to comment
Using (four) different softwares (on the Mac) to decode/convert FLAC into WAV files (no change of bit and sample rate!), I noticed differences in file size (as well as subtle but consistent sonic differences)? Checksum of files isn't identical either. I'd figured conversion softwares are all doing the same thing when it comes to lossless conversions, and am puzzled to find they're not. Anyone have an explanation? Thanks in advance!

 

Greetings from Switzerland, David.

 

I'm a windows guy I know nothing about the mac but it should be the same in this regard. This is a case of bits are bits and to transcode flac to wave at the same sample rate I would expect the same results from any valid program. I personally use xrecode II and do this all the time. Most common results are double the file size upon completion. I can only guess if you are seeing a difference between software that the difference is small and the programs are including some additional data into the files that probably has no effect on the actual transcoding. For example some type of log reports specific to the software. Perhaps artwork or included pdf files in the files are handled differently by different programs. If the actual music files are not of the same size then there is a very real problem. I have not seen this before unless the sample rate was changed.

 

I actually do this every day as I transcode flac to wave and place the file in my ram disk for playback in HQPlayer. Just my personal choice as it certainly sounds better than playing from hard disk. I would be interested to know how you resolve this.

Link to comment
Using (four) different softwares (on the Mac) to decode/convert FLAC into WAV files (no change of bit and sample rate!), I noticed differences in file size (as well as subtle but consistent sonic differences)? Checksum of files isn't identical either. I'd figured conversion softwares are all doing the same thing when it comes to lossless conversions, and am puzzled to find they're not. Anyone have an explanation? Thanks in advance!

 

Greetings from Switzerland, David.

 

David,

FLAC conversion has adjustable (in codec settings) level of compression. It impact to CPU resource consuming only. Sound quality is equal for any level of compression.

AuI ConverteR 48x44 - HD audio converter/optimizer for DAC of high resolution files

ISO, DSF, DFF (1-bit/D64/128/256/512/1024), wav, flac, aiff, alac,  safe CD ripper to PCM/DSF,

Seamless Album Conversion, AIFF, WAV, FLAC, DSF metadata editor, Mac & Windows
Offline conversion save energy and nature

Link to comment
David,

FLAC conversion has adjustable (in codec settings) level of compression. It impact to CPU resource consuming only. Sound quality is equal for any level of compression.

 

That was exactly the point but if I understand him correctly he is saying the resulting wave file is not the same size using different programs. Since wave is uncompressed they should all be the same regardless of the software used for the conversion. Of course you are correct the original flac files could have been set at any compression and they might be different sizes but the resulting wave files at the same sample rate as the original should be identical.

Link to comment

Keep in mind that different programs might write different "meta data" (aka. resource fork) to a file. The data fork will be identical.

The resource fork contain info such as date and time, icons and so on. So overall file size can differ.

Roon client on iPad/MacBookPro

Roon Server & HQPlayer on Mac Mini 2.0 GHz i7 with JS-2

LPS-1 & ultraRendu → Lampizator Atlantic → Bent Audio TAP-X → Atma-sphere M60 → Zero autoformers → Harbeth Compact 7 ES-3

Link to comment
That was exactly the point but if I understand him correctly he is saying the resulting wave file is not the same size using different programs. Since wave is uncompressed they should all be the same regardless of the software used for the conversion. Of course you are correct the original flac files could have been set at any compression and they might be different sizes but the resulting wave files at the same sample rate as the original should be identical.

 

Sorry, I understand. I supppose, different digital signal processing (filtration or other) applied (truncatinng of internal buffers). May be different versions of FLAC's programming libraries used (also different buffer truncating or aligment to internal block size).

 

Anyway possibly check sameness of result WAV via using FLAC with special developed test signal (having obvious reference points).

 

With such file possible check full circuit: WAV - encode - FLAC - decode - WAV.

 

Periodically (when anybody say about difference FLAC and WAV) I want learn sameness of FLAC and WAV. But while have no time.

 

I suppose possible some time shifting between different coders and encoders.

 

As example, in my audio converter possible 2 mode for unchanged sample rate:

1. With DSP (what made "half of DAC's work")

2. Without DSP - input as is.

 

In case 2 (if used FLAC compression) input WAV (passed thru FLAC) must be same to output WAV. I use standard FLAC library and while don't checked case for WAV-FLAC-WAV.

 

Other manufacturers, I think, also use the standard FLAC library and difference may be in environment of the library.

 

Also possible difference of unpacking for different FLAC library's versions.

 

But (theoretically, as declared by FLAC developer) for any library samples input and output WAV must be same.

 

P.S. Also need concider possibility of rounding errors (depend on internal software's processing) when writed output WAV. It don't influence to length but affect to sound quality when says about difference quality WAV and FLAC.

AuI ConverteR 48x44 - HD audio converter/optimizer for DAC of high resolution files

ISO, DSF, DFF (1-bit/D64/128/256/512/1024), wav, flac, aiff, alac,  safe CD ripper to PCM/DSF,

Seamless Album Conversion, AIFF, WAV, FLAC, DSF metadata editor, Mac & Windows
Offline conversion save energy and nature

Link to comment

Thanks everyone for your replies!

 

What I did is convert FLAC files (from high-resolution downloads) into WAV using Max, XLD, eXACT and Korg AudioGate (the reason is that my favorite media player won't play back FLAC in its current version). I'd expected to get an identical WAV file from each one of these conversions, but to my surprise, I didn't.

 

I'll try the recommended back and forth conversion to see if each of the softwares does this consistently.

 

Greetings from Switzerland, David.

Link to comment
Thanks everyone for your replies!

 

What I did is convert FLAC files (from high-resolution downloads) into WAV using Max, XLD, eXACT and Korg AudioGate (the reason is that my favorite media player won't play back FLAC in its current version). I'd expected to get an identical WAV file from each one of these conversions, but to my surprise, I didn't.

 

I'll try the recommended back and forth conversion to see if each of the softwares does this consistently.

 

Greetings from Switzerland, David.

 

Actually, I did understand it as you describe. I am curious as to what sort of difference you are seeing in the size of the resulting wave files? Perhaps it is a very small difference? That might be the result of my comments above? If it is a large difference something is really off. I would like to suggest you try xrecode II which you can get as a trial. I use it every day for the same purpose and it is very reliable.

Link to comment
Actually, I did understand it as you describe. I am curious as to what sort of difference you are seeing in the size of the resulting wave files? Perhaps it is a very small difference? That might be the result of my comments above? If it is a large difference something is really off. I would like to suggest you try xrecode II which you can get as a trial. I use it every day for the same purpose and it is very reliable.

 

Thanks! Appears to be a Windows-only software, however?

Link to comment

Did the following:

 

Using Track 1 from my HDtracks 24/96 FLAC download of Keith Jarrett's Köln Concert, extracted WAVs with following results in file size:

 

Max: 907.6 MB

XLD: 907.4 MB

Korg AudioGate: 907.4 MB

xACT: 907.4 MB

 

Converting all of these back and forth, first to FLAC compression level 5 and back again to WAV, only Max will preserve the integrity of the 907.6 MB file it originally created - all the others will make a 907.4 MB WAV file out of either their own 907.4 MB or Max's 907.6 MB file.

 

I tried other FLACs (the first three tracks from a Le Nozze di Figaro download from HighResAudio) - the differences in file size were even bigger, both in relative as well as absolute terms.

 

I also tried on of my own WAVs. All four software convert that back and forth with consistent results.

 

What this means is that Max (the obsolete, since no longer updated since 2009, but handy converter by Stephen Booth that in contrast to e.g. XLD would allow one to modify tags etc. prior to conversion) "adds" something of differing size while decoding FLAC to WAV, which the other converters in turn strip from the resulting file during back and forth conversion.

 

In short, only Max will back and forth convert its own conversion, or that from one of the other three softwares, and preserve whatever the integrity of a WAV regardless of source. It does, however, something differently decoding FLAC files.

 

Now, I would never have noticed unless the files sounded differently. Yes, a WAV decoded from FLAC with the old Max software has sonic character that's different from all the other converters. That's how I noticed in the first place - all my old WAV conversions from FLAC downloads on my storage SSD were done with Max. Then a couple of days ago, I used another converter and noticed the difference in sonic character.

 

Jarrett's Köln Concert was the LP that, as a teenager, sent me on this audiophile journey or never-ending quest, an album I've owned in all formats, and taken everywhere to listen to it, a recording that I know by heart.

 

Now, I won't say that one conversion sounds per se "better" than the other, but for sure one sounds more similar to the original LP (the same is true of Kleiber's Nozze).

 

And I find that highly irritating. I'd hoped digital would be more straightforward. Less tweaking, no nonsense that makes me scratch my head. Literally, I was used to analogue sounding different however irrelevant a modification might seem, but digital continues to surprise me in this respect.

 

Back to topic: anyone have an idea what it is that Max does differently during FLAC to WAC decoding?

 

(That it is the only software to handle its own conversions properly, preserving whatever it does differently as well as preserving the integrity of WAVs regardless of source, is less surprising - and ultimately perhaps irrelevant. Except that what the other converters strip appears to be what Max adds, and what appears to affect a file's sound.)

 

Greetings from Switzerland, David.

Link to comment
Using (four) different softwares (on the Mac) to decode/convert FLAC into WAV files (no change of bit and sample rate!), I noticed differences in file size (as well as subtle but consistent sonic differences)? Checksum of files isn't identical either. I'd figured conversion softwares are all doing the same thing when it comes to lossless conversions, and am puzzled to find they're not. Anyone have an explanation? Thanks in advance!

 

Greetings from Switzerland, David.

 

There’s a good reason for different converters producing WAV files with different sizes when a source FLAC contains tags (metadata and art): the WAVs can contain identical audio samples but contain different tags.

For example:

xACT does not write tags into WAV.

Max writes id3 tags into WAV.

XLD has options to either not write tags, write tags in INFO format, write tags in id3 format, or write tags in both formats into WAV.

Furthermore, programs process cover art differently when embedding it into a file. XLD will store the provided image directly into an audio file, whereas Max will re-encode the image in PNG format before storing into the audio file, which can result in very significant file size differences.

 

You can use an audio editor such as Audacity to arithmetically determine whether files contain absolutely identical audio samples. WAVs containing identical audio samples sound identical to me, regardless of the size of the files or whatever tags may be in them.

 

Link to comment
There’s a good reason for different converters producing WAV files with different sizes when a source FLAC contains tags (metadata and art): the WAVs can contain identical audio samples but contain different tags.

For example:

xACT does not write tags into WAV.

Max writes id3 tags into WAV.

XLD has options to either not write tags, write tags in INFO format, write tags in id3 format, or write tags in both formats into WAV.

Furthermore, programs process cover art differently when embedding it into a file. XLD will store the provided image directly into an audio file, whereas Max will re-encode the image in PNG format before storing into the audio file, which can result in very significant file size differences.

 

You can use an audio editor such as Audacity to arithmetically determine whether files contain absolutely identical audio samples. WAVs containing identical audio samples sound identical to me, regardless of the size of the files or whatever tags may be in them.

 

Thank you! I'll look into that!

 

I'd never thought about this until a couple of days ago, I mean, how could they not sound the same - however if they did, I'd never have noticed. It was hearing a new to me sonic character in Keith Jarrett's Köln Concert that made me discover the files are differently sized etc.

 

Used a different converter (xACT, because one doesn't even have to check or change settings to decode FLAC) for the purpose for the first time since I have a CA setup, put on the music turning my attention to something else, and looked up hearing the first few notes: "What the…?!"

 

First I thought I'd done something wrong, but I've double- and triple-checked since, and tried all of the above. Now I feel like I need to decide which conversion software to use when until a couple of days ago I'd merrily assumed there couldn't be a difference.

 

Greetings from Switzerland, David.

Link to comment
[...] Max (the obsolete, since no longer updated since 2009, but handy converter by Stephen Booth that in contrast to e.g. XLD would allow one to modify tags etc. prior to conversion) [...]

 

XLD does allow editing before conversion.

 

"Open folder as a disc..." command (or Cmd-drag folder onto XLD icon): processing doesn't start until you click the Transcode button, giving you opportunity to edit metadata first. Also, the editor there allows simultaneous editing of multiple tracks when you've made a multi-selection in the tracklist before invoking "Edit metadata..." command (Cmd-i).

 

"Open..." command (or drag file/folder onto XLD icon): processing doesn't start until you click the OK button, giving oppportunity to edit metadata first, if you've enabled "Edit tags before convert" in Metadata tab of XLD's Preferences.

Link to comment
XLD does allow editing before conversion.

 

"Open folder as a disc..." command (or Cmd-drag folder onto XLD icon): processing doesn't start until you click the Transcode button, giving you opportunity to edit metadata first. Also, the editor there allows simultaneous editing of multiple tracks when you've made a multi-selection in the tracklist before invoking "Edit metadata..." command (Cmd-i).

 

"Open..." command (or drag file/folder onto XLD icon): processing doesn't start until you click the OK button, giving oppportunity to edit metadata first, if you've enabled "Edit tags before convert" in Metadata tab of XLD's Preferences.

 

Thanks again for providing the most specific and, to a computer dummy like me, helpful answers!

 

XLD is my favorite CD ripper on the Mac thanks to AccurateRip (= I don't even know if it's still the only one to do this, but it was when I first got into CA, and I'm not really keeping up to date in these matters, as you'll have noticed, generally using tools to get the job done and go on to the to me essential part, which is listen to music).

 

Seriously, very much appreciated! I hate being confused…

 

Greetings from Switzerland, David.

 

P.S.

I hope you're not going to frown upon a potentially silly question, but the way you're putting it further above, it would seem that xACT does the most "puristic" FLAC decoding (even if, as I understand you, this shouldn't result in an audible difference)?

Link to comment
Thanks everyone for your replies!

 

What I did is convert FLAC files (from high-resolution downloads) into WAV using Max, XLD, eXACT and Korg AudioGate (the reason is that my favorite media player won't play back FLAC in its current version). I'd expected to get an identical WAV file from each one of these conversions, but to my surprise, I didn't.

 

Yes this is very clear. My understanding of WAV and AIFF is that they are "containers" and as such maybe there can be ways to produce different file outputs that really represent the same bitstream. It puzzles me that you say you hear consistent sound differences.

 

So I will add a question to yours: How does one compare that two files really have the same bitstream?

mini+Roon > RAAT > dCS Rossini DAC + dCS Rossini Clock 

SME 20/3 + SME V 9” + Dynavector XV-1s > vdH The Grail

Audio Note Kondo Ongaku > Avantgarde Duo Mezzo

system pics

Link to comment
Thanks again for providing the most specific and, to a computer dummy like me, helpful answers!

 

XLD is my favorite CD ripper on the Mac thanks to AccurateRip (= I don't even know if it's still the only one to do this, but it was when I first got into CA, and I'm not really keeping up to date in these matters, as you'll have noticed, generally using tools to get the job done and go on to the to me essential part, which is listen to music).

 

Seriously, very much appreciated! I hate being confused…

 

Greetings from Switzerland, David.

 

P.S.

I hope you're not going to frown upon a potentially silly question, but the way you're putting it further above, it would seem that xACT does the most "puristic" FLAC decoding (even if, as I understand you, this shouldn't result in an audible difference)?

 

 

You're welcome! :)

 

XLD and xACT can produce identical "pure" WAVs with no embedded tags.

 

XLD: disable writing of tags in the options for WAV output.

xACT: no action required.

Link to comment

I hope you're not going to frown upon a potentially silly question, but the way you're putting it further above, it would seem that xACT does the most "puristic" FLAC decoding (even if, as I understand you, this shouldn't result in an audible difference)?

 

Most "puristic" FLAC decoding impossible due FLAC definition as lossless.

 

All FLAC decoders for unchanged bit-depth and sample rate must have fully identical sounding (bit-perfect, may be with some time shifting or different size - it don't impact to sound).

 

FLAC's matatags is not same to id3 matatag format used for WAV.

 

Thus tag converting may be very different.

 

id3 tags may have different length of metatag simbol string due different simbol's encoding (ISO-8859-1, UTF8, UTF16,...). Each simbol UTF8 have 1 byte length, UTF16 - 2 byte.

AuI ConverteR 48x44 - HD audio converter/optimizer for DAC of high resolution files

ISO, DSF, DFF (1-bit/D64/128/256/512/1024), wav, flac, aiff, alac,  safe CD ripper to PCM/DSF,

Seamless Album Conversion, AIFF, WAV, FLAC, DSF metadata editor, Mac & Windows
Offline conversion save energy and nature

Link to comment
So I will add a question to yours: How does one compare that two files really have the same bitstream?

Audacity can be used to check whether two files decode to identical audio samples. Do a null test, i.e. subtract one track from another to see if the result is absolute silence:

 

Open first file in Audacity.

Drag second file from Finder into same Audacity window.

Select one track then do Effect > Invert.

Select no track, or both tracks, then Press Shift+Cmd+M to mix to a third track.

The default display waveform is a linear vertical scale which won't reveal very low levels, so change the mix track's display to "Waveform (dB)" to see whether it is truly silent, i.e. flat at -infinity dB.

A flat result indicates that the files decode to identical audio samples, as in the attached image.

null.png

Link to comment

 

So I will add a question to yours: How does one compare that two files really have the same bitstream?

 

In binary view (as example it can Total Commander - Compare files by content).

 

1. In WAV need find word "data" (4 bytes), pass 4 bytes, further will audio info.

 

2. In AIFF or AIF need find word "SSND" (4 bytes), pass 4 bytes, further will audio info.

 

When comparing possible need consider shifting audio info for compared files. For this case need find reference points (byte sequences).

AuI ConverteR 48x44 - HD audio converter/optimizer for DAC of high resolution files

ISO, DSF, DFF (1-bit/D64/128/256/512/1024), wav, flac, aiff, alac,  safe CD ripper to PCM/DSF,

Seamless Album Conversion, AIFF, WAV, FLAC, DSF metadata editor, Mac & Windows
Offline conversion save energy and nature

Link to comment

Audacity can be used to check whether two files decode to identical audio samples. Do a null test, i.e. subtract one track from another to see if the result is absolute silence:

 

Open first file in Audacity.

Drag second file from Finder into same Audacity window.

Select one track then do Effect > Invert.

Select no track, or both tracks, then Press Shift+Cmd+M to mix to a third track.

The default display waveform is a linear vertical scale which won't reveal very low levels, so change the mix track's display to "Waveform (dB)" to see whether it is truly silent, i.e. flat at -infinity dB.

A flat result indicates that the files decode to identical audio samples, as in the attached image.

[ATTACH=CONFIG]16286[/ATTACH]

 

Thanks again!

 

In lay terms, the assumption appears to be that as long as the DAC "sees" the same bits, the container format and unpacking process for the raw audio data should (can?) have no impact on the sound - I'm wondering if one day we'll learn that the assumption itself is wrong (after all, it wouldn't be the first, nor the last)?

 

In the meantime, your explanation makes perfect sense, appreciate it!

 

Greetings from Switzerland, David.

Link to comment
In binary view (as example it can Total Commander - Compare files by content).

 

1. In WAV need find word "data" (4 bytes), pass 4 bytes, further will audio info.

 

2. In AIFF or AIF need find word "SSND" (4 bytes), pass 4 bytes, further will audio info.

 

When comparing possible need consider shifting audio info for compared files. For this case need find reference points (byte sequences).

 

Yuri, thanks for chiming in! Sorry to have to admit your contributions are beyond my scope - my bad! :(

 

Needless to say, your input is much appreciated!

 

Greetings from Switzerland, David.

Link to comment

I'm here to report that I have, thanks to the input I've gotten in this thread and the help of an audiophile electronics engineer friend, figured it out. The differences I hear between files has nothing to do with the conversion softwares I tried and mentioned further above, but tags/metadata.

 

The following will be of no relevance to those who don't hear a difference between PCM file formats, or another PCM file format (e.g. FLAC and AIFF) and WAV. However, if you, like me, convert your downloads and rips to WAV because you do hear a difference, be sure to convert to WAVs that contain no tags/metadata.

 

This is also true of some WAV downloads. The XLD converter, for example, can be set to strip a WAV from tags by unticking/unchecking the "write tags" box (thereby disabling the "INFO chunk"/"id3 chunk"/"Both" options, which then appear in pale print). Resulting files, provided they contained any tags/metadata in the first place (= even if seemingly empty), may be fractionally smaller in size as a result.

 

Needless to say, I encourage anyone interested in why some crazy audiophiles (like myself) claim they can hear a difference to compare their preferred format to WAV that is free from tags/metadata. In the same vein, I recommend to all those who love pics, metadata, the convenience factor (whatnot), to refrain from comparative listening, in fact, feel free to think of and/or call me (and anyone else who prefers WAV, such as the good people at the Sound Liaison label) delusional.

 

As to technical explanations: don't ask me. I perfectly realize the messenger is the first to be beheaded, but it may be worth remembering whom (and thus what) the mythical messenger's was warning of: Lucullus! ;)

 

Greetings from Switzerland, David.

Link to comment
I'm here to report that I have, thanks to the input I've gotten in this thread and the help of an audiophile electronics engineer friend, figured it out. The differences I hear between files has nothing to do with the conversion softwares I tried and mentioned further above, but tags/metadata.

 

The following will be of no relevance to those who don't hear a difference between PCM file formats, or another PCM file format (e.g. FLAC and AIFF) and WAV. However, if you, like me, convert your downloads and rips to WAV because you do hear a difference, be sure to convert to WAVs that contain no tags/metadata.

 

This is also true of some WAV downloads. The XLD converter, for example, can be set to strip a WAV from tags by unticking/unchecking the "write tags" box (thereby disabling the "INFO chunk"/"id3 chunk"/"Both" options, which then appear in pale print). Resulting files, provided they contained any tags/metadata in the first place (= even if seemingly empty), may be fractionally smaller in size as a result.

 

Needless to say, I encourage anyone interested in why some crazy audiophiles (like myself) claim they can hear a difference to compare their preferred format to WAV that is free from tags/metadata. In the same vein, I recommend to all those who love pics, metadata, the convenience factor (whatnot), to refrain from comparative listening, in fact, feel free to think of and/or call me (and anyone else who prefers WAV, such as the good people at the Sound Liaison label) delusional.

 

As to technical explanations: don't ask me. I perfectly realize the messenger is the first to be beheaded, but it may be worth remembering whom (and thus what) the mythical messenger's was warning of: Lucullus! ;)

 

Greetings from Switzerland, David.

 

I actually agree with everything you said and that is exactly how I listen to my music. I use Xrecode II to accomplish all this easily. I just wanted to jump in and say that you are not alone in your thinking. I also realize there are many who say this is madness, but I believe what I hear.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×
×
  • Create New...