Download: FLAC, WAV ?

Ah, but I did say easy rather than easier...

Innocent Bystander posted:

imperialline posted:

Innocent Bystander posted:

Well, I can't say I can tell any difference between flac and wav - may be my ears, may be the fact that my renderer is not Naim and is on a computer with plenty of processing power (and sounds better than did renderer stage of ND5XS).

IB, you may be surprised to know that the output of Audirvana to a DAC is always in WAV.

Indeed I was unaware. May be why I could't hear a difference when I compared...

I think you should be pleased with yourself and your clearly objective listening skills!

best

David

I can easily hear the difference between wav and flac. To me also wav sounds better. I feel transcoding must have some musical loss, although ive never tried this myself. If you have enough storage then why not just rip to wav in the first place?

Mr Happy posted:

I can easily hear the difference between wav and flac. To me also wav sounds better. I feel transcoding must have some musical loss, although ive never tried this myself. If you have enough storage then why not just rip to wav in the first place?

Loss of what? No bits are lost in transcoding. 

ChrisSU posted:

Mr Happy posted:

I can easily hear the difference between wav and flac. To me also wav sounds better. I feel transcoding must have some musical loss, although ive never tried this myself. If you have enough storage then why not just rip to wav in the first place?

Loss of what? No bits are lost in transcoding. 

Musical loss. Im not very techy but ive heard differences when some people quote its all ones and zeros so there can’t  possibly be a difference. 

Ive never tried transcoding so I cannot say whether there is a difference or not, but im sure, as has been the case before, it won’t be a simple case of bits lost or not. But hey, maybe it would be identical sounding, as I said ive never tried.

Mike-B posted:

Just to be sure its all understood - I have a hint of slight misunderstanding in some posts.

dBpoweramp is a software package (www purchase) loaded onto your laptop/PC/Mac,  that amongst many things like ripping CD's,   is used to convert music file formats e.g. FLAC to WAV (it will convert any lossless file type to any other - it can also convert lossless to lossy e.g. MP3)      The converted file is bit-for-bit & nothing is lost.     The file - in your case WAV - is stored as a WAV on your NAS & your streamer plays that as & when you choose.   dBpoweramp has no more involvement.  

The other suggestion is to store the FLAC or other lossless format on your NAS & set the NAS media software to "transcode" to WAV,  this feeds a WAV stream to the player each & every time its played, or as long as the NAS transcode setting is retained.   The stored file remains FLAC & dBpoweramp is not involved. 

I always use method one

Hi Mike-B: Thanks for that; my original query concerned the SQ between various file-formats. This was soon followed by my first (attempted -lol) download of a 24/96 file only available in a flac container (MQA was the album download ).

Since I just began 'streaming' (as in  a few weeks back). As an experiment, I decided to record about 100 of my CD's in various file formats including wav., flac (Levels 0,1,5,8), AAC and most recently AIFF -something my computer (and definitely my ears) do not like; there's a disturbing "phase-y-ness" to the AIFF file download (dB poweramp and Foobar 2000 for playback) that is truly bizarre and un hi-fi like. Very strange.    

 In any case, WAV files has proven (to me) the far superior file-format than, specifically FLAC  (the supposed lossless encoder). I've compared (blind) all the file-formats "on-the-fly"  about 20-25 songs from basic rock & roll (poorly recorded no less) to, well more sophisticated fare. In all but one case, WAv was the best sounding. By far.

Therefore, if I was to download a file (music), and I see that it is"packaged" in a FLAC container, I become suspicious since my listening tests clearly and absolutely preferred WAV.

That leads to my current question: Does the FLAC file/container hold the entire CD information, bit-for-bit as a WAV file ? I guess my question is there a difference between a flac-file and flac "container"?

Remember, I wish all of my music-files to be encoded in wav. (for the reasons noted above).

peter 

Hungryhalibut posted:

I use dbpoweramp to rip CDs and manage metadata. Everything is ripped to flac. All my downloads are in flac as well. I then use Asset to transcode to WAV when music is played. WAV sounds better to me, and I can’t tell the difference between a flac played as WAV, and a WAV played in its native format. To me this gives the best of both worlds - smaller files with easy to manage metadata, and best sound. 

Hi HH: You say: 

" ....and I can’t tell the difference between a flac played as WAV, and a WAV played in its native format."  (That would be great If there were no audible (to me) difference at all. It wold really simplify things. Yet, I remain stubborn -and always doubtful of claims -a hazard of aging lol !)  

" ....To me this gives the best of both worlds - smaller files with easy to manage metadata, and best sound."  (So true. I hope I can verify -the SQ component)

Thanks,

pj

allhifi posted:

Hi Mike-B: Thanks for that; my original query concerned the SQ between various file-formats. This was soon followed by my first (attempted -lol) download of a 24/96 file only available in a flac container (MQA was the album download ).

Since I just began 'streaming' (as in  a few weeks back). As an experiment, I decided to record about 100 of my CD's in various file formats including wav., flac (Levels 0,1,5,8), AAC and most recently AIFF -something my computer (and definitely my ears) do not like; there's a disturbing "phase-y-ness" to the AIFF file download (dB poweramp and Foobar 2000 for playback) that is truly bizarre and un hi-fi like. Very strange.    

 In any case, WAV files has proven (to me) the far superior file-format than, specifically FLAC  (the supposed lossless encoder). I've compared (blind) all the file-formats "on-the-fly"  about 20-25 songs from basic rock & roll (poorly recorded no less) to, well more sophisticated fare. In all but one case, WAv was the best sounding. By far.

Therefore, if I was to download a file (music), and I see that it is"packaged" in a FLAC container, I become suspicious since my listening tests clearly and absolutely preferred WAV.

That leads to my current question: Does the FLAC file/container hold the entire CD information, bit-for-bit as a WAV file ? I guess my question is there a difference between a flac-file and flac "container"?

Remember, I wish all of my music-files to be encoded in wav. (for the reasons noted above).

peter 

How different files will sound will depend on the renderer - in your case Foobar (and the computer on which it is installed), so what you find could be quite different from people using other renderers. Transcoding between different lossless formats with dBPoweramp and then playing should make no difference (e.g download .flac and change it to .wav), unless possibly to metadata if the tags formats are different. Note MQA flac compression is not lossless like standard .flac files.

Of course, how any will sound overall will depend on the DAC and rest of your system, and if your DAC is susceptible to RF modulation it will also depend on any isolation between the computer and DAC.

Mike-B posted:

I use dBpoweramp to convert any FLAC, or other format files, to WAV.   Its essentially a bit-for-bit format conversion & also carries over the metadata.    You do not lose anything - SQ or bits - in the conversion process.   

You can store albums as FLAC,  saving storage space, & transcode FLAC (or any other lossless format)  to WAV that feeds the Naim streamer/player.  This is done in your NAS wit its media server software,  but it depends on your NAS & its UPnP media server software capabilities,  some are better than others.   So what is your NAS & its media server software ???   Personally I prefer to convert & store as WAV (per 1st paragraph) rather than transcode.

Hi again (Mike-B): You really packed a lot of possibilities within your reply (above); let's unpack a bit.

 " I use dBpoweramp to convert any FLAC, or other format files, to WAV.   Its essentially a bit-for-bit format conversion & also carries over the metadata.    You do not lose anything - SQ or bits - in the conversion process."  (I must ask; did you actually listen/compare this to be true/ confirm?)

"You can store albums as FLAC,  saving storage space, & transcode FLAC (or any other lossless format) ....but it depends on your NAS & its UPnP media server software capabilities,  some are better than others.  

(Excellent insight. But there is the "but" qualifier. I have no doubt it is completely accurate. For  practical concerns, I too would prefer to convert & store as WAV -so long as someone has undertaken listening comparisons to verify if no identifiable differences exist between a straight PCM/wav. file and the same when converted/manipulated and played back)

pj

As for the accuracy or otherwise converting with dBPoweramp, if you can hear the difference between the .wav and .flac (ignoring the lossy MQA flac), then you can try yourself to see if dBPoweramp does as it claims:  

Take whichever type of file you have that sounds best to you (wav), use dBPoweramp to transcode to flac, then use it again to transcode the flac back to wav, then compare the sound of the two wav files: if they sound different the conversion is not perfect. If you can't, it is (or any imperfections are inconsequential) Of course best done blind to avoid any bias someone else doing the conversion and naming the original and recoded wav files identifiable only by them.

[@mention:72103499056291076]: you can convince yourself that transforming a .wav file to a .flac file and than back yields the original file by simply doing:

- flac afile.wav -o afile.flac (encodes afile.flac to afile.wav)

- flac -d afile.flac -o afile.1.wav (decodes afile.flac to afile.1.wav)

- diff afile.wav afile.1.wav (checks whether afile.wav and afile.1.wav differ)

This does not mean that afile.wav sounds the same as afile.flac, of course. For instance on a DAC that can only process .wav files, .wav and .flac files will sound very differently!

For me the best approach towards managing a music collection is the one put forward by HH: keep all your music as .flac files and transcode to .wav if your equipment sounds better with .wav files. I do this not primarily to save disk space (I have a rather small music collection) but because the tools that I use to edit metadata can handle .flac files much better than .wav files.

Mr Happy posted:

I can easily hear the difference between wav and flac. To me also wav sounds better. I feel transcoding must have some musical loss, although ive never tried this myself. If you have enough storage then why not just rip to wav in the first place?

 

Mr. H: Indeed, I too believe there must be some type of loss in any conversion process -for reasons not understood at this time. As a crude example, my flac/Wav comparison proves this beyond any reasonable doubt -employing listening skill.

And yes, I've recorded (about 100 so far) using the three/four file formats most used. The reason is simple and straight-forward; as a (streaming/file transfer) noobie, there are two things I expect to gain:

1) Verification: Determine myself if claims of a "lossless compression" (FLAC) are in fact accurate. So far, my listening tests/comparisons say NO WAY -in many cases, not even close. Since I've only started with this streaming/conversion process (as in a few weeks), I'm one who must continually do the comparisons (for at least up-to one year), for a greater "sampling" time and perhaps in the interim learn why this is/may be so?

2) Sharing: Music can do amazing (very healthy) things; it stands to reason (for me) that if anything can be gained by interest/effort, it was well worth the experience. Ultimately, anything that can both clarify/confirm while enhancing our musical enjoyment is something I strive to accomplish -and share.  I take hi-fi sound seriously (as in seriously enlightening), not 'head-in-vice' serious. It's imperative to me that any experiences and/or discoveries are shared, tested, discussed, quantified, verified and/or at least stamped "approved" -by passionate listeners. Posterity depends on us ! lol

peter 

Mr Happy posted:

ChrisSU posted:

Mr Happy posted:

I can easily hear the difference between wav and flac. To me also wav sounds better. I feel transcoding must have some musical loss, although ive never tried this myself. If you have enough storage then why not just rip to wav in the first place?

Loss of what? No bits are lost in transcoding. 

Musical loss. Im not very techy but ive heard differences when some people quote its all ones and zeros so there can’t  possibly be a difference. 

Ive never tried transcoding so I cannot say whether there is a difference or not, but im sure, as has been the case before, it won’t be a simple case of bits lost or not. But hey, maybe it would be identical sounding, as I said ive never tried.

Naim (legacy) streamers are said to have been developed to play WAV, and the theory goes that if you send one a FLAC stream, it has more work to do to unpack the data....and more work can create more electrical noise, which may affect the sound you hear. 

The point of transcoding to WAV on the fly is that your NAS does it remotely, so the data it sends to your streamer over the network is WAV, just the same, regardless of the format it used to be stored in. I’m not suggesting for one moment that you should blindly follow the ‘bits are bits’ creed that some people tend to push when discussing any digital matters, but in this case, I think it’s fair to say that bits actually are bits - or at least, packets of WAVs are packets of WAVs. 

dBpoweramp can tag and convert WAV files just fine. As I'm sure can most of the alternatives. The proposition that WAVs are somehow hard to tag doesn't hold up in practice.

As for objective versus subjective versus theory versus practice, all very interesting and ultimately moot. Stick with what you prefer the sound of. If you have no preference, please yourself.

NBPF, unfortunately that test is only valid for the file not specifically for the music payload, hence can reveal false negatives.

On conversion from FLAC to WAVE, the format the software uses for the metadata blocks in the WAVE file may not be precisely the same as the format for the original FLAC file, and some change may occur.  On going back from WAVE to FLAC, if the metadata in format the software uses for the WAVE file are not in a format that is directly compatible with the stricter format definition of metadata for FLAC, there will again be some data conversion of the metadata going WAVE to FLAC.  Unlike the definition of the music data, this conversion of the metadata isn't mathematically defined as lossless, so changes can occur here without affecting the music data.

S-i-S has checked the data in this conversion process and found them to be identical when decoded to LPCM (as is indeed guaranteed by the maths!).

Innocent Bystander posted:

Simon-in-Suffolk posted:

allhifi posted:

Do you feel, believe, know (lol) if a FLAC file can be 'un-packed' to a precise, bit-for-bit accurate WAV file ?

When I have done technical analysis of transcoded FLAC and Wav using WireShark and using Asset or MinimStreamer the PCM rendered has been bit for bit identical over gigabits of material. The technology used is prevalent in the IT world and digital communications - and if it didn't work much of our use of IT services would collapse into a big messy puddle.

Now here is the interesting bit - if there was some future encoding format or transmission format that didn't quite do bit by bit identical and added continuous low level digital errors, as opposed to analogue audio errors, then the resultant audio would sound a mess....

S

Or indeed present (MQA).

S.I.S.: Thanks for sharing. That's impressive (i.e. technically accurate) technical credentials.

Yet (for me), two things stand out:

1) Why can I (and many others) consistently, accurately pick-out/identify the better sounding format ?  (if above statements apply - I really don't know)

2) The "system" indeed works, no one doubts its (digital transmission) robustness. However, music (sound) reproduction authenticity has been shown to be exquisitely sensitive to unnatural sounds. And a such, one cannot use general digital file/transmission theory (or even practice) when evaluating authenticity (or simple enjoyment) of recorded music that our  remarkably accurate hearing mechanisms routinely exploit.        

Another MQA "shamer"  ! lol. I have never heard/experienced an MQA file, but the technical specifications of the format intends to/proves? that it does the opposite (of introducing digital errors). In fact, it claims (and shows/demonstrates) to correct any previous "errors".

Other than the floating around "money-grab" (primary) intentions of MQA (that I seriously doubt), I tend to pay attention when such a respected talent as Stuart (and team) has something to say. With that said, I'm really enjoying what's currently available (non MQA). If MQA does not (obviously) improve upon current  sound quality, I shall continue on as per usual. Nothing but gains can come from a deeper analysis into digitally recorded music, something MQA has invested considerably.  

pj

Harry posted:

I use dBpoweramp also. Life would be easier if I couldn't hear a difference between FLAC and WAV. But I do.  So I convert when I can't obtain a WAV file, which is rare.

Hi Harry: No kidding -about life being easier if ... !

Since Wav is so readily available, what wold be your (or communities) thoughts of an MQA file sent in a FLAC container ?

peter 

ChrisSU posted:

Mr Happy posted:

ChrisSU posted:

Mr Happy posted:

I can easily hear the difference between wav and flac. To me also wav sounds better. I feel transcoding must have some musical loss, although ive never tried this myself. If you have enough storage then why not just rip to wav in the first place?

Loss of what? No bits are lost in transcoding. 

Musical loss. Im not very techy but ive heard differences when some people quote its all ones and zeros so there can’t  possibly be a difference. 

Ive never tried transcoding so I cannot say whether there is a difference or not, but im sure, as has been the case before, it won’t be a simple case of bits lost or not. But hey, maybe it would be identical sounding, as I said ive never tried.

Naim (legacy) streamers are said to have been developed to play WAV, and the theory goes that if you send one a FLAC stream, it has more work to do to unpack the data....and more work can create more electrical noise, which may affect the sound you hear. 

The point of transcoding to WAV on the fly is that your NAS does it remotely, so the data it sends to your streamer over the network is WAV, just the same, regardless of the format it used to be stored in. I’m not suggesting for one moment that you should blindly follow the ‘bits are bits’ creed that some people tend to push when discussing any digital matters, but in this case, I think it’s fair to say that bits actually are bits - or at least, packets of WAVs are packets of WAVs. 

And yet (if this applies-I really don't know), why then is Flac (to this chap), clearly inferior to Wav?

Are bit not bits in this example (flac vs. Wav) ? As in I have a CD. I make copies of the CD as a Flac, Wav, AAC and AIFF file. I playback using Foobar 2000 (dB poweramp to rip). AND, WAV is easily identifiable (and i am "listening" through my desktop (outboard, though authentic 3-ways lol) speakers. This, over several days/weeks, in the AM, PM etc.

 Don't beat me up, I need to know whats what moving forward. If I can garner some time-saving experience/expertise on this forum, I'll be step or two ahead  ...

pj 

Is anyone here using AIFF files? I recall Jason Gould demonstrating both that and WAV at Bristol a number of years ago and, although the audience was split, it was clear to me that AIFF sounded better.

My NS01 rips to WAV btw, so that's what I mainly use at home. 

allhifi posted:

Mr. H: Indeed, I too believe there must be some type of loss in any conversion process -for reasons not understood at this time. As a crude example, my flac/Wav comparison proves this beyond any reasonable doubt -employing listening skill. 

Actually, what you've proved is that with your renderer .flac sounds different from .wav, not whether .flac is truly lossless or whether music can be transposed between the two formats without change. I refer you to my previous post for a way to assess the latter.

Huge posted:

NBPF, unfortunately that test is only valid for the file not specifically for the music payload, hence can reveal false negatives.

On conversion from FLAC to WAVE, the format the software uses for the metadata blocks in the WAVE file may not be precisely the same as the format for the original FLAC file, and some change may occur.  On going back from WAVE to FLAC, if the metadata in format the software uses for the WAVE file are not in a format that is directly compatible with the stricter format definition of metadata for FLAC, there will again be some data conversion of the metadata going WAVE to FLAC.  Unlike the definition of the music data, this conversion of the metadata isn't mathematically defined as lossless, so changes can occur here without affecting the music data.

S-i-S has checked the data in this conversion process and found them to be identical when decoded to LPCM (as is indeed guaranteed by the maths!).

Agree, my post was meant to illustrate that a reversible conversion between .wav and .flac formats is possible and therefore there is nothing, in principle, that speaks against transcoding if one wishes to do so. Of course, transcoding requires in practice some CPU time and might generate some additional heat. Thus, environment aware listeners will prefer to transcode in winter and refrain from doing so in summer. And hard boiled environmentalists might prefer not to transcode at all. I am probably a rather inconsistent envirinmentalist and an happy transcoder :-)  

allhifi posted:

Harry posted:

I use dBpoweramp also. Life would be easier if I couldn't hear a difference between FLAC and WAV. But I do.  So I convert when I can't obtain a WAV file, which is rare.

Hi Harry: No kidding -about life being easier if ... !

Since Wav is so readily available, what wold be your (or communities) thoughts of an MQA file sent in a FLAC container ?

peter 

Not answering for Harry, who no doubt will respond appropriately, but as already mentioned, MQA is not lossless. If you search the forum you'll find discussions about MQA earlier this year.

Innocent Bystander posted:

allhifi posted:

Mr. H: Indeed, I too believe there must be some type of loss in any conversion process -for reasons not understood at this time. As a crude example, my flac/Wav comparison proves this beyond any reasonable doubt -employing listening skill. 

Actually, what you've proved is that with your renderer .flac sounds different from .wav, not whether .flac is truly lossless or whether music can be transposed between the two formats without change. I refer you to my previous post for a way to assess the latter.

Absolutely...

As an example of a lossless conversion, take the ZIP file conversion,  it will compress most things, but what you always get after unzipping a file is the same data you put in in the first place.

Think of FLAC as being similar to ZIP, except is applies to the music content ONLY.  The metadata aren't compressed.  The music data are encoded (and compressed as part of the encoding) in such a way that you ALWAYS get the same data back out when you decode (and hence uncompress) it.  WAVE is uncompressed anyway, so as far as the music data are concerned, coded and then decoded FLAC  data are identical to WAVE data.

The reason WAVE sounds better is the extra internal electrical interference in the player, caused by the extra computing activity needed to decode the FLAC data stream.

allhifi posted:

ChrisSU posted:

Mr Happy posted:

ChrisSU posted:

Mr Happy posted:

I can easily hear the difference between wav and flac. To me also wav sounds better. I feel transcoding must have some musical loss, although ive never tried this myself. If you have enough storage then why not just rip to wav in the first place?

Loss of what? No bits are lost in transcoding. 

Musical loss. Im not very techy but ive heard differences when some people quote its all ones and zeros so there can’t  possibly be a difference. 

Ive never tried transcoding so I cannot say whether there is a difference or not, but im sure, as has been the case before, it won’t be a simple case of bits lost or not. But hey, maybe it would be identical sounding, as I said ive never tried.

Naim (legacy) streamers are said to have been developed to play WAV, and the theory goes that if you send one a FLAC stream, it has more work to do to unpack the data....and more work can create more electrical noise, which may affect the sound you hear. 

The point of transcoding to WAV on the fly is that your NAS does it remotely, so the data it sends to your streamer over the network is WAV, just the same, regardless of the format it used to be stored in. I’m not suggesting for one moment that you should blindly follow the ‘bits are bits’ creed that some people tend to push when discussing any digital matters, but in this case, I think it’s fair to say that bits actually are bits - or at least, packets of WAVs are packets of WAVs. 

And yet (if this applies-I really don't know), why then is Flac (to this chap), clearly inferior to Wav?

Are bit not bits in this example (flac vs. Wav) ? As in I have a CD. I make copies of the CD as a Flac, Wav, AAC and AIFF file. I playback using Foobar 2000 (dB poweramp to rip). AND, WAV is easily identifiable (and i am "listening" through my desktop (outboard, though authentic 3-ways lol) speakers. This, over several days/weeks, in the AM, PM etc.

'This chap' is comparing WAV to FLAC when the streams are received as such by his streamer. My point is that if the FLAC files are converted to WAV on the NAS, before they are sent to the streamer, then you are comparing WAV with WAV, and these can be proved to be bit perfect regardless of any lossless storage format that may have been used in the past - including the very recent past if transcoding on the fly.

Huge posted:

Innocent Bystander posted:

allhifi posted:

Mr. H: Indeed, I too believe there must be some type of loss in any conversion process -for reasons not understood at this time. As a crude example, my flac/Wav comparison proves this beyond any reasonable doubt -employing listening skill. 

Actually, what you've proved is that with your renderer .flac sounds different from .wav, not whether .flac is truly lossless or whether music can be transposed between the two formats without change. I refer you to my previous post for a way to assess the latter.

Absolutely...

As an example of a lossless conversion, take the ZIP file conversion,  it will compress most things, but what you always get after unzipping a file is the same data you put in in the first place.

Think of FLAC as being similar to ZIP, except is applies to the music content ONLY.  The metadata aren't compressed.  The music data are encoded (and compressed as part of the encoding) in such a way that you ALWAYS get the same data back out when you decode (and hence uncompress) it.  WAVE is uncompressed anyway, so as far as the music data are concerned, coded and then decoded FLAC  data are identical to WAVE data.

The reason WAVE sounds better is the extra internal electrical interference in the player, caused by the extra computing activity needed to decode the FLAC data stream.

From a completely non technical standpoint, and to agree with the final sentence from HUGE, when I first got into streaming ripped CDs, the theory was that this might well sound better than playing the original CD, as there was no ongoing error correction process being employed, as was the case with my CD player.

Similarly, whilst FLAC conversion to WAV, on the fly, can be shown to contain identical data to the original WAV file, I've always thought that the extra computing activity involved might somehow degrade the sound arriving at my aural receptors.

This might well be just another example of expectation bias on my part, but is part of the reason why, to me, WAV just sounds "better".