Article explaining why wav and flac sound different

Interesting indeed. Can someone explain what this "estimated distance above B&W 802D MR" is all about ? It this the perceive location of the stereo image ?

I understand it literally means the distance above the speakers where a certain chord appears to come from in a specific recording of  "España". The higher above the speakers the better the sound quality (according to the authors).

Yes interesting,  it seems my decision to go WAV was the right one, I did it for what I thought were logical technical reasons & the info available at the time.  In testing I can hear a small difference between FLAC (5) & WAV,  I can't hear any difference when transcoding FLAC (5) to WAV.   And that is probably where the bottom line is,  irrespective of what the measurements & technical reasons are,  can the listener hear the difference? 

Very interesting, as you can imagine it is going down well on the forums with the usual bits are bits arguments raging fiercely 

I turned off at...

Part three reported that a FLAC file sounded inferior to the WAV file from which it was made, and we found to our surprise that when these FLAC files were reconverted, the resulting WAV file did not recover the full sound quality of the original. We repeated these conversion steps five times and observed a hyperbolic decline in WAV sound quality, the greatest loss occurring in the first two or three conversions.

Sorry but (excepting the possibility of corruption and possibly fragmentation) it is fundermentally impossible for a computer to "know" if a file has been converted into a flac file and back again.  

Worth noting IMHO is that these results were found using computer audio replay with JRiver Media Center. The results may or may not apply to streamers from Naim or others. If I'm not mistaken my dealer still recommends FLAC for Linn streamers.

jfritzen posted:

I understand it literally means the distance above the speakers where a certain chord appears to come from in a specific recording of  "España". The higher above the speakers the better the sound quality (according to the authors).

Whist stereo imaging per se might be a relevant "test" of quality, when is a broader image "good" vs "bad?"  Or is the assumption that no matter how something was recorded, or what the producer / engineer wanted, "bigger is better?"

Yes - I have found similar things - BUT it depends on the FLAC encoder and decoder. There are many variables to FLAC encoding and no two encoders for a given compression setting may be implemented the same or use the same settings even if they use the same underlying libraries.

So when you encode to FLAC - do a test with your encoder and decoder to check the files are the same and measure the same afterwards - there are tools to do this. Now yes we don't know how Naim decodes FLAC and I have no way of measuring that - but you can measure transcoded WAV to FLAC and checking the media data in reconstructed PDUs. But to my mind there is a question mark to purchased media in FLAC - I am now buying download in WAV and NOT FLAC.

And there are more variables with transcoding - I am investigating this now - and it is to do with media frame timing across the network. I am analysing this at the moment when I have time - and there are DEFINITE differences between media servers and I wonder if this is what the Melco guys stumbled across.....

Good game 

Interesting findings concerning the effect of metadata in the audio files. Perhaps this might explain Naim's decision to have this data separate on their US.

I think naim would be disingenuous to suggest thats the reason, WAVs are a total arse to tag and 'sidecars' of tag data is not at all uncommon.

I have not read the article, I have never heard any difference, this is a good thing because FLACS are so much easier to handle.

An interesting article indeed, of which I (think) I followed, at least, some of it!

Simon - you're point about only buying WAV downloads. Whilst the likes of Qobuz offer a choice, many other do not have that option - FLAC only. I use DBpoweramp, and can obviously convert downloaded FLAC to WAV, but on reading the article, am I actually gaining any benefit in doing so?? By the way, I currently hold my "library" on PC and copy to USB to play in the nDac. 

When I extract a ZIPped PDF, some of its o's are changing to ø, ö, ó, etc. I swear this happens more aggressively the first few times a reZIP the PDF.
After a few ZIPs and unZIPs it settles to some semi-readable form. 
Now if I decide to use different ZIP algorithms, then the result is, well, anybody's guess.

Tongue in cheek all of that ;-)

I would accept that further processing inside the decoder of a DAC or a server, may potentially mean more work for the CPU, more heat, more noise, etc, at the expense of sound quality, though I haven't witnessed it myself. But that a (lossless) decompression algorithm will have an effect on SQ I think it's more of a bug than a feature.

Very interesting. This may explain why Naim UnitiServe / HDX rips to WAV and stores the metadata separately from the music files.

I will continue with WAV and AIFF.

What has meta data to do with this??? Its completely unrelated

I am not disputing what the article says (yet) but there are things I simply cannot understand. The size of the album art having an effect on the sound ?

Why can't we simply do an A/B comparison by comparing the spectra output of the analog signal rather than using height of stereo image ?

Simon-in-Suffolk posted:

What has meta data to do with this??? Its completely unrelated

 

Simon, page 2 of the report's findings states

"We can say with some confidence that the metadata-associated cover art is primarily responsible for the decline in height reproduction and associated sound quality when WAV and FLAC files are interconverted"

"While the metadata effect and its audible losses can be minimised or reversed, the FLAC compression and decompression effects appear to be permanent unless replaced by clean metadata that has never been exposed to prior conversions"

"On the basis of our results thus far, we can make specific recommendations to the music download industry for issuing improved music files:

• Erase metadata containing historical records of previous repeated WAV-to-FLAC-to-WAV conversions and replace with fresh metadata prior to final distribution."

Interesting stuff. I've posted before that I felt that FLAC sounds a bit flat in comparison to WAV. However I settled on AIFF as my preferred file format as I could not tell the difference between it [AIFF] and WAV, but it still retains the ability to attach metadata thus providing the best of both worlds. 

I also read the article and find some sense in the difference after encoding, regardless if it's tags or another reason. But I am also getting at the point where I have the feeling the complexity of all components to the optimal sound is getting too high. So I will definitely stay with flax which is then transcoded on the fly to wav, I might once to a test on comparison wav and transcode wav......., but don't have a lot of appetite for doing that....

sjbabbey posted:

Simon-in-Suffolk posted:

What has meta data to do with this??? Its completely unrelated

 

Simon, page 2 of the report's findings states

"We can say with some confidence that the metadata-associated cover art is primarily responsible for the decline in height reproduction and associated sound quality when WAV and FLAC files are interconverted"

"While the metadata effect and its audible losses can be minimised or reversed, the FLAC compression and decompression effects appear to be permanent unless replaced by clean metadata that has never been exposed to prior conversions"

"On the basis of our results thus far, we can make specific recommendations to the music download industry for issuing improved music files:

• Erase metadata containing historical records of previous repeated WAV-to-FLAC-to-WAV conversions and replace with fresh metadata prior to final distribution."

 

But as I have pointed out before, that is down to the decoder implementation, and differences between encoder and decoder. Meta data itself, or more including  Vorbis Comments is an integral part of the format structure, but is seperate from the media itself. However the way the decoder handles that can possibly lead to different code sequencing execution and therefore noise produced.... but it is seperate from the FLAC encoding itself. The same can equally hold true for WAV, ALAC and AIFF. If this is an issue for you , then arguably this is an advantage of streaming via UPnP where the metadata is decoupled from the media within the file before going to the streamer from the media server.

if you want to get a better feel for the FLAC variations and why I feel it's important match encoder to decoder have a peruse of the FLAC standard itself.

https://xiph.org/flac/format.html

Simon

HardBop posted:

An interesting article indeed, of which I (think) I followed, at least, some of it!

Simon - you're point about only buying WAV downloads. Whilst the likes of Qobuz offer a choice, many other do not have that option - FLAC only. I use DBpoweramp, and can obviously convert downloaded FLAC to WAV, but on reading the article, am I actually gaining any benefit in doing so?? By the way, I currently hold my "library" on PC and copy to USB to play in the nDac. 

Food for thought.. If you are purchasing FLAC, then to my mind you should ideally know what FLAC encoder was used... that way you should be in a position to confirm you can fully accurately decode it.. and of course many quality encoders allow you to form a correlation check with its decoder to confirm no residual errors have crept in.

As far as converting to WAV, there is probably no advantage doing that for storage, but there are often perceived advantages to do this transcoding prior to play back, as less local electrical noise is produced per second of audio when decoding FLAC as opposed to WAV-PCM, and if this noise has a risk of modulating ground planes or clocks that are important in the audio playback chain then almost certainly WAV will be beneficial.

Mayor West posted:

Interesting stuff. I've posted before that I felt that FLAC sounds a bit flat in comparison to WAV. However I settled on AIFF as my preferred file format as I could not tell the difference between it [AIFF] and WAV, but it still retains the ability to attach metadata thus providing the best of both worlds. 

Coming from an iTunes setup, I also now use AIFF. I did do a comparison for my own satisfaction between ALAC and AIFF, and the latter definitely sounded better. Doing a similar comparison but with WAVE files instead of ALAC, I couldn't detect any difference. Not scientific I know, but good enough for me.

tonym posted:

Mayor West posted:

Interesting stuff. I've posted before that I felt that FLAC sounds a bit flat in comparison to WAV. However I settled on AIFF as my preferred file format as I could not tell the difference between it [AIFF] and WAV, but it still retains the ability to attach metadata thus providing the best of both worlds. 

Coming from an iTunes setup, I also now use AIFF. I did do a comparison for my own satisfaction between ALAC and AIFF, and the latter definitely sounded better. Doing a similar comparison but with WAVE files instead of ALAC, I couldn't detect any difference. Not scientific I know, but good enough for me.

AIFF and WAV differ in word order. Maybe iTunes (and/or other Mac tools) are 'optimised' for AIFF and playback of WAV requires an additional 'thing' to do which might have a sufficient effect for you to hear a difference between AIF and WAV?

just a thought ;-)

I always buy uncompressed and rip in AIFF. Compressing to ALAC when using my old MM setup and with the Naim and Linn streamers never sounded quite as good - Just a bit greyer and flatter than full fat AIFF so it always stayed that way.

Doing the same comparison with the Melco, i found no difference between ALAC and AIFF so i've gone the whole hog and converted my whole library to ALAC. 

I can see the temptation to "purchase wav files," but how are we to know if they are "virgin" wav files, or have been transcoded who-knows-what-ways-and-how-many-times prior to being uploaded onto some server by a technician?  

This is one area in which, if somehow the provenance of the file is important, we have little hope of accurate knowledge regarding purchased files.  As we all know, online sellers are not all that reliable regarding just what we're buying (HDTracks comes to mind).

Based on my own, albeit tin-eared, testing, I started ripping my CDs to AIFF, after I realised the tagging in WAV format was shite. This came after I had already ripped some hundreds of CDs in FLAC and, previously in iTunes, ALAC 

I'll have finished re-ripping them in a few years. No doubt, by then, some new 'improved' format will have come along to blow these away.