AudioCircle
Audio/Video Gear and Systems => The HiRez Music Circle => Topic started by: testsound on 1 Jul 2020, 04:03 am
-
I read that WAV is the best quality. But uncompressed FLAC should be same as WAV in quality, correct?
-
Correct, though people still debate it. So many other things will have more impact on the sound of computer audio that will make it insignificant. Choice of computer server, digital cables, linear power supplies, nas, software, switches, galvanic isolation.....
Enjoy the uncompressed FLAC files.
-
WAV and FLAC are essentially equivalent. You can take a WAV file, convert it to FLAC (compressed or uncompressed - it doesn't matter), then convert it back to WAV - the initial and resultant WAV files will be identical. Run an MD5 hash on both WAV files for confirmation that they are identical.
As long as your computer server, digital cables, linear power supplies, nas, software, switches, galvanic isolation are capable of supplying a bit perfect stream into your DAC then you are golden. (Something in that chain would need to be misconfigured or severely broken to not get a bit perfect signal to the DAC.)
Martin
-
In the early days of computer audio WAV files did sound marginally better than FLAC because FLAC had to be expanded and played on the fly, everything was happening at the same time. Slow processors and error correction could have subtly effected the sound.
Current players like JRiver have Memory Playback, the FLAC file is first expanded into a WAV and stored in RAM then played back from memory. The computer is always playing a WAV no matter what the storage format.
-
Yes, FLAC and WAV are identical except in size. We use lossless compression all the time for money. You could have ten $1 bills, a dime, two nickles, and five pennies for a total of $10.25. It can be compressed into one $10 bill and a quarter. Further compression would be to write a check for $10.25. No matter which one you deposit at your bank, it gets converted to the number $10.25 as a credit to your account.
WAV files store the left and right channels sequentially with separate date for each. One thing lossless audio codecs does is take data that is identical in the left and right channel and stores it in a way that removes redundancy.
Current players like JRiver have Memory Playback, the FLAC file is first expanded into a WAV and stored in RAM then played back from memory. The computer is always playing a WAV no matter what the storage format.
JRiver's "Load decoded file into memory" option decodes both FLAC and WAV to PCM prior to the start of playback. The computer is always outputting PCM no matter what the storage format and regardless of whether it decoded prior to the start of playback or on the fly during playback.
-
WGH, mojave. Would the two of you make up my mind? :scratch::lol:
-
My mind is shot, I'll let mojave have a crack at it.
-
My library is about 80% flac and 20% wav (this doesn't include random MP3s that I have left out of the percentages). I tend to rip CDs to flac with lower compression but it doesn't matter, even highly compressed flac and wav sound indistinguishable to me. I briefly made the switch to wav a couple of years ago but after seeing just how much more space wav files were costing me, I did a pretty extensive AB comparison and went back to flac. I was running Jriver and had it set to load decoded files into memory before playing back then but 6 months ago I ditched my pc for a raspberry pi. I have had memory player turned off while running Jriver from the PI and still heard no difference in sq between flac and wav (my recent comparisons are quite limited though since I only have a few albums in both formats).
The Pi with Pi2AES however sounds dramatically better than my pc ever did. So like mhconley suggested don't forget about hardware.
If you are neurotic about it, just quadruple your hard drive storage budget and go wav.
-
If you are neurotic about it, just quadruple your hard drive storage budget and go wav.
Tagging really sucks with WAV files. JRiver plays WAV files fine but tagging and album covers are a PIA.
The internet's opinion seems to be: "FLAC is a PCM compressed file. Lossless. Like .zip, .rar, 7z or any other files compressors. The audio players which run the FLAC formats, decompress the FLAC on-the-fly and run the decompressed PCM."
https://forums.stevehoffman.tv/threads/is-pcm-the-same-as-flac.803948/ (https://forums.stevehoffman.tv/threads/is-pcm-the-same-as-flac.803948/)
Both FLAC and WAV are PCM containers.
Thanks mojave for prodding me to do more research :thumb:
-
I find they are close, but, IMHO, .wav is slightly better sounding than .flac with high resolution playback systems.
With memory so cheap nowadays, may as well use the best sounding file format.
-
but, IMHO, .wav is slightly better sounding than .flac with high resolution playback systems.
What is in the high resolution system you mention? How high does a person need to go, Devialet 400 ($17,495)?
Don't confuse the newbie, a common mistake is reversing memory and storage. Storage is cheap.
-
I find they are close, but, IMHO, .wav is slightly better sounding than .flac with high resolution playback systems.
With memory so cheap nowadays, may as well use the best sounding file format.
Is that true even with Uncompressed FLACs which have same megabytes as WAV?
-
Is that true even with Uncompressed FLACs which have same megabytes as WAV?
Depends on the system and where/how the decoding is being done. With something like Roon and isolated networked endpoints, the final device only sees PCM, so you can start off with any lossless format (compressed or uncompressed) and they'll sound identical as the final device receives PCM and not any of the formats. That's why for my Roon and master/backup library, I use the highest compression FLAC to save the most space and have the correct metadata. I always download and rip in FLAC.
In audio critical devices where the decoding is happening, check it out for yourself. If you're on Mac, you can use something like XLD to batch convert FLAC compressed, FLAC uncompressed, ALAC, AIFF/AIFF-C (make sure to check which kind it is - little or big endian), and WAV.
I went with WAV after testing in my secondary library for playback with BDP-1 (where the decoding is happening aka non Roon) that I've batch transcoded via XLD. I'd have probably gone with FLAC uncompressed next, which was just like WAV almost (wouldn't put money on it in a blind test). Unexpectedly, AIFF on the other hand stood out in a weird way. Not sure if something was broken somewhere or had a bug. Although all uncompressed files were the same size.
Theoretically and mathematically, they are all the same and have identical data. If there are any SQ differences due to CPU/IO byproducts, you'll simply have to try that out for your own rig.
-
I read that WAV is the best quality. But uncompressed FLAC should be same as WAV in quality, correct?
Yes, the only difference is the "shell", not the data itself. The amount of "additional processing" going on is basically zero.
To all those people who claim they hear a difference - show me in a blind test.
-
Memory storage devices are very inexpensive these days, so the size of music files is not as much of an issue nowadays.
There are slight deltas regarding processing between .flac and .wav files. Over time, have noticed that while the .flac files can sound excellent, there are some recordings where comparing the two formats, the .wav files do sound a bit more open and slightly more natural. This is noticeable with both speakers and headphones playback. I'll grant the fact it's subtle, and depending on the recording, may or may not be noticeable. Having said that, it's enough of a delta that I've gone back and changed a fair amount of stored recordings from .flac to .wav.
YMMV.
-
If you rip to FLAC and then later convert it to WAV, is there any loss in quality compared to if you ripped it in WAV in the first place?
-
If you rip to FLAC and then later convert it to WAV, is there any loss in quality compared to if you ripped it in WAV in the first place?
No. See my post above on how to irrefutably prove it.
Martin
-
If you rip to ALAC and then later convert it to WAV, is there any loss in quality compared to if you ripped it in WAV in the first place?
I read somewhere that ALAC ripped through iTunes doesn't sound good as FLAC.
-
If you rip to ALAC and then later convert it to WAV, is there any loss in quality compared to if you ripped it in WAV in the first place?
I read somewhere that ALAC ripped through iTunes doesn't sound good as FLAC.
ALAC like FLAC is lossless. Both will produce the same WAV file. I ripped a CD track to ALAC and FLAC, converted both to WAV then ran an MD5 checksum on both. The checksum is identical therefore the resultant WAV files are identical.
Anyone hearing a difference between FLAC, ALAC and WAV files of the same track is either imagining things (google expectation bias) or has something in the playback chain that is severely broken. A blind test would confirm which of the two is the cause of the difference they hear.
Martin
-
Yes Martin. Ripped same CD to FLAC and WAV. I'll be darned to my ears WAV took the cake.
charles
-
Yes Martin. Ripped same CD to FLAC and WAV. I'll be darned to my ears WAV took the cake.
charles
Do an internet search on the subject. From strictly a sound quality metric, the .wav files do sound better than .flac.
Agree with your assessment.
-
If you rip to ALAC and then later convert it to WAV, is there any loss in quality compared to if you ripped it in WAV in the first place?
I read somewhere that ALAC ripped through iTunes doesn't sound good as FLAC.
No, you can go back and forth between lossless formats infinite times and they'll carry the same data. It's entirely in the digital domain and can be verified. If data started changing, we'd have to worry about far bigger problems. There is no room or need for subjective interpretation or validation.
Any differences in SQ between lossless formats has nothing to do with the data itself. PCM streams are identical. SQ differences can only arise due to the byproducts of the processing of these files that somehow make their way across the circuits and impact audio circuitry.
The data is identical. The only thing I've heard in the data realm being different was something to do with how the padding can vary between different ripping softwares. I can't recall any further information or conclusion about that discussion.
-
Do an internet search on the subject. From strictly a sound quality metric, the .wav files do sound better than .flac.
Agree with your assessment.
My internet search on the subject turns up the exact opposite results. There is no difference in sound between .wav files and .flac.
As with politics it all depends upon whom you choose to believe.
I believe ABX testing on my own system with my own ears. I hear no difference. And before you say my hearing is not good enough or my system is not resolving enough I suggest you to try true ABX testing yourself. You may be surprised...
Martin
-
My internet search on the subject turns up the exact opposite results. There is no difference in sound between .wav files and .flac.
As with politics it all depends upon whom you choose to believe.
I believe ABX testing on my own system with my own ears. I hear no difference. And before you say my hearing is not good enough or my system is not resolving enough I suggest you to try true ABX testing yourself. You may be surprised...
Martin
We shall agree to disagree then. I'll post links that support deltas between the two. There is a delta with the processing, which means there should indeed be some difference. Many of the modern DACs have different filtering options, which DOES change the sound.
I wish there wasn't a delta, as now I have to redo a fair amount of my stored recordings. The differences are heard with both tube an SS, speakers and headphones. If you don't hear it, fine. I do, and it's real, not imagined.
Here's an argument regarding ABX testing
https://audiophilereview.com/audiophile-news/why-double-blind-testing-cant-work-for-audio.html (https://audiophilereview.com/audiophile-news/why-double-blind-testing-cant-work-for-audio.html)
-
We shall agree to disagree then. I'll post links that support deltas between the two. There is a delta with the processing, which means there should indeed be some difference. Many of the modern DACs have different filtering options, which DOES change the sound.
I have never heard of a dac that decodes flac and wav. That sounds like a digital transport. Filtering options for wav and flac files? Never heard of that either. Filtering for PCM? Yep, I have heard of that and those definately change the sound for PCM streams. They have nothing to do with flac, wav alac or anything else.
I dont think that anyone ever argued that there wasnt a processing delta either. Flac is compressed and has to be decompressed before playback. How large is the delta though?
A person would have to have a really old computer with terrible software or some old obsolete "audiophile" digital transport with poor implementation if it cant handle flac without a performance hit that affects sound quality. Lack of performance on that scale would likely result in playback dropouts not the usual comparison speak like "I thought the treble was less extended".
What you are really arguing is that there is some flac "artifact" left over in the pcm stream after it is decompressed or that flac is actually a lossy format. This is false and was covered in previous posts.
-
Here's an argument regarding ABX testing
https://audiophilereview.com/audiophile-news/why-double-blind-testing-cant-work-for-audio.html (https://audiophilereview.com/audiophile-news/why-double-blind-testing-cant-work-for-audio.html)
I love it. Toss out a bunch of questions about variables that have zero bearing on the actual ABX test, what is being tested, and the validity of the results, and claim they throw the entire process into question. That's how audiophilia works - obfuscation and slight of hand. "Pay no attention to the man behind the curtain."
Pop in a CD you know very well and rip a track to FLAC and WAV on your system. Use foobar2000's ABX comparator on the tracks and report how you did. Ripping software, cables, headphones or speakers, DAC, amplifier, ad infinitum - the playback chain does not matter one iota in this test. Have fun!
Martin
-
I have compared .flac and .wav of many recordings. With some, the deltas are not noticeable. However, with others, there indeed is a noticeable delta, and it's always in favor of .wav. The deltas can be best described as a more open and transparent playback in the presence region.
I'm also quite familiar with processing chain. Of course the DAC filters the incoming bitstream after the codec does it's thing. BTW, my Sony can perform additional processing prior to the DAC, which is DSEE-HX. That can also impact the sound.
The issue regarding sound quality deltas is likely hardware processing related. The hardware can be working as designed (not broken), and still, there can be a noticeable difference in sound in some recordings.
This link provides some additional insight:
https://www.thewelltemperedcomputer.com/KB/WAV-FLAC.htm (https://www.thewelltemperedcomputer.com/KB/WAV-FLAC.htm)
-
I think we're at the :deadhorse: point. Also, some seemed to have missed the OP was curious about ripped .WAV files compared to files ripped (presumably) with dbPoweramp in Uncompressed FLAC.
I think most will agree that a good system will expose the differences between .WAV and Compressed FLAC presumably because of the unfolding of the compressed file on the fly. If however, you can easily discern the differences between Uncompressed FLAC and .WAV you are a golden ear to be sure or have a simply amazing system. Either way, I continue to stand by my original comment that EVERYTHING else in the digital chain will have far greater impact on the sound quality than this. :D
-
I think we're at the :deadhorse: point. Also, some seemed to have missed the OP was curious about ripped .WAV files compared to files ripped (presumably) with dbPoweramp in Uncompressed FLAC.
I think most will agree that a good system will expose the differences between .WAV and Compressed FLAC presumably because of the unfolding of the compressed file on the fly. If however, you can easily discern the differences between Uncompressed FLAC and .WAV you are a golden ear to be sure or have a simply amazing system. Either way, I continue to stand by my original comment that EVERYTHING else in the digital chain will have far greater impact on the sound quality than this. :D
Sounds good to me. :D