Why do my CD's still sound better than flac files?

Ignoring the moral ramifications, why would anybody have to repurchase anything in such a situation?

 

Convert FLACs back to WAV, buy stack of TY CD-R's, invest in CD binders, go nuts.

Problem solved.

I mean, you bought em once, right? *arrested*

 

Because the evidence seems to be slowly mounting that FLAC files may be sonically inferior to the original WAV files and we all know how fussy us audiophiles can be at obtaining the best possible version of any music that's available.

 

Besides there being little evidence that that's actually the case, converting from FLAC to WAV is trivial. Certainly much more so than re-ripping from the original CD.

 

-

 

I hope you’re being sarcastic because there’s no "evidence" whatsoever.

Additionally, lossless (in any format) will decompress to bit-perfect copies of the original.

 

Exactly. And, to take that a step further:

CD -> WAV

Will produce exactly the same file as:

CD -> WAV -> FLAC -> AIFF -> SDII -> SHN -> [insert 100 additional conversions between uncompressed or lossless compressed formats here] -> [send around the globe via the Internet a few times] -> WAV

 

Trivial? Yes. Identical? No.

If you have dbPoweramp installed, right-click on a WAV file and convert it to a FLAC file, then right-click on that FLAC file and convert it to a differently-named WAV file. You may notice that the file size of the WAV copy is slightly smaller than the file size of the original WAV.

I was the hoping the file size would be identical so that the DOS file compare command would settle the matter:

FC /B Orig.WAV Copy.WAV

 

Actually, identical: yes.

I just compressed a WAV using FLAC, then decompressed it back to WAV. File sizes are 100% identical. FC /B yields "no differences encountered".

There can be differences in header information in WAV files, but those have nothing to do with the audio data itself, nor with the FLAC process.

 

The tag info, metadata, and some optional WAV format chunks can change during a round-trip conversion like that. You need to do a compare that compares just the audio data. The DOS FC style compare won't work. EAC has a WAV compare tool that will work.

 

Wow. I didn't intend to set off such a debate! Oh well, very interesting. (I actually read them all!)

First, clarification of my signal path: I rip CD's to FLAC on my music room pc using dbpoweramp. Compression level is left at the default "5" because I don't know enough about these matters (other than what is stated by the creator of dbpoweramp) to know why I should change it. When a rip session is over, I send these files wirelessly to a pc in another room that "serves" as the music server. All command and control is done via my Squeezebox Classic located in the music room. The actual server software is Squeezebox Server (At least I think that is what they call it now.) I think all pertinent settings are at their defaults. I try to not mess with things that are working. But I digress...

FLAC files are streamed by SB Server via ethernet to my router. The router sends them via wi-fi to the SB Classic player which in turn outputs to my NAD CD players DAC via Toslink. The NAD C565 supposedly has a pretty good DAC, but I'm no expert on the subject. From there it goes to the NAD pre, on to the Adcom amp and out to my B--- (rhymes with Nose) speakers.

I guess I don't really know where in playback, conversion from flac to wav takes place...or if it takes place.

I am not offended by comments that I may be imagining things. I think I may have offered that as a possible explanation at the outset. In fact, I admit that is quite possible.

BUT...I was not listening critically, trying to discern differences in the sound of a CD straight from my CD player from the sound of the same music after having gone through this rather circuitous (!) route from CD to flac through two pc's, a router, an ethernet connection, two wi-fi connections, into a network player, and then back into the original CD players same Wolfson DAC. Geez, when you look at that it seems it would have to sound worse. But honestly, going in I was of the opinion my flac files would sound identical to the CD's from which they were ripped. My point is, I was not looking for even a difference, let alone a clear preference one way or the other.

But a preference is what seems to have developed. And this is not good news as far as I'm concerned, having ripped exactly 3744 tracks so far in lossless flac. And I'm just starting the L's. BTW, I don't rip CD's whole. I cherry pick, with the intention of just ripping the good stuff and leaving the filler off. More digression!

The thing I have to look at is whether or not I'm going to have to retool my process here and start over, perhaps with uncompressed flac, or maybe examine the signal chain. The only way I could change that much is to take the SB out of the picture and play direct from pc to dac. I don't really relish that. The pc in the music room is too noisey. If I have it in an adjoining room, then how about "command and control"? That's what the Squeezebox is all about in the first place - playing your tunes w/o having hte pc in the listening room...which it does really well.

I think the next order of business is a little experimentation on my part to try to determine if it's real or if it's Memorex!

to be continued...

 

There are some USB to SPDIF converters that would allow you to do USB to your DAC. Like the Musical Fidelity V-Link II. There's others as well. That will let you test what a more direct connection sounds like through your DAC.

Computers can be made more silent. Some options are to use a passive cooled video card, replace the CPU cooler with a model that is quieter, replace the case fans with quieter fans, isolate the hard drive or power supply to reduce vibration noise.

There's options too for controlling a remote PC. You can use a remote desktop app running on a tablet or laptop or netbook to control the big PC. Some media player software have Android or iOS control apps that will allow you to browse and play your music collection. So if you do find that something like the V-Link sounds better than the Squeezebox then there will be ways to make that sort of setup work conveniently.

 

Dear Barry,

yeah, I know that and here in Vienna we have done that a 1000 times. As I bought my equipment my dealer and I have made many double blind shootouts. The same question is discussed here also. People say, that they hear a difference. We have invited this people (from a audiophile trade show) and made double blind tests. No way that they a right on about 10%.

But you say that you hear a difference and now I will find out the reason. I don't say I don't believe you, but for me there is no experience that a good setup (Foobar - WASAPI Driver - DAC, who can handle 44,1KHz native and so on - which is bit perfect) WAV and FLAC files sounds different!

 

@dat56: Please look in your chain if your signal get's at any point resampled! Does the Squeezebiox resample the 44,1KHz FLAC files to 48KHz? Then it is clear, why it can sound different!

 

Evidence seems to be slowly mounting that static images can move...

Do I see it moving? Yes! Does it really move? You answer it...

That's the difference between theory and practice that Barry mentions... If you go by the practice then the answer is yes the image is really moving... if you go by the theory then it isn't. Am I wrong if I say it is moving?

 

Again, if you're playing back through the CD players D/A, vs. the Squeezebox's D/A converter, it's not a fair comparison.

My advice: rip the same CD in FLAC and WAV, listen to both kinds of files through the Squeezebox, and see if you can tell the difference. That's a fair test -- same signal path, same D/A, only different files. My bet is that they'll sound the same.

 

You are completely misunderstanding. There is no evidence, no argument, no one claiming that FLAC files are inferior to uncompressed WAV or AIFF. It has been proven beyond doubt that when you convert a FLAC or ALAC file to WAV or AIFF the result is bit-perfect. It's called lossless compression for a reason. As posted above, you could convert hundreds of times between lossless and uncompressed formats and the final file will still be bit-perfect to the original.

Anyone who has ripped their CDs to FLAC and then decides they want uncompressed files can simply convert the files and get the same result as if they'd ripped directly to WAV or AIFF in the first place. No need to repurchase and re-rip anything at all.

Where the difference arise, for those who hear any (not me, but I'm not going to tell someone else what they can or cannot hear), is in the playback of lossless files vs. playback of WAV or AIFF. Totally different issue, and one that has no bearing on the file itself. As I mentioned earlier, I think the only way to test this is to capture the steaming digital signal during playback and then bit-compare the FLAC/ALAC playback to the WAV/AIFF. (This would have to be done on a system where a difference is heard, of course.) If the ones and zeros are identical then I have no other explanation as to why they would sound different (i.e. if the DAC is receiving an identical steam of data how could the resulting audio be different?). If they are not bit-perfect then something in the playback mechanism is altering the data, handling it differently than it would for WAV/AIFF.

That been said, assuming that this is the case at least on some playback systems, you would think that it would eventually be resolved. There should be no practical reason, given today's average desktop computing power, that they shouldn't be able to decompress the file on-the-fly and get bit-perfect results.

 

These timing issues youre referring to sound like something wrong with the clock sending the bits into the DAC, or something wrong with the CD itself

 

Dear utenteanonimo64,

perfect example how our perception can be fooled!
Thank you very much!

Two things we have to split:

1. Is the whole chain from the HD to the DAC REALLY bit-perfect?
2. If this is the case, we get fooled by ourself!

 

Yeah, I had a debugger attched to the WASAPI Driver and I get the same ones and zeros regardless if I playback WAV or FLAC files. So, my DAC gets every time the same bitstream. So, why should this sound different?

I don't hear any difference and many of my audiophile friends also hear no difference. I say not that Barry can hear no difference. But for me there will be no reason that the identical bitstream will sound different. This is not possible!

 

Agree with Luke here. FLAC seems to be sonically identical in my listening.

 

Some level of jitter or time-based distortion exists in all digital playback. The key is to make sure it is below a threshold of hearing.

So in a sense, everything is "working" but better parts and design allow the process to have less jitter and even better sound.

 

One other note: in my experience power cables can greatly influence sound quality on digital gear. I suspect this is related to EMI and RFI but it is possible that all else being equal in terms of lossless coding being perfect, etc. the devices with better ac cords (assuming there are shielding issues) will sound better.

Digital is a funny animal. All sorts of things impact the sound even when the bits are perfect.

 

Yep, I don't know why but the same experience I also have made. Any one who can bring light into this?

 

I wouldn't think that it should. (Note that I didn't say "could" -- I'm not an electrical engineer, and don't pretend to have a full working knowledge of these things.) What I'd like to see is the same test performed on a system where audible differences have been noted.