Can you hear a difference between .wav and .flac of the same ripped CD?

[sandyk]

If a Mac Mini running iTunes, is so good, why are there so many other, usually expensive, software player programs available for an OS with perhaps only 10% of the overall market ?

 

[Downrange]

has not been favorable...

 

I've yet to find a program that works with Lion, (tried Bitperfect, Audirvana PLus, and Amarra) that can hold a candle to Asio/Media Monkey on an old XP Dell...

 

I've got both a brand new Mac Book Air and a Mac Mini (8gb). They fall far short using the same back end equipment. Just saying...

 

[mitchco]

http://www.computeraudiophile.com/blogs/FLAC-vs-WAV-Part-2-Final-Results

 

Lots of good comments too. Cheers!

 

 

 

[Paul R]

Pretty much for two reasons:

 

(1) A great deal of the music recorded in the last 20 years has been recorded using Macs. The artists and producers liked the sound. Windows machines just were not up to the task, at least not reliably. "Oh, let's do that take again, the computer just blue screened for no reason."

 

(2) Modern Macs, meaning Macs since OSX was introduced, as UNIX machines and don't have all the issues that Windows machines have, or rather, had up until Windows 7.

 

But in truth, it doesn't matter that much today. You can get really good sound out of a PC these days - JRiver MC is really really good. I get frustrated with it sometimes, but that is most because of the odd and annoying Windows issues. Like importing file names on a HTFS disk from Linux, you can easily run into file names that are too long to copy, or have a character in them Windows chokes on, etc., etc., etc.

 

Also, Macs running iTunes are easy to make bit perfect, which is, again in my opinion, the necessary starting point for any CA system. Windows are much less so, unless of course...

 

You buy a third party music player like JRiver.

 

To be honest, Unix/Linux machines are where I seem to feel the best sound is at, right now.

 

-Paul

 

 

 

[jhwalker]

I don't there are any more third-party players for Mac than there are for Windows - they're just better

 

Seriously, the only deficiency I note with iTunes is that it doesn't automatically switch sampling rate to follow the file being played - I'd rather not have every file I play be resampled to some arbitrary starting point set in Audio MIDI. I use BitPerfect just for the automated sample rate switching - otherwise, I'm pretty happy with the sound I get straight from iTunes.

 

That said, I think there's something to be said for the integer / direct modes the Mac developers are building. I like the idea of talking more directly to the DAC than is possible thru all the Core Audio rigamarole, convenient as I'm sure it is.

 

[sandyk]

Downrange

Do you have the deluxe version ?

There are 2 very nice live rehearsal performances in 16/48 on the DVD. You can also rip just the audio with a suitable program.

Regards

Alex

 

Assuming you are using Windoze, have you tried re-ripping it in Safe Mode ?

 

[Downrange]

The Deluxe version of what?

 

And, yes, you can convert from FLAC to WAV and vice versa, of course. I've actually been doing that this afternoon in "batches" to save some time. There is some debate about this, but that's another thread - for my purposes, the restored .wav files sound as good as those I'm ripping over again, so I'm going to leave it at that.

 

Now, back to listening. As I've said, I've been using MediaMonkey with Asio4All to get bit-perfect playback, which just frankly blows away my two new Mac machines (Lion) using anything (have them all). But, here's the thing. I'm hearing a very large difference in every single CD rip playing back as .wav vs. flac. I don't know if it's the process of having to convert "on the fly" or what, but this is not a subtle difference. Just reporting here and continuing to evaluate, but so far, it's across the board. Every CD rip sounds better played back as a .wav than as a .flac.

 

I'm going to experiment with my hi-res flacs next...

 

[sandyk]

downrange

There is a deluxe version of "Nora Jones-Not Too Late"

On the DVD there are live rehearsal performances of "Until the End" and "Sinkin'Soon" as well as other material.The live performance videos are not compressed.

The Absolute Sound recently reported about .wav vs. .flac, and as expected it didn't go down too well with some here.

(TAS 220 and TAS 221)

 

Regards

Alex

 

[Downrange]

No, I have the standard CD of Norah's "Not Too Late." Yes, been following the TAS debate - another thread, though.

 

What I'm trying to ascertain through this thread if others can hear significant difference in playback of .flac vs. .wav. I understand that many programs "spool" the files ahead and "supposedly" take any issues of the conversion out of the "real time" domain, and bit streams should be equal, and all that. I'm just saying I'm not hearing that - not with this setup, which is my reference.

 

I'm hearing a very large difference between flac and wav playback of 16 bit material.

 

[elcorso]

Welcome!

 

You are one more of the privileged CA member than can listen to the difference.

 

The other ones listen to checksums...

 

BTW, there is another format called Apple CAF (A-Law 2:1) where you can format the file as Linear "PCM, 16 bit little-endian signed integer", that by coincidence has exactly the same SQ as WAVE.

 

As always Apple went for the way they wanted with AIFF, making this format iTunes favorite.

 

But I would't recommend it, since it is away from the standard

 

As you can see I like better the little endian, since big endian could be dangerous. (At least in cowboys & indians movies).

 

Roch

 

[elcorso]

Hi Alex,

 

For the convenience of use it, that' the only reason that comes to my mind. Not all the users are crazy music fans, or maybe the proper term is 'purist', like you and me.

 

Kind regards,

 

Roch

 

[Downrange]

I've satisfied myself about the advantages of playback of .wav vs. .flac for 16 bit, now I'm taking it up a notch. I've got a good bit of 24/44.1 material here, and I purchased it as .flac. I'm beginning the process of comparing extensively those .flacs with .wav conversions.

 

First up: Yevgeny Sudbin's wonderful Beethoven Piano Concertos (4 and 5). 24/44.1. I've been using these as reference material since December, and they are truly great sounding downloads (from eClassical). Well, tonight I converted No. 4 to wav. Yes, big difference - not as large perhaps as 16 bit rips, that go from sometimes as small as 400 bitrate to 1411 bitrate, but still a big change, in this case from around 1100 to 2116 bitrate.

 

Who knew? I thought playback of FLAC was just fine... Turns out it's not. At least, not in my system. Your mileage may vary. Let me know what you find.

 

[Downrange]

I've just about finished converting all my flacs (24/44.1) to .wav, and it's simply unanimous - everything, absolutely everything sounds better.

I will try using XLD and other software to restore my 24/48 and 24/96 flacs to full waves next. This is fun!!

And cheap - memory is cheap these days!

No one is trying this? Huge difference...

 

[Paul R]

I convert FLAC, ALAC, WAV, and AIFF to PCM just before I stream it to a player.

 

Sounds glorious and absolutely eliminates the vey minor differences between formats I believe I seem to hear. (grin)

 

As a side benefit, the player has a very small power supply, and is galvanically isolated frome the network cable and the connection to the DAC.

 

It just keeps getting better too, as the player is utterly noiseless and the storage machine is a honking big server stuffed in an equipment closet, out of sight, and thankfully, out of hearing!

 

The sound to me is about as good as Amarra on a Mac. Which is pretty darn good.

 

I decided to just get rid of the whole format problem that way. Works - very well - for me.

 

By the way, works well even using JRMC as the player,mwith an SMB mount for the music, and one for movies ...

 

Paul

 

 

 

[Downrange]

Cheers!

 

Glad you found a solution that worked. I'm just blown away at the difference here between flac and wav. Never suspected it until I got the V-Link 192. That thing is the schizznit.

 

[prufrock]

I would be interested to hear the results of an ABX listening test. Do you intend giving it a go?

 

[Downrange]

seem to be in what I would term "inner detail," especially in the upper octaves of the human voice.

But the same applies to accompaniment and orchestration, it's just harder to hear. What is the .flac doing to the music? In some cases, it's dulling a certain "edginess" - so if the material or the playback system (either or both) is adding here, the .flac may actually sound "better." But it's better because of an undesirable characteristic of the source material/playback system that's being masked. I feel this is where a lot of testers can err. There are a few cuts on "Not Too Late," for example where Jones' voice is just bit on the hot side. Some of this is probably due to the bit depth being 16, rather than 24, as I've seen this problem in virtually all female vocals. Still this is a "reference disc," according to HP, so I'm going to hold it to a high standard here.

 

I want to drop some of these test segments into Audacity and try to see if there's anything at all that can be measured in the waveforms. I feel that if the tools is sufficiently analytical, what I'm hearing must be able to be seen, as well as heard.

 

Moving to 24/44.1 material, the differences can be more subtle. But at least when the upper frequencies are "unmasked" they aren't over-etched like CD material is. The inner detail of symphonic orchestration, especially with massed choral voices accompanying, are even easier to hear than with CD.

 

As for ABX listening tests, I'd really rather not get into that hairy ball of wax. I'll just say that many find this type of testing to be highly unreliable as one simply cannot establish adequate controls between two different testing venues. (If you wish to "duplicate" a test, you must have the exact same components, players, etc.

 

Let me say just a bit about how I test for A/B. The easiest person to fool is yourself, so you have to take some precautions to ensure there's no confirmation bias. All you need is one assistant.

 

1. Select test material - ensure you take VERY SHORT selections - no more than fifteen seconds, often three or four seconds will do - from each piece. It's extremely important that these snippets contain the specific musical material that diffentiates the two samples in your hearing. If your results are like mine, you'll find there are a few places in any recording where the differences really contrast. Choose those.

 

2. Run through the test yourself, listening for what is different, and what makes one better than the other for you. Have the assistant randomly play the two short selections and try to identify them as a "blind listener." If you're "imagining things" or suffering from confirmation bias, this will root it out quickly. Be sure to avoid "listener fatigue" and "over-testing syndrome," as this will increase your error rate.

 

In general though, subjectivity must always be the rule. You're never going to establish one standard for a given musical reproduction that everyone will agree upon. The best system to test in is your own. Perfect reproduction is an "impossible ideal," but we get closer by testing and listening. The differences I'm hearing have to do primarily with INNER DETAIL. Either you can hear them in your system, or you cannot. Here, it's very obvious, but as said before, only recent improvements to a fairly high end system really brought them out.

 

So, enough on testing. People love to argue about that, and I'm simply not going to engage them. If we believed that subjective testing was not valid, and that engineering measurements were everything, we'd all be (still) using Dynaco Stereo 120s and Pat-4s.

 

Back to music. With YOUR system (that can play both .flac and .wav) can you hear the difference in playback? How do you characterize it? And, if you hear a difference, what do you suppose is responsible? My thesis for now is that the "unspooling in real time" of a FLAC file is not as musical for some reason as the same playback of the "already unspooled" WAV file. Nothing magical, really. It's an extra step, and may even be that only my playback software shows this difference. (Media Monkey with Asio 4 All - still the best thing I've heard, much better than anything out of my MAC/Lion platforms, but I haven't tried JRiver or JPlay in the Dell yet.) We all have work within our references, and that's mine, at present.

 

[prufrock]

I have just started down the computer audio path and about two thirds the way through ripping our CDs to Wav atm. Had originally intended to do it in Flac due to the fact that its supposed to be better for embedding metadata. Changed my mind when I read that Apple doesn't support it.

I intend to start buying some hi res downloads, so am curious regarding your findings. You mention that only high end systems might pick it up. Have just bought a Stello DA100SIG DAC and am picking up the U3 DDC to run between it and the laptop tomorrow. Haven't got the dedicated laptop yet.

Did you do any of your testing blind, i.e with an assistant?

 

[Downrange]

Thanks for the link to the diff testing thread - very good stuff!

 

Just a quick comment - I don't think the differences I'm "hearing" are down 90 db! (as the flac vs. wav differentials seemed to be). I'm not sure how the increased "inner detail" would appear in terms of waveform analysis, but I'd like to find out.

 

Edit: Prufrock, I'm not familiar with your equipment and cannot say anything about it. At any rate, we all go with what we have. I'm still making improvements here, so nothing is fixed in stone. Just listen with "what you brought" and find out what you hear.

 

Edit two: Oh yeah, I buy DLs in flac. I pay for bandwidth by the GB here, so that's an easy decision. I haven't found any compelling reason to purchase them and download as .wav (even if available). Frankly, thus far, I've been more than happy with the conversions to .wav I'm getting here.

 

[goldsdad]

"I have just started down the computer audio path and about two thirds the way through ripping our CDs to Wav atm. Had originally intended to do it in Flac due to the fact that its supposed to be better for embedding metadata. Changed my mind when I read that Apple doesn't support it."

 

Why not AIFF instead of WAV? AIFF is Apple's uncompressed alternative to WAV, and has the advantage of embedded metadata. AIFF can be played by all audio players on a Mac and most players on other platforms.

 

Also, since you were prepared to use FLAC except that it isn't well supported by Mac software, Apple's ALAC format is similar to FLAC, being lossless compressed. It doesn't have an uncompressed option like FLAC has, though.

 

 

 

[sandyk]

downrange

Diffmaker, and other methods using Audacity etc. will show no differences.You would probably need to use something like a Digital Storage Oscilloscope to examine the waveforms at analogue Out of the DAC.Even then.... Have you tried copying one of your better sounding high resolution .flacs to a USB memory stick, then converting it to.wav saved at the same location ? Try playing both .flac and .wav from a program like cPlay etc. that plays from System Memory, and compare that against playing both types from the original location.

Regards

Alex

 

[Downrange]

Yes, I believe you're right. The changes to waveforms are probably very subtle and have to with complex waveforms in time and with tiny phase differences between files. Still, it would be nice to be able to quantify what is heard.

As to other experiments, right now I'm just comparing flac vs. wav, without looking at much else. I realize that's an area of some interest for many, but for me I'm just trying to nail down this one comparison.

For my money, I'm putting EVERYTHING into .wav. It's much better playing back here.

 

[mitchco]

If you are hearing a difference when making a comparision, then the waveform must have been altered from the original. And if it has been altered, it can be measured. That is what Audio DiffMaker was designed for: http://www.libinst.com/Audio%20DiffMaker.htm

 

Good info here to see the level of resolution at which Audio DiffMaker measures at: http://www.computeraudiophile.com/blogs/1644-vs-24192-Experiment#comment-137251

 

If you plan on performing ABX listening tests, try using Foobar2000 with the ABX plugin:

 

Cheers, Mitch

 

[sandyk]

Audio DiffMasker is unable to show these differences.

That does NOT mean that there aren't any though.

 

[4est]

And there certainly is a difference in the sound of lossless files around here...