 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
75.93.2.155
In Reply to: RE: Well said posted by kgiessler on May 28, 2007 at 15:37:24
At least on my laptop, WMAL files don't sound as good as a WAV version. However if converted from WMAL to WAV, they sound the same. So, while no data is lost, recovering the data in real time may introduce artifacts on some machines. Perhaps it would be wise to verify that one's computer is able to deliver the goods before taking the plunge.
Rick
Follow Ups:
"At least on my laptop, WMAL files don't sound as good as a WAV version."This is a playback or ripping software issue. Not a hardware or source data (WMAL/WAV) issue.
Sure it is a playback issue, but that is the point I think. Lossless will get all the data, but that doesn't mean that the real time "unzipping" doesn't muck up the sound vs. wav.
Due to all the extra work the hard drive (which is much slower than any modern CPU) has to do to read all that extra data (not to mention all the extra RFI caused by all that extra data moving across an unshielded hard drive cable, and all the added potential for read errors from the hard drive, which is much more error-prone than a CPU).I wonder, has anyone ever ABX'd a WAV file vs. FLAC (or some other lossless compression format), or is all this talk about differences between the two just someone's imagination?
I've done quite a few comparisons between WAV and FLAC in different system configurations.
And yes I CAN hear a difference under some circumstances and not in others.
In my current setup using a squeezbox I can definately hear a difference between streaming wav or flac (flac sounds worse) but I cannot hear a difference between wav or flac stored on the server. In all cases there are no bit errors, I have played a DTS recording stored originally as a wav file under all these situations and sent the digital out from the SB to a DTS decoder and I get the proper music out of it. If I do anything which I know changes bits (digital domain volume control, resampling etc) all I get is noise outof the decoder. This is a pretty good indication that whatever is causing the sonic differences is not bit changes.
With a USB DAC connected to a computer I can tell the difference between wav and flac under some situations. With short cables or the Optics USB cable I cannot tell the difference, but with long passive cables I can hear a difference. Checking bit accuracy is not so easy in this case. My guess is that the noise/jitter due to the long cable is getting to the point where it is significant to sonics, and changes in playback runtime (processing, disk seeks whatever) are causing a change in that noise/jitter which becomes audible. On the short or Optics cable the runtime differences are small enough they are not audible without the extra noise/jitter from the cable.
So I think its pretty well determined that from a bit perspective they are identical, but that under some circumstances in some systems there is an audible difference durring playback.
What I don't think we know yet is what the actual mechanisms are for those differences, there are a lot of hypothesis floating around, but testing them is going to be very tough. Especially when they are not audible in all cases.
I don't have any MACs so I have not done any tests with MAC formats. I know people that HAVE heard differences in MAC formats, but I did not take part in those tests.
John S.
Thanks for your informative post.
The results seem very odd to me, because the software performance should be unrelated to the cable that is being used. The software should be sending the same data at the same rate through that cable regardless of which format is used, unless we are dealing with buggy software, in which case the result would be random, not related to a specific cable.
What software were you testing with? Have you tried using the ABX component in Foobar2000 for testing, which makes this kind of comparison extremely simple (a positive result would put to the rest most arguments that there is a difference between the sound produced by the two formats)? I have tried this test, using some extremely resolving headphones in a very high-end setup, and cannot tell the difference between the two formats (nor could my wife, who has pretty good ears and really could care less about the whole argument).
ABX testing with Foobar2000 is about the easiest way to compare two files that I know of...just select two files, right click, and select "ABX Two Tracks..." from the Utils menu that pops up when you right-click.
My point is that playback will be dependant upon software / drivers. It is pointless to state that WAV sounds better that WMAL because it is system / software dependant.
I was wrong anyways. It could be a hardware issue also, given that de-compressing the audio stream may require a certain level of hardware capability. Results could also be affected by the ripping software. Different ripping software / options will deal with read errors differently.
Fair enough, but thats not a problem with the file, thats a problem with the computer your using for playback.
Given the parallels that I've drawn above, thats like being able to download the .exe of a program your computer is not capable of running. It happens all the time, but it doesn't mean there's a problem with the program.
But the result is the same as corrupt data: less sonic goodness. What's the hoary old line..."the operation was a success but the patient died".
Like most everyone I love the convenience of computer audio, and am very pleased with how good it can sound. But it seems that getting to the right software combinations and settings for a given hardware platform is every bit as demanding as setting up a good sounding analog system. Fortunately it's also as rewarding.
Rick
How about a better analogy then.
A poster posts to audioasylum and says "I'm going to start buying media. I'm either going to buy all 16bit/44.1khz CDs, or I'm going to buy all DVD-Audio and SACD or whatever hi res format is available. Which do I choose, I don't buy this hi-res format stuff"
And somebody posts and says "I've found that on my Mark Levinson No 39 normal CDs sound better than the either DVD-Audio or SACD when I play it on my Marantz DV6400"
Are we supposed to chalk it up to the fact that hi res formats are hogwash, or to the fact that the person playing them hasn't figured something out yet? Its pretty easy to see.
In the case of hi res formats, you *could* make a subjective argument if you wanted to.
In lossless compression, you simply cant. No matter what, you will be able to get your original data back. You can do a checksum, it can be verified. In fact it has been verified. It is lossless. You get the same data back.
As for playback, we can debate that. But as to whether or not lossless is lossless, well, its lossless.
I know this much...he'll start ripping that data, and presumably keep it for some time, probably through many computer changes. Whatever is lacking in his playback computer now will catch up with his files.
I made the mistake 5 years ago of going to MP3 instead of monkey's audio. I did it because of expense of storage. Now, I wish I had compromised to have a smaller collection of music so that I wouldn't have to go back and rip thousands of CD's again to get them in FLAC.
> > the mistake 5 years ago of going to MP3 < <
Actually, variable Bit mp3 is quite good. I doubt most people would be able to tell the difference between VB MP3 and WAV. I went VB MP3 a few years back and don't regret it at all. I went with the current technology at the time. Storage was an issue back then; now it's not. Now I only use Apple lossless.
Jim,
Mp3 is easy to distinguish from .wav.
I made 192 vbr files using lame (whatever their transparent setting was (-2 I believe)).
Anyhow, I ABX'd this quite easily (14/15):
foo_abx 1.3 report
foobar2000 v0.9.3.1
2006/10/13 17:49:17
17:49:17 : Test started.
17:58:14 : 01/01 50.0%
17:58:30 : 02/02 25.0%
17:58:48 : 03/03 12.5%
17:59:16 : 04/04 6.3%
18:00:06 : 05/05 3.1%
18:00:18 : 06/06 1.6%
18:00:33 : 07/07 0.8%
18:00:55 : 08/08 0.4%
18:01:12 : 08/09 2.0%
18:02:03 : 09/10 1.1%
18:02:22 : 10/11 0.6%
18:02:37 : 11/12 0.3%
18:03:13 : 12/13 0.2%
18:03:28 : 13/14 0.1%
18:03:44 : 14/15 0.0%
18:04:00 : Test finished.
If your system has enough resolution, it is pretty easy. I did it on my general computer with average head phones.
Here is the same file but Apple lossless vs. mp3:
foo_abx 1.3.1 report
foobar2000 v0.9.4.2
2007/03/31 22:12:43
File A: C:\Documents and Settings\Dawnrazor\My Documents\test mp3\Love Under Will.mp3
File B: C:\Documents and Settings\Dawnrazor\My Documents\My Music\Fields Of The Nephilim\The Nephilim\08 Love Under Will.m4a
22:12:43 : Test started.
22:16:18 : 01/01 50.0%
22:16:32 : 02/02 25.0%
22:16:48 : 03/03 12.5%
22:17:18 : 03/04 31.3%
22:17:48 : 04/05 18.8%
22:18:10 : 05/06 10.9%
22:18:31 : 06/07 6.3%
22:18:46 : 07/08 3.5%
22:19:02 : 07/09 9.0%
22:19:26 : 08/10 5.5%
22:19:36 : 09/11 3.3%
22:20:00 : 10/12 1.9%
22:20:14 : 11/13 1.1%
22:20:28 : 12/14 0.6%
22:21:00 : 13/15 0.4%
22:21:57 : 14/16 0.2%
22:22:16 : 15/17 0.1%
Easy.
Good report! You can make it harder on yourself if you are willing to take up a little more space and change some settings.
Figuring that perceptual compression was likely to be quite individualistic, I messed about trying to find the best compromise for me that would still be ~100MB/CD. I ended up with 256K, CBR, Stereo, Bit res. on, as I recall. I found the sound, especially the percussion, much better not using Joint Stereo or VBR. Anything less than 256K hurt the highs.
While I can hear a slight difference between this encoding and the source, the essence of the recording and performance comes through nicely and gives me a good listening experience. Using the recommended settings or speeds had far worse outcomes.
Memory space is now so cheap that none us us are likely to bother with MP-3 in the future for CD resolution audio. However I'm very impressed with how good it can sound and the room it can save if you are willing to take the time to find the appropriate settings for your ears and music.
Rick
sounds awfully good. Danny Kaey of Positive Feedback Online is writing an article about this codec, I am told, and hopefully it will be published soon.
Danny made a CD-R of various codecs for me to listen to a few months ago. All tracks were error-corrected prior to burning. Included were WAV, LAME 320 and LAME 320 hi-rez. Without a doubt, the LAME 320 (non hi-rez) was the best sounding of all....fuller and more natural-sounding, far more so than the original error-corrected WAV track. All tracks were ripped and encoded via Poikosoft.
Food for thought.
It does indeed!
I found it virtually indistinguishable from the CD, however it generated larger files that I needed for my plans at the time and I felt that since it's file sizes were approaching those of lossless encoding, why not go that route if one wanted even better.
So I use 256K for most stuff and lossless for especially 'good' recordings. After a year's listening I'm so pleased with it that I rarely use lossless. But the settings are very important, the recommended settings, that we are informed can't be beat, caused significant degradation to the transients.
It's eerie that you found that LAME320 sounded better than the WAV source. I've noticed very significant variations amongst software players and even different versions of the same software on my PC. You may find it interesting to do the comparisons using several different players. If they all produce the same result then I suppose the encoder caused the improvement. Computer audio certainly has added greatly to the audiophile's repertoire of tweakable things...
Regards, Rick
therefore equalizing the playing field for all codecs...
Using a single CDR should cancel out most drive problems however even using the same SW player doesn't mean that the codecs it uses for different formats have equal performance. Even multi-format codecs may have varying quality between formats.
If you were to transcode all of the samples to a single format THAT would eliminate problems due to the player and it's codec. The crumby thing is that while it eliminates playback as a variable, you end up with the same potential problems at the transcode step.
The only thing I can suggest is to repeat the tests using a wide variety of playback software and see if the results remain the same.
That ol' playing field may be flat, but it's got potholes...
Good luck, Rick
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
