Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: [NONSENSE] Blind Abx Test Of Wma Lossles... (Read 4760 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

[NONSENSE] Blind Abx Test Of Wma Lossles...

In trying to rip my 250+ CD Collection to the PC, I chose Windows Media Lossless as my chosen format, for various reasons, mainly the library, media player support, and the fact that I use Sound Forge in my profession - and it supports WMA9 Lossless - meaning I can convert back when and if I need to

But in the process I also believe that the sound of the lossless file is different to that of the extracted wav file in EAC or Sound Forge, so...

Ok, so here's what I did.

I know that most here know what foobar is and what explanation that is given is not to patronise those who know, simply to explain to those who don't yet know the process. 

I extracted Track 2, The Logical Song from Sound Forge 7 and saved as a 16bit 44.1khz Stereo wav in a folder called blindtests.

I then ripped the same track in Windows Media Player 10 to WMA Lossless into the same folder.

I then opened the lossless file in Sound Forge and saved that as a 16bit 44.1khz wav file in the blindtests folder.

I used the program foobar2000 http://www.foobar2000.org

which has an ABX comparator, where it allows you to blind test 2 wave files, very handy when like me you are concerned about the placebo effect, which is hearing things that aren't there.

So I downloaded the ABX component from here http://www.foobar2000.org/foo_abx.zip

and put the dll in the components folder of foobar2000 under c:/programfiles/foobar2000/components.

Opening Foobar2000, I added the two files from the blindtests folder, and Ctrl-A selecting them both then right clicked to select the ABX Two Files option.

This brings up a series of tests, and BY GOD is it tough to concentrate on a short piece of music for a long time, but alas, the results are below, proving that for the first 10 minutes or so, I could hear differences in the test, but after a long period of listening, my ears became too accustomed to it and I started making mistakes.

My theory is the following:

If you rip in EAC to wav, ensuring one of the best rips possible, and then convert that file using the command line utility and selecting lossless, you are getting a lossless file.

If you rip directly from CD inside of WMP10 to lossless, I believe it is doing a lousy job at ripping correctly in an uncompressed format, capturing all the data correctly before then converting to WMA Lossless.

So, for now, I'll continue as I have been, avoiding direct rips inside WMP10.

I welcome anyone to try this again properly, and let me know if you have any opinions on it all.

I have not included a test sample as it was the entire song that I used, and copyright laws do not permit that...

Results:

foo_abx v1.2 report
foobar2000 v0.8.3
2005/02/18 11:46:14

File A: blindtests\Extract from CD 1 - Track 2.wav
File B: blindtests\02 The Logical Song.wav

(percentage is the chances that I was guessing, score is how many times i was right when choosing)

11:46:23 : Test started.
11:48:30 : 01/01  50.0%
11:51:23 : 02/02  25.0%
11:53:32 : 03/03  12.5%
11:54:36 : 04/04  6.3%
11:58:04 : 05/05  3.1%
11:58:49 : 06/06  1.6%
12:00:10 : 07/07  0.8%

12:00:58 : 07/08  3.5%
12:01:20 : 07/09  9.0%
12:01:43 : 07/10  17.2%
12:02:22 : 08/11  11.3%
12:03:38 : 09/12  7.3%
12:04:30 : 09/13  13.3%
12:04:47 : 09/14  21.2%
12:05:15 : 09/15  30.4%
12:05:39 : 10/16  22.7%
12:06:17 : 10/17  31.5%
12:06:52 : 11/18  24.0%
12:07:24 : 11/19  32.4%
12:07:27 : Test finished.

----------
Total: 11/19 (32.4%)

Thanks for reading!

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #1 –
Did you try to use foo_bitcompare before wasting time on the ABX test?
"To understand me, you'll have to swallow a world." Or maybe your words.

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #2 –
No, I didn't try that.

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #3 –
I'll rerip now and compare those, can I post the results?

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #4 –
So these are the results of the bit compare...

INFO (foo_bitcompare) : Comparing:
INFO (foo_bitcompare) : location: "file://C:\blindtest\The Logical Song.wav" (0)
INFO (foo_bitcompare) : location: "file://C:\blindtest\wma\The Logical Song.wav" (0)
INFO (foo_bitcompare) : No differences in decoded data found.
INFO (foo_bitcompare) : Finished successfully.

Does this mean the wma file is identical to the wav file in every way?

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #5 –
Quote
Does this mean the wma file is identical to the wav file in every way?


[a href="index.php?act=findpost&pid=274588"][{POST_SNAPBACK}][/a]


Sure sounds like it. I've been messing with WMA Lossless lately, so I'll try it on a few files tonight.

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #6 –
Quote
Does this mean the wma file is identical to the wav file in every way?

Yes, and your ABX result was negative. There was a 1/128 (i.e. very possible) chance that you would get the first 7 trials right in such a test. However, subsequent trials revealed that it was no more than a lucky streak.

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #7 –
You have to agree to the max number of results in advance, or the ABX test is invalid.  If you just go until you're ahead and stop, you don't prove anything.

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #8 –
Quote
So these are the results of the bit compare...

INFO (foo_bitcompare) : Comparing:
INFO (foo_bitcompare) : location: "file://C:\blindtest\The Logical Song.wav" (0)
INFO (foo_bitcompare) : location: "file://C:\blindtest\wma\The Logical Song.wav" (0)
INFO (foo_bitcompare) : No differences in decoded data found.
INFO (foo_bitcompare) : Finished successfully.

Does this mean the wma file is identical to the wav file in every way?


[a href="index.php?act=findpost&pid=274588"][{POST_SNAPBACK}][/a]

Yes.
"To understand me, you'll have to swallow a world." Or maybe your words.

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #9 –
Thanks for your reply.

[NONSENSE] Blind Abx Test Of Wma Lossles...

Reply #10 –
Quote
But in the process I also believe that the sound of the lossless file is different to that of the extracted wav file in EAC or Sound Forge, so...


how did this thread not make it to the recycle bin yet?