Log in

View Full Version : AAC 5.1 vs WMV9 Pro 5.1


Pages : [1] 2

Sagittaire
14th April 2004, 20:00
KpeX edit: There were many misconceptions introduced in this thread. For an accurate summary and debunking of myths, please see the last two posts in this thread (http://forum.doom9.org/showthread.php?s=&postid=478160#post478160).

Edit: New test download & compare

[B]Source:
http://jfl1974.free.fr/Video/English.ac3

HE-AAC:
http://jfl1974.free.fr/Video/VBR-51-120-Tape.mp4 -> 1
http://jfl1974.free.fr/Video/VBR-51-153-Radio.mp4
http://jfl1974.free.fr/Video/VBR-51-175-Internet.mp4 ->2
http://jfl1974.free.fr/Video/VBR-51-232-Streaming.mp4 ->3

WMA9 Pro:
http://jfl1974.free.fr/Video/VBR-51-82-10.wma
http://jfl1974.free.fr/Video/VBR-51-109-25.wma
http://jfl1974.free.fr/Video/VBR-51-126-2pass.wma -> 1
http://jfl1974.free.fr/Video/VBR-51-185-50.wma -> 2
http://jfl1974.free.fr/Video/VBR-51-250-75.wma -> 3

HE-AAC encoding:
aac.dll 2.5.8.1 (Nero 6.3.0.0)
aacenc32.dll 2.6.1.1 (Nero 6.3.0.0)
besweet encoding



BeSweet v1.5b25 by DSPguru.
--------------------------
Using azid.dll v1.9 (b922) by Midas (midas@egon.gyaloglo.hu).
Using Shibatch.dll v0.24 by Naoki Shibata & DSPguru (shibatch.sourceforge.net).
Using bsn.dll v0.21 by DPeshev,Richard,DSPguru (http://DSPguru.Doom9.org).

Logging start : 04/15/04 , 17:02:32.

C:\Program Files\RMFactory\besweet\besweet.exe -core( -input G:\Mes dossiers\B.A\Les deux tours\English.ac3 -output G:\Mes dossiers\B.A\Les deux tours\English.mp4 -logfilea G:\Mes dossiers\B.A\Les deux tours\English.log ) -azid( -g max -n1 -L -3db -c normal ) -ota( -hybridgain ) -bsn( -6chnew )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : G:\Mes dossiers\B.A\Les deux tours\English.ac3
[00:00:00:000] | Output: G:\Mes dossiers\B.A\Les deux tours\English.mp4
[00:00:00:000] | Floating-Point Process: No
[00:00:00:000] +-------- AZID -------
[00:00:00:000] | Input Channels Mode: 3/2, Bitrate: 448kbps
[00:00:00:000] | Total Gain: 12.473dB, Compression: Normal
[00:00:00:000] | LFE levels: To LR -3.0dB, To LFE 0.0dB
[00:00:00:000] | Center mix level: BSI
[00:00:00:000] | Surround mix level: BSI
[00:00:00:000] | Dialog normalization: -4dB
[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +---------------------
[00:02:58:016] Conversion Completed !
[00:02:17:000] <-- Transcoding Duration

Logging ends : 04/15/04 , 17:04:49.



WMA9 Pro encoding:
AC3 -> 6 WAV with besweet and same setting that AC3 -> HE-AAC encoding
WMEncoder with 6 mono wav

bond
14th April 2004, 20:27
Originally posted by Sagittaire
Nero HE-AAC mode VBR quality "streaming : medium":
2.0 -> 1601 Ko
5.1 -> 5056 Ko

WMA9 Pro quality 50%:
2.0 -> 1554 Ko
5.1 -> 3737 Ko

AAC is optimized for 5.1 encoding ?lol, Sagittaire what are you trying again :D

if you want to compare two multichannel encoders and their quality you will have to do this at the same bitrate
comparing what output filesize two settings bring shows nothing ;)

note that just because a setting is labelled the same for multichannel and stereo, it doesnt mean that they are related somehow :rolleyes:
would it change anything for you if ahead would change the preset name to "multichannel : medium" for multichannel? names mean nothing...

to answer your question: yes, aac is optimized for multichannel, use the sbr option (called he-aac in nero) if you want maximum quality with aac and multichannel (still this doesnt mean of course that there isnt room for improvements)

Sagittaire
14th April 2004, 20:44
Make the test bond ... AAC isn't optimised for 5.1 ... Do you want I post examples ...

bond
14th April 2004, 20:48
if you say so :rolleyes:

when will you learn that the statements you make are totally useless, when will you learn to post what settings, what encoders you used aso...
ever heard about abx? at hydrogenaudio everyone will laught at you if you make quality claims without abx

i can imagine a marketing guy paid by m$ making such statements, but not someone serious

KpeX
14th April 2004, 22:01
Originally posted by Sagittaire
Make the test bond ... AAC isn't optimised for 5.1 ... Do you want I post examples ... I don't see a reason for this thread to remain open unless you clarify your statement very soon. Obviously two encoders are not going to output the same bitrate at the unrelated settings.

jimmy basushi
14th April 2004, 22:40
try rm 5.1 audio to see that they can come out fairly small. its quality is less then desired for me but.

whats the sound quality like on wmv 5.1 out of curiousity? nero aac streaming is very good...

Sagittaire
14th April 2004, 22:40
WMA9 Pro 5.1 48 KHz with WME9encodeur
HE-AAC 5.1 48 KHz with Nero Digital Audio Encodeur and graphedit
AC3 -> Normalisation and boost=4 with HeadAC3He -> Ogg 5.1 1500 Kbps or q10 ~LossLess -> WMA9 & HE-AAC

And exemple with 5.1 ...

Source AC3 5.1 448 Kbps
http://jfl1974.free.fr/Video/English.ac3

HE-AAC quality "Tape:lowest"
http://jfl1974.free.fr/Video/VBR-51-128.mp4

WMA9 pro 5.1 128 Kbps VBR 2 pass
http://jfl1974.free.fr/Video/VBR-51-128.wma

bond
15th April 2004, 08:33
Originally posted by Sagittaire
[B]WMA9 Pro 5.1 48 KHz with WME9encodeur
HE-AAC 5.1 48 KHz with Nero Digital Audio Encodeur and graphedit
AC3 -> Normalisation and boost=4 with HeadAC3He -> Ogg 5.1 1500 Kbps or q10 ~LossLess -> WMA9 & HE-AACwhy did you use ogg vorbis in between?
its absolutely not necessary (ar least with aac) and will lead to worse quality
also the preprocessing could bias the comparison, can you plz upload two samples, which are directly ac3->wma9pro (as you surely know 5.1 isnt possible with normal wma9) and he-aac

maybe i should try to convince rjamorim to conduct a 5.1 listening test to clarify this :)

bond
15th April 2004, 09:52
Sagittaire,
i dont know what you are trying here, but i now encoded your .ac3 myself in besweet ("the normal way": hybridgain, no other preprocessing, also with he-aac and tape:lowest ) and the output file has clearly higher quality than the sample you posted

of course i dont assume that you wanted to bias the test, but i think you maybe simply used outdated nero encoder dlls, used the wrong channel mapping, maybe your strange preprocessing lead to your output file, who knows...

all in all grap my aac sample here (http://8ung.at/bond/sample.zip)

Sagittaire
15th April 2004, 10:41
I use for this test the latest Nero Digital Encoder with graphedit (encodeur use by Nero Recode for AAC encoding) ...

http://jfl1974.free.fr/Video/Nero.jpg

And i use Ogg 5.1 q10 ~LossLess for have exactly the same source in entry ...

bond
15th April 2004, 10:48
Originally posted by Sagittaire
And i use Ogg 5.1 q10 ~LossLess for have exactly the same source in entry ...you dont understand
this ogg step in between
1) leads to potential loss of quality (yeah even if its at -q10 info is lost)
2) who knows what channel mapping ogg vorbis uses and if it can be used correctly with the nero encoder (if the channels are mixed up somehow during encoding, this will surely lead to low quality
3) who knows if 5.1 in ogg vorbis is stable, who knows what hidden bugs might be there, what not known problems (as good as noone uses vorbis for 5.1 atm) could occur...

the point is there is no reason to use ogg vorbis in between, so why risk anything
or asking more specifically: why do you insist on using ogg vorbis in between?

SeeMoreDigital
15th April 2004, 11:01
Yes and here is my test sample (http://homepage.ntlworld.com/seemoredigital/AAC-HE_6Ch@128kbps_VBR_24kHz_(SMD).zip), generated using the newest version of Foobar2000 and Nero with no filters etc!

It should sound very much like bonds except my encode has a sample rate of 24,000Hz and bonds is 22,050Hz (and it does not display any title information - I really must learn how to do that)

Anyway, if you used HeadAC3He anywhere in the mix, in my opinion this is not a very good move, as I don't think it maps the 6Ch wav output correctly and chops of all the low frequencies...

I agree with bond. Don't make 'rash' statements without asking for a little help and advice first. Just because you've managed to generate a poor sounding AAC encode, does not instantly mean WMA is better!

Cheers

EDIT: My NeroDigital GUI looks like this: -

http://82.2.167.24/Uploaded_Files/Doom9_Forum_files/NeroDigital_GUI.gif

I see that the Nero GUI look very different in GraphEdit!

Tuning
15th April 2004, 18:10
Sagittaire,
Plz use foobar or Besweet for AAC encoding. Use same bitrate for both AAC and WMA9Pro, otherwise this comparison is of no use.


PS: Thread title should be changed to AAC 5.1 vs WMA9 Pro 5.1 (Now it is WMV..)

Sagittaire
15th April 2004, 21:19
New Test: download and compare

Source:
http://jfl1974.free.fr/Video/English.ac3

HE-AAC:
http://jfl1974.free.fr/Video/VBR-51-120-Tape.mp4 -> 1
http://jfl1974.free.fr/Video/VBR-51-153-Radio.mp4
http://jfl1974.free.fr/Video/VBR-51-175-Internet.mp4 ->2
http://jfl1974.free.fr/Video/VBR-51-232-Streaming.mp4 ->3

WMA9 Pro:
http://jfl1974.free.fr/Video/VBR-51-82-10.wma
http://jfl1974.free.fr/Video/VBR-51-109-25.wma
http://jfl1974.free.fr/Video/VBR-51-126-2pass.wma -> 1
http://jfl1974.free.fr/Video/VBR-51-185-50.wma -> 2
http://jfl1974.free.fr/Video/VBR-51-250-75.wma -> 3

HE-AAC encoding:
aac.dll 2.5.8.1 (Nero 6.3.0.0)
aacenc32.dll 2.6.1.1 (Nero 6.3.0.0)
besweet encoding



BeSweet v1.5b25 by DSPguru.
--------------------------
Using azid.dll v1.9 (b922) by Midas (midas@egon.gyaloglo.hu).
Using Shibatch.dll v0.24 by Naoki Shibata & DSPguru (shibatch.sourceforge.net).
Using bsn.dll v0.21 by DPeshev,Richard,DSPguru (http://DSPguru.Doom9.org).

Logging start : 04/15/04 , 17:02:32.

C:\Program Files\RMFactory\besweet\besweet.exe -core( -input G:\Mes dossiers\B.A\Les deux tours\English.ac3 -output G:\Mes dossiers\B.A\Les deux tours\English.mp4 -logfilea G:\Mes dossiers\B.A\Les deux tours\English.log ) -azid( -g max -n1 -L -3db -c normal ) -ota( -hybridgain ) -bsn( -6chnew )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : G:\Mes dossiers\B.A\Les deux tours\English.ac3
[00:00:00:000] | Output: G:\Mes dossiers\B.A\Les deux tours\English.mp4
[00:00:00:000] | Floating-Point Process: No
[00:00:00:000] +-------- AZID -------
[00:00:00:000] | Input Channels Mode: 3/2, Bitrate: 448kbps
[00:00:00:000] | Total Gain: 12.473dB, Compression: Normal
[00:00:00:000] | LFE levels: To LR -3.0dB, To LFE 0.0dB
[00:00:00:000] | Center mix level: BSI
[00:00:00:000] | Surround mix level: BSI
[00:00:00:000] | Dialog normalization: -4dB
[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] | Source Sample-Rate: 48.0KHz
[00:00:00:000] +---------------------
[00:02:58:016] Conversion Completed !
[00:02:17:000] <-- Transcoding Duration

Logging ends : 04/15/04 , 17:04:49.



WMA9 Pro encoding:
AC3 -> 6 WAV with besweet and same setting that AC3 -> HE-AAC encoding
WMEncoder with 6 mono wav

bond
15th April 2004, 21:33
thanks for the samples!

Originally posted by Sagittaire
aac.dll 2.5.8.1 (Nero 6.3.0.0)
aacenc32.dll 2.6.1.1 (Nero 6.3.0.0)latest version is 6.3.1.6, why dont you use it?
in contrary to some other codec providers ahead is always working on improvements of their aac encoder and also releases them every month

-azid( -g max -n1 -L -3db -c normal )these preprocessing isnt really necessary (also because you use hybridgain already)
using -3db is not good as afaik this is only needed for stereo output (defines how the lfe channel is mixed into l/f)

[00:00:00:000] | Rear channels filtering: No
[00:00:00:000] | Source Sample-Rate: 48.0KHzthe newest aac encoder dlls resample to 44.1khz output normally, cause 48 doesnt seem to be necessary anyways at such low bitrates...

Sagittaire
15th April 2004, 21:48
aac.dll and aacenc32.dll version with Nero 6.3.1.6 ... ?

bond
15th April 2004, 22:06
Originally posted by Sagittaire
aac.dll and aacenc32.dll version with Nero 6.3.1.6 ... ? hm latest i downloaded the latest recode and it comes with an aacenc32.dll version 2.6.2.0

anways it always makes sense to use the latest dlls coming with the latest nero burning rom

Sagittaire
15th April 2004, 22:51
It's good besweet profil ... ?

-core( -input "G:\Mes dossiers\B.A\Les deux tours\English.ac3" -output "G:\Mes dossiers\B.A\Les deux tours\English.mp4" -logfilea "G:\Mes dossiers\B.A\Les deux tours\English.log" ) -ota( -hybridgain ) -bsn( -6chnew ) -ssrc( --rate 44100 )

Sagittaire
16th April 2004, 00:00
That doesn't change: The WMA9 pro 5.1 is always better than AAC 5.1 with 44.1 Khz or 48 Khz ... and especially for the very low bitrate ~128 Kbps

jk888
16th April 2004, 15:53
Hm... The DirectSound filter reports that the AAC sample has 6 channel but why does it report that the wma has only 2 channels? Is it that the DirectSound filter can't understand the new WM9 codec? Or something went wrong with his encode?

The wma encode is really really quiet, I can hardly hear it even with my volumn at peak. The AAC sample sounds like crap on the tape profile encode, the sound just isn't sharp and there's a distinct hushing sound when people speak. All my encodes were done with the tape profile, 5.1 channel AAC HE, and they sound really good, almost like the original DVD audio.

Here's what I do (with the latest nero dlls and Besweet 1.5b26):
"F:\BeSweet\BeSweet.exe" -core( -input "f:\DVD2\test.ac3" -output "f:\DVD2\test.mp4" -logfilea "F:\BeSweet\BeSweet.log" ) -azid( ) -bsn( -6chnew -config ) -ssrc( --rate 44100 )

SeeMoreDigital
16th April 2004, 16:14
Hmm...

Correct me if I'm wrong but when I generate 6Ch AAC-HE VBR (or LC) stream it's only done in one pass. Sure it's VBR but still only one pass was used.

There's no doubt that the 6Ch WMA encode sounds very nice. However, it was generated using two full passes. This has got to make a big difference!

Cheers

Sagittaire
16th April 2004, 17:49
@ jk888

Install WM9 filter for WMP 6.4 or WMP9

@ SeeMoreDigital

2 pass for predictibility ... not for quality: encoding with 82, 109, 185 and 250 Kbps in quality 1 pass mode.

Everyone says ACC is best audio codec. Everyone says XviD is best video codec ... that doesn't seem to be really the truth ...

SeeMoreDigital
16th April 2004, 18:15
Originally posted by Sagittaire
@SeeMoreDigital

2 pass for predictability ... not for quality: encoding with 82, 109, 185 and 250 Kbps in quality 1 pass mode. Pardon!

It would seem fairly obvious to me that any codec (whether it be an video or audio codec) will generate a more accurate encode when an full 'analysing/predictability' pass is done prior to an 'writing/encoding' pass!

If you are able, why not generate another 6Ch WMA9 encode in 1pass VBR (like AAC does) - that is, if it's possible to do with WME9... I can't remember. And I don't have it installed on this PC!

Just my 2 cents but valid I think!

Cheers

Tuning
16th April 2004, 18:25
Sagittaire,

If WMA9 Pro is done in 1-pass, which will be the fastest encoding? AAC or WMA9 Pro? Curious to know the result.:) Thanks.

jk888
16th April 2004, 19:56
Originally posted by Sagittaire
@ jk888

Install WM9 filter for WMP 6.4 or WMP9


I have WMP9 installed already, that plays the file fine, but WMP6.4 doesn't play it at all. I used Media Player Classic and it clearly shows these wma files are 2 channel only.

VBR-51-126-2pass.wma
http://randy.abstractstudios.ca/JK888/temp/Capture02.jpg

VBR-51-120-Tape.mp4
http://randy.abstractstudios.ca/JK888/temp/Capture01.jpg

Sagittaire
16th April 2004, 21:50
For WMP 6.4, MPC and other Player
http://www.microsoft.com/windows/windowsmedia/format/codecdownload.aspx

Or Sazuki Pack Codec
http://satsuki.yatoshi.free.fr/sysofts/satsuki.decoder.pack.1.0.3.0.exe

SeeMoreDigital
16th April 2004, 23:14
Originally posted by Tuning
Sagittaire,

If WMA9 Pro is done in 1-pass, which will be the fastest encoding? AAC or WMA9 Pro? Curious to know the result.:) Thanks. Yep, it would be VERY interesting ;)

Originally posted by jk888
I have WMP9 installed already, that plays the file fine, but WMP6.4 doesn't play it at all. I used Media Player Classic and it clearly shows these wma files are 2 channel only. It sounds like (no pun intended) you need to re-configure your players output settings.

Sometimes they can reset themselves. Which is very annoying!

Cheers

Sagittaire
17th April 2004, 00:20
AC3 5.1 Duration: 2 min 58
Celeron II 1470 Mhz 384 Mo

AC3 -> HE-AAC 5.1
With besweet and no azid pre-pass

for HE-AAC profil (tape, radio, internet, streaming)
Encoding duration ~2 min 20


AC3 -> WMA9 Pro 5.1
With WMEncodeur

for 1 pass quality profil (10%, 25%, 50% ....)
Encoding duration ~1 min 10

for 2 pass bitrate profil
Encoding duration ~2 min 25

Tuning
17th April 2004, 03:13
Is original ac3 is at 48Khz, then resampling takes some time.(in AAC conversion)

jk888
17th April 2004, 04:02
Originally posted by Sagittaire
For WMP 6.4, MPC and other Player
http://www.microsoft.com/windows/windowsmedia/format/codecdownload.aspx


I installed all 3 codecs from the MS site and still WMP 6.4 can't play it. Only WMP 9 can play it. MPC still reports 2 channel onlys. I'm using W2k for OS and I don't think that affects anything. So can anyone else play this on WMP 6.4?

Originally posted by SeeMoreDigital

It sounds like (no pun intended) you need to re-configure your players output settings.

Sometimes they can reset themselves. Which is very annoying!

Not an output settings problem, because the error reported by WMP 6.4 is "An Invalid Media Type Was Specified". Also I checked my output settings, they are on the Default DirectSound, so I tried the other ones SB Live and Wave mapper but exactly the same results.
-------------------
I used VideoLAN to test the wma file and it doesn't play it. So this is mostly likely not compatible with Linux. Strangely it reports that the AAC encode is 2 channel while the unplayable wma is 6 channel, this is probably a bug with the player.

SeeMoreDigital
17th April 2004, 12:45
Originally posted by Sagittaire
AC3 5.1 Duration: 2 min 58
Celeron II 1470 Mhz 384 Mo

AC3 -> HE-AAC 5.1
With besweet and no azid pre-pass

for HE-AAC profil (tape, radio, internet, streaming)
Encoding duration ~2 min 20


AC3 -> WMA9 Pro 5.1
With WMEncodeur

for 1 pass quality profil (10%, 25%, 50% ....)
Encoding duration ~1 min 10

for 2 pass bitrate profil
Encoding duration ~2 min 25 I'm not quite sure what you are trying to say here!

Please correct me if I'm wrong but....

When generating an 6Ch AAC file with BeSweet (or Foobar2000), you're able to convert from AC3 directly to AAC in one application. However, the process will go something like this: -
6Ch AC3-->6Ch WAV-->6Ch AAC
ie: a three step process with one application.

When generating an 6Ch WMA file with WME9. It's obviously not as simple as the above AC3-->AAC process. Because you first have to create suitable WAV streams, prior to feeding them into WME9. So the process goes something like this: -
6Ch AC3-->6Ch WAV or 6No mono WAV's
6Ch WAV or 6No mono WAV's-->6Ch WMA
ie: a three step process with two applications.

Please confirm what you are suggesting?

Cheers

Sagittaire
17th April 2004, 13:57
AC3 -> AAC with besweet without azid prepass gain ~ 2 min 20 (+ 10 s with azid prepass)
AC3 -> WMA9 with WMEncodeur 1 pass quality ~ 1 min 10
6 WAV -> WMA9 with WMEncodeur 1 pass quality ~ 1 min 15

WMEncoder work directly with ac3 files or multiple mono wav ...

SeeMoreDigital
17th April 2004, 16:12
Originally posted by Sagittaire
WMEncoder work directly with ac3 files or multiple mono wav ... Sagittaire, I sincerely apologise. I was aware that WME9 could do this!

Has it always been able to do this. As I'm sure I used to have to piss about making 6No mono wavs at one time or another?

That said, my main point still stands. The 6Ch WMA file you posted has been generated using 2pass VBR. But the 6Ch AAC-HE encodes bond and I posted were generated using 1pass VBR.

Cheers and my apologies again!

SeeMoreDigital
18th April 2004, 12:36
Originally posted by hans-jürgen
...Including 5.1 AAC Low Complexity files in any multichannel transcoding test makes sense, because the first mass-production 5.1 hardware decoders (e.g. in DVD players) will very likely be able to decode only this profile, not HE AAC. Hmm. I don't know about this.

However, unless you have a Digital Surround Sound (DSS) amplifier which can decode AAC (and there are currently only a couple of them), owning an DVD player (or Mpeg4/DVD player) that can puke out 6Ch AAC streams is pretty pointless.

There is of course the prospect that the 'new players' could include an 'on-board' AAC decoder, but this would mean the audio signals would have to be sent out of the player in the 'analogue' format!

Cheers

Sagittaire
18th April 2004, 13:34
For the WMA9 5.1 it is only the version pro of the WMA9. Between WMA9 pro and LC-AAC there is no possible comparaison with low bitrate for 5.1 and even for 2.0: WMA9 pro is really better than LC-AAC ...

SeeMoreDigital
18th April 2004, 14:08
Originally posted by Sagittaire
For the WMA9 5.1 it is only the version pro of the WMA9. Between WMA9 pro and LC-AAC there is no possible comparaison with low bitrate for 5.1 and even for 2.0: WMA9 pro is really better than LC-AAC ... Yep, I would agree with you there Sagittaire. I generated some 6Ch AAC-LC encodes and they were not brilliant!

That said, I've just generated some 6Ch WMA 1pass VBR(Q25) encodes (at 16 and 24bit) and to my ears they don't sound as good as 6Ch AAC-HE 1pass (Tape::Lowest) VBR.

So it's back to you.

Cheers

Tuning
18th April 2004, 14:11
Can we mux WMA9 Pro with any of container formats? :) (Except WMV )encodes (at 16 and 24bit) and to my ears they don't sound as good as 6Ch AAC-HE 1pass (Tape::Lowest) VBR.

SMD, is it that bad? I mean Is AAC/WMA9 Pro not comparable..?

SeeMoreDigital
18th April 2004, 14:29
Originally posted by Tuning
Can we mux WMA9 Pro with any of container formats? :) (Except WMV )Matroska probably. The guys keep telling us it can contain anything :D

Originally posted by Tuning
SMD, is it that bad? I mean Is AAC/WMA9 Pro not comparable..? No, don't get me wrong. 6Ch WMA 1pass VBR still sounds very good indeed. But if we are to be as 'nit-picky' as Sagittaire is leading us all to be. To my ears it's not 'quite' as good!

If you hang on a while I'll zip the two I did up, and upload them to my 'speedy server' for download (does that make sense?)...

Cheers

EDIT: Files uploaded (http://homepage.ntlworld.com/seemoredigital/WMA_6Ch@Q25_1pass_VBR_(SMD).zip). You will find that the front channel steering is not as accurate for starters!

Tuning
18th April 2004, 14:57
@SMD, Currently downloading...approx 20 min..:o (64kbps connection)
Thanks for the files.
:)

@All, can any one tell the channel order of WMA9 Pro ?:)


PS: Now I think Sagittaire's comments lead us to bond's signature ...:D :D


$EDIT$
24bit files seems good.:)

btw, I'm currently out of multichannel equipment: only having 2.1 system now for power saving :D, so i cannot test the 5.1 now...may be tomorrow.. ;)

SeeMoreDigital
18th April 2004, 15:39
Originally posted by Tuning
@All, can any one tell the channel order of WMA9 Pro ?:) It would appear that WME9 correctly steers the audio channels from an 6Ch AC3 source.

Before I began encoding with WME9 I generated the following Channel Check Speaker Test (http://82.2.167.24/Uploaded_Files/Doom9_Forum_files/WMA_6Ch_Speaker_Test.zip).

Cheers

Tuning
18th April 2004, 15:45
Originally posted by SeeMoreDigital
It would appear that WME9 correctly steers the audio channels from an 6Ch AC3 source.

Then WMA9 Pro is of same channel order of ac3.

Concerning size and quality obtained, WMA9Pro 5.1 is good for multichannel. As encoding can be done in 2-pass the movie sound may be well compressed.

PS: Time to follow bond's signature... :D

SeeMoreDigital
18th April 2004, 16:33
Originally posted by Tuning
Then WMA9 Pro is of same channel order of ac3.

Concerning size and quality obtained, WMA9Pro 5.1 is good for multichannel. As encoding can be done in 2-pass the movie sound may be well compressed.

PS: Time to follow bond's signature... :D Yep, WMA9 maps in the same channel order as AC3.

With regard to 2pass WMA VBR encoding I have to say it does sound very good indeed. Dare I say it, better even than 1pass AAC-HE VBR. But that's the main benefit of 2pass encoding!

I have read bonds post and I'm in total agreement with it. That said, part of accessing any given codec's merits is to try and remain impartial. Which does not mean automatically hating everything Micro$oft produces... We just have to make sure that we (here on the forum) all work together to help make open-source products better than anything M$ can offer.

Cheers

hans-jürgen
18th April 2004, 16:36
Originally posted by Sagittaire
For the WMA9 5.1 it is only the version pro of the WMA9. Between WMA9 pro and LC-AAC there is no possible comparaison with low bitrate for 5.1 and even for 2.0: WMA9 pro is really better than LC-AAC ... When I look at your first posting in this thread, you seem to have tested the streaming profile of Nero 6 which resulted in 232 kbps/6ch and WMA9 Pro at VBR quality 75% which resulted in 250 kbps/6ch. Or do I interpret your file names in the wrong way?

Anyhow, these bitrates are also possible with the Low Complexity Profile of AAC (a stereo bitrate of 90-100 kbps VBR is considered to be the "sweet spot" of AAC LC), that's why I asked if someone could test FAAC at a VBR quality of 75% and a cutoff at 13 kHz with your original sample. Now I'm downloading that 9.5 MB AC-3 file and would like to know if this has been already transcoded to Vorbis in some way or really is the direct AC-3 track from the DVD?

avih
18th April 2004, 18:24
Originally posted by Sagittaire
For the WMA9 5.1 it is only the version pro of the WMA9. Between WMA9 pro and LC-AAC there is no possible comparaison with low bitrate for 5.1 and even for 2.0: WMA9 pro is really better than LC-AAC ...

Dear Sagittaire. I've asked you not to repeat such claims without proofs.

rjamorim made an independant listening test with 12 broad samples, every sample listened to by 15-20 people and the overall result was this:
http://www.rjamorim.com/test/128extension/plot12z.png
which clearly shows that lc-aac is on par qualitywise with wma9 pro.

wma9 pro isnt usable for bitrates lower than 128kbps. it simply doesnt offer these bitrates, thats why its not possible to compare it on lower bitrates with other formats. however, you make it sound as IF it's not comparable since it's obvious than WMA9 pro is just superiour.

for repeating such claims eventhough i asked you to avoid it, i strike you now for rule 4 (respect the moderators). it's not the respect i'm after though. it's the truth. and spreading FUD just doesn't make me happy.

now, after we've put this behind us, can i please ask you to start acting like a pro? again, many of your claims are VERY non-specific, fuzzy, and might mislead less knowledgable members. i know you have the deep technical knowledge and enthusiasm to be a really valuable member of our community. will you please try to be specific and professional? help users, back your claims, use various samples when comparing quality and not make it look as if you're blindly promoting WMx?

thx for the understanding.
avih

ps.
modifying many of your posts as in the modification of my quote of your original post toThat doesn't change: The WMA9 pro 5.1 is always better than AAC 5.1 with 44.1 Khz or 48 Khz ... and especially for the very low bitrate ~128 Kbpsdoesn't help either. don't make broad claims. be specific. back up your claims. be a professional.


my apologies. the above mentioned post was not modified. these are 2 different posts, and they are both still there. sorry. my points remain though.

Sagittaire
19th April 2004, 13:57
wma9 pro isnt usable for bitrates lower than 128kbps. it simply doesnt offer these bitrates, thats why its not possible to compare it on lower bitrates with other formats. however, you make it sound as IF it's not comparable since it's obvious than WMA9 pro is just superiour.

WMA9 pro 2.0 Quality 50% for ~64 Kbps
For me low bitrate for 2.0 is ~64Kbps
For me low bitrate for 5.1 is ~128Kbps

WMA9 pro is very better than WMA9 sdt and WMA9 sdt is better than best LC-AAC encoder (quick time) for low bitrate:
http://www.rjamorim.com/test/64test/results.html

no possible comparaison between WMA9 pro and LC-AAC for low bitrate ... make the test for 5.1 with 128 or 192 Kbps for exemple.

For my sample (LOTRII) WMA9 pro is better than HE-AAC for me. Perhaps for other sample HE-AAC will be better then WMA9 pro. But when some say "HE-AAC is the best audio codec and by far" once again it isn't the reality ...

SeeMoreDigital
19th April 2004, 14:23
Sagittaire,

Why are you doing this? You must know that when avih wrote the statement below, he was refering to 6Ch audio (not 2Ch audio): -

wma9 pro isnt usable for bitrates lower than 128kbps. it simply doesnt offer these bitrates, thats why its not possible to compare it on lower bitrates with other formats. however, you make it sound as IF it's not comparable since it's obvious than WMA9 pro is just superiour.

I'm completely bemused why you started comparing the quality of 2Ch audio in your 6Ch audio thread!

You've made some good points, some of which I support. But even I'm starting to get fed up with this thread.... may be it should be closed before it starts to confuse our newbie members.

Cheers

tigre
19th April 2004, 14:28
Originally posted by Sagittaire
WMA9 sdt is better than best LC-AAC encoder (quick time) for low bitrate:
http://www.rjamorim.com/test/64test/results.html
That interpretation is wrong. The error bars overlap which means that it's not possible to conclude from the test which one is better or worse. So they must be regarded equal for now.

Sagittaire
19th April 2004, 14:32
Correction ... ;-)

WMA9 pro is very better than WMA9 sdt and WMA9 sdt is comparable with best LC-AAC encoder (quick time) for low bitrate ...

SeeMoreDigital
19th April 2004, 16:21
Why is every suddenly talking about 2Ch comp tests in this 6Ch thread... they are a complete irrelevance and totally useless to the subject at hand!

Teegedeck
19th April 2004, 16:32
...not to mention that it is illegitimate comparing a codec that did not participate in the test with the worse of two participating AAC versions (at that bitrate). That argumentation is just a total mess.

Maybe WMA Pro is better than AAC, who knows. But that would have to proven in a listening-test, not by speculations.

Maybe WMA Pro is better than HE AAC at 5.1 encoding. But that would have to be proven in a listening-test as well.

Without one we'd just have to believe you. And this, forgive me, is asked a bit much considering what you have shown so far.