NEW CODEC! 128 kbps and less for 5.1 surround encoding for MPEG4 (Xvid/Divx) movies [Archive] - Page 4

Cyberace

5th February 2006, 15:03

Cool stuff, but why not release the full source code and license it under standard GPL or LGPL?, (if not the full codec with the encoder then how about just the decoder), that way open sourced players that are licensed under GPL/LGPL can also incooperate the decoder to be able to play Aud-X encoded audio files and soundtracks, ...example of such players are MPlayer, XINE, VideoLan-Client (VLC), and the FFmpeg/FFdshow codec-suit.

3dsnar

5th February 2006, 18:12

I don't know about the source code, but I know that Aud-X DLL
encoder and decoder will be released probably before the end of this month
(for incorporating it in third party software). And ofcourse it will be free.
-----------------------------------------------------------------------
Maybe Aud-X owners will decide to release the source code as well...
I do not know...

3dsnar

5th February 2006, 18:16

Perhaps that VBR quality mode could be better ... if possible

VBR mode with lame
Quality mode VN -> base mp3 is VN lame join stereo (VN too for spatial info)
and
Quality mode VN -> base mp3 is VN lame mono (VN too for spatial in
Thanx for suggestions.
Yeah. But we would rather focus on extending the spatial information processing
with mpeg-4 audio, or ogg vorbis. But this is for the future, which will depend on the performance of the CBR mp3 based release. CBR was chosen for one particular reason. AVI (crap) compatibility ;)

Gabriel_Bouvigne

6th February 2006, 09:52

The core bitrates that we use
1) STRQ 80 kbps -> base mp3 is 56 kbps mono (24 kbps for spatial info)
2) STDQ 128 kbps -> base mp3 is 64 kbps mono (64 kbps for spatial info)
3) HGHQ 192 kbps -> base mp3 is 128 kbps stereo (64 kbps for spatial info)
4) SPBQ 192 kbps -> base mp3 is mono 64 kbps (128 kbps for spatial info)
OMG! Up to 128kbps of parametric information!
Isn't it a little too much?
SPBQ is especially odd to me, as the mp3 base layer is only 64kbps for a total of 192kbps for a 5.1 reconstructed stream.
Mp3@64kbps mono should provide comparable quality to mp3@96kbps stereo. If you give the same bitrate to front and rear (which would be a little odd), it means that @192kbps you could have a 4.0 stream with the same quality as the SPBQ base channel, but without any parametric info.
If you reduce a little the bitrate of the rear channels down to 80kbps, we would have a 176kbps stream for 4 full channels. Don't you think that the results could be better by adding 16kbps of parametric info to a 176kbps 4 full channels encode than by using 128kbps of parametric info to a single 64kbps channel?

3dsnar

6th February 2006, 10:12

Gabriel, thanx for ideas and suggestions.
This may seem a little odd, but reconstructing the total panorama
is a balance between the perceptual encoder settings, and ability to reconstruct phase in all channels.

Therefore so much bitrate has to be used for the encoding the spatial information.
(this is also the major difference between our approach, and the one applied in mp3surround
and AAC 5.1).

Ofcourse everything is a matter of subjective taste. After hundreds hours of listening tests (all in 5.1 speakers set in a studio) the balance between
the main audio stream quality and spatial perception was established.
Such parameters tuning are very, very time consuming and decision is also not easy. We have tested several configurations on approx. 30 people of various
listening experiance, and based on their feelings the decision was made.

Probably the listeners preferences (resulted in high bitrate for spatial info),
were probably caused by the fact, that the chosen test material was only movie soundtracks
with very reach 5.1 panorama.
And creatomg codec tuned for movies was our primary target.

---

But you have a point here. A ~160 kbps bitrate would fill a gap between
128 and 192 kbps. And therefore might be useful.
This is something for the future, which we will definitely consider.

elenhil

7th February 2006, 14:23

I tested it again.;)
Uncheck the mixer itself, where is only(I believe) place to downmix the decoded(upmixed) multichannel by 'dolby decode'.
It has never been checked in my configuration. Ever.

elenhil

7th February 2006, 14:24

3dsnar, will the new Aud-X DirectShow filter support other formats besides mp3?

3dsnar

7th February 2006, 15:06

In fact our filter supported also AC3, but I have disabled this.
Also, the mp3 compatibility is limited, to allow for FFDshow and AC3filter
usage with our DSfilter.

Very soon, we want to add a control panel of our DSfilter,
which will enable to choose formats decoded by the filter,
and it will also allow for disabling the direct show module.

I hope to have this ready in February.

But maybe you have some specific problem
and maybe I will be able to sort it out?
Please let me know

elenhil

7th February 2006, 17:42

Very soon, we want to add a control panel of our DSfilter,
which will enable to choose formats decoded by the filter,

The question is whether it will support OGG Vorbis (and thus decode a DPL-encoded ogg file)?

3dsnar

7th February 2006, 19:33

Why not.
So far, we have implemented mp3 and AC3 decoding,
so decoging Ogg would require a bit of extra work.
Therefore I cannot specify when exactly Ogg compatible
direct show could be released, but it will be considered for 100%.

Rockaria

7th February 2006, 19:38

It has never been checked in my configuration. Ever.OK... it seems not be either my or FFDShow's fault.
The 'dolby decoder' does the DPL II decoding(to 6ch) no problem here(but with somewhat too much center).
Anyway, read the Audio-FAQ which is the best help you can find here.

What I can do is just helping each other if we can and want.. voluntary.
So if you don't call me again (ever), this thread is none of my interest.

3dsnar

14th February 2006, 13:18

We have just lounched Aud-X forum
www.aud-x.com/forum

So everyone who has qustions, suggestions,
requests, bug reports, etc. related to the project

is welcome to share her/his opinion :)

sigma_sigma

20th February 2006, 21:40

very good project. generally a movie DVDRIP on 2 cd's with ac3 sound (6ch) has a good video quality but with your codec, 5.1 sound will be smaller and video percent for video will be bigger and this results a better quality.
Greta job|!

3dsnar

21st February 2006, 18:10

sigma_sigma,
thanx for interest in our project.

Please note,
you may find very useful and user friendly guides
for encoding, decoding and transcoding on our web site:
http://www.aud-x.com/images/stories/downloads/aud-x_dvd_backup_guide_(full).pdf
and also
http://www.aud-x.com/index.php?option=com_remository&Itemid=34&func=download&filecatid=2
http://www.aud-x.com/index.php?option=com_remository&Itemid=34&func=download&filecatid=10

3dsnar

10th May 2006, 11:27

Short info:
Aud-X MP3 5.1 support added to FFDshow
http://bittekeinspam.googlepages.com/
:thanks:

shon3i

10th May 2006, 11:59

Nice, Thank You

buzzqw

10th May 2006, 16:35

@3dsnar

Could be possible to implement stdin support ?

this is very useful with bepipe/behappy encoding

thanks

BHH

3dsnar

10th May 2006, 16:41

@3dsnar

Could be possible to implement stdin support ?

this is very useful with bepipe/behappy encoding

thanks

BHH
This is already done... :)
By Dimzon. Please check this out:
http://www.gotdotnet.com/workspaces/workspace.aspx?id=1bb59ddf-901b-43a5-bd54-b0999e8e223e

buzzqw

10th May 2006, 17:44

:) :thanks:

i missed the link !

BHH

3dsnar

11th May 2006, 15:09

I have requested some interface changes regarding Aud-X usage
in FFDshow.
I think this should be more straight forward and stable.
Please take a look at this post
and share your thoughts.
http://forum.doom9.org/showthread.php?t=98600&page=60
:thanks:

raquete

12th May 2006, 04:05

hy all!
:stupid: but i:readguid: and can't find answers!
excuse my ignorante,newby and maybe stup question:
can i encode and play mp3 5.1(my preference is vbr) in my standalone without video?
i only want the sound.

thanks !
:o

3dsnar

12th May 2006, 05:46

You can encode sound with our HGHQ encoding option
(192 kbps). This is Dolby Pro Logic II compatible, so you will
get some sort of surround sound on standalone DVD (assuming the DPLII decoder will be deployed).
When decoding this sound on PC, Aud-X auxiliary information contained in the stream will be used for upmixing to 5.1
(this results in much better, than DPLII quality).
So this approach is doulbe-commpatible (Aud-X MP3 5.1 and DPLII)
--
Alternatively, you can create a DPLII downmix and encode it as normal mp3 VBR, and also you will be able to get a sort of surround on standalone DVD player.
But it will not be compatible with Aud-X.

raquete

12th May 2006, 11:43

wonderful and clear explanations 3d

thank you so much!
:)

TFM_TheMask

12th May 2006, 22:25

How do I know if the avi that I have contains Aud-X audio or just normal MP3 audio?? Is there a way to check this.

When I open an Aud-X file it always shows MP3 audio with only 1 channel (mono).

Pomyk

13th May 2006, 01:19

Please fix the bugs first.. :|

3dsnar

13th May 2006, 07:30

How do I know if the avi that I have contains Aud-X audio or just normal MP3 audio?? Is there a way to check this.
When I open an Aud-X file it always shows MP3 audio with only 1 channel (mono).
1) You have to use a direct show compatible player (e.g. MediaPlayerClassic, BSplayer etc. ).
2) Than, make sure that in CONFIG tab
[start->programs->Aud-X->Aud-X.exe] of our decoder
Aud-X/MP3 output is enabled.
3) Play the avi, and Aud-X icon shoud appear in the tray. Double click on it, and you will acces direct show filter window.
4) at the bottom of the DSfilter window it should be indicated what type of stream is currently decoded. This appears only during playback.
---------------------------------
In FFDshow it is not possible to determine that (no appropriate description is displayed).
Please request this feature through SourceForge:
http://sourceforge.net/tracker/index.php?func=detail&aid=1485338&group_id=53761&atid=471492
Cheers, 3d

3dsnar

13th May 2006, 07:32

Please fix the bugs first.. :|
What bugs do you have in mind?
Actually the latest versions are very stable.
However, if there are any problems that you would like to report,
please follow this link:
http://www.aud-x.com/forum/viewforum.php?f=2
:thanks:

TFM_TheMask

13th May 2006, 07:51

1) You have to use a direct show compatible player (e.g. MediaPlayerClassic, BSplayer etc. ).
2) Than, make sure that in CONFIG tab
[start->programs->Aud-X->Aud-X.exe] of our decoder
Aud-X/MP3 output is enabled.
3) Play the avi, and Aud-X icon shoud appear in the tray. Double click on it, and you will acces direct show filter window.
4) at the bottom of the DSfilter window it should be indicated what type of stream is currently decoded. This appears only during playback.
---------------------------------
In FFDshow it is not possible to determine that (no appropriate description is displayed).
Please request this feature through SourceForge:
http://sourceforge.net/tracker/index.php?func=detail&aid=1485338&group_id=53761&atid=471492
Cheers, 3d

Can't you build in something that it can be detected before playing, like with mediainfo or with AviInfo.

3dsnar

13th May 2006, 07:56

Well, it is backwards mp3 compatible, therefore it appears as a regular MP3. However there is a 'Aud-X1.0' string in the audio stream.
Based on this an application can recognize that it is dealing with Aud-X.
--
When you use our DLL, it returns an appropriate information,
indicating whether this is or isn't the Aud-X stream.
You can download the SDK here:
http://www.aud-x.com/index.php?option=com_remository&Itemid=34&func=download&filecatid=11

BTW. I have PM-ed you. Please chack your mailbox.

Pomyk

13th May 2006, 12:33

What bugs do you have in mind?
Actually the latest versions are very stable.
However, if there are any problems that you would like to report,
please follow this link:
http://www.aud-x.com/forum/viewforum.php?f=2
:thanks:

It's crashing on some normal mp3's. I reported it some time ago and sent a sample.

3dsnar

13th May 2006, 16:39

IMHO this bug is fixed. Please do the following:
1) make sure that you are using v. 1.21 (help tab)
if not
2) if not, please uninstall the current version
3) manually delete audxdshow.ax from c:\windows\system32 folder
4) install the new version

If the problem persists,
please let me know.
Thank you.

3dsnar

13th May 2006, 16:45

Big kudos to Jacquers
for (unofficial) AVI.NET
support for Aud-X MP3 5.1
http://forum.doom9.org/showthread.php?p=826587#post826587

Pomyk

13th May 2006, 18:34

The problem is still in v1.21. I think reported it after 1.21 was released so how could it be fixed? And I would really like to use this codec :/

edit: it crashes on every 192kb 48kHz mp3 here.

3dsnar

13th May 2006, 18:57

Please send us some example (problematic) file once more.
helpATaud-x.com
Thanks for the bug report.

dimzon

13th May 2006, 19:07

3dsnar
Any news about licencing problem (MPG123 still GPL)

3dsnar

13th May 2006, 19:45

So far we have provided information
(as u suggested) regarding this issue:
http://www.aud-x.com/content/view/5/27/
Cheers, 3d

dimzon

18th May 2006, 13:44

So far we have provided information
(as u suggested) regarding this issue:
http://www.aud-x.com/content/view/5/27/
Cheers, 3d
Please, dont get me wrong, but this is only short-term solution.
You must get explitit permissions from MPG123 author OR make your code open source

http://lame.sourceforge.net/license.txt

Can I use LAME in my commercial program?

Yes, you can, under the restrictions of the LGPL. The easiest
way to do this is to:

1. Link to LAME as separate library (libmp3lame.a on unix or
lame_enc.dll on windows)

2. Fully acknowledge that you are using LAME, and give a link
to our web site, www.mp3dev.org

3. If you make modifications to LAME, you *must* release these
these modifications back to the LAME project, under the LGPL.

*** IMPORTANT NOTE ***

The decoding functions provided in LAME use the mpglib decoding engine which
is under the GPL. They may not be used by any program not released under the
GPL unless you obtain such permission from the MPG123 project (www.mpg123.de).

PS. Can You provide me link to Aud-X Logo?

dimzon

19th May 2006, 16:39

BeHappy extension to support Aud-X 5.1
http://img105.imageshack.us/img105/6730/untitled7ir.png

3dsnar

20th May 2006, 05:45

Hi Dimzon,
yes, the explanation is a temporary solution.
---------
Looks good. Thanks for utilizing the graphics.
Cheers, 3d.

Rockaria

21st May 2006, 21:23

3dsnar

22nd May 2006, 06:08

I come to think the spatial info(ancillary data) is imbeded in the 'phase-out' area, controlling the frequency band anyhow, based on the FV-LAME-DPL II technology.

The ffdshow imbeding is an unexpected implementation for me, especially on a deviation(?) version, losing some functionalities except the core codec.
I didn't test the quality of the core codec because there was no mention of any improvement.

Now some usability issue :

<the ffdshow imbeded codec>
. the mp3 decoding works well, and supposedly the aud-x also(I didn't test)
. the ffdshow version has lost the 'check DTS signal in wav' feature.(reinstalled the formal version)

<the audx.exe cli encoder>
. perfectly 'no help', in a hope to integrate to foobar2k cli encoder front end.

<the aud-x dsfilter>
. insistantly sticks to the sound-out device only...
. not tested the crosstalk issue.

The hrtf feature is a good plan especially for LARGE headphone users.

Good Luck.
Hi Rockaria,
thanks for your comments.
It is not similar to DPLII, because the way the signals are being mixed is not used for creating 5.1 sound. In other words, there is not information hidden in the signal, which is later used for decoding, but only in the ancillary data. Moreover I have found out that people think that this is matrix (like DPLII) encoding. This is incorrect, because information describing other channels (beside the ones that are audible when played with standard mp3 player) is also present there.

Please address the FFDshow issues to FFDshow developers:
http://sourceforge.net/tracker/?group_id=53761&atid=471492

And audx cli issues to Dimzon:
http://www.gotdotnet.com/workspaces/workspace.aspx?id=1bb59ddf-901b-43a5-bd54-b0999e8e223e

---
Indeed the DSfilter still has the limitation that you mentioned.
When the HRTF thing will be implemented, I will report it here.

:thanks:

Rockaria

22nd May 2006, 06:25

3dsnar

22nd May 2006, 06:40

Still got a bit puzzled, not that much impacting my life yet though.;)

The audx.exe issue, I have probably figured out the parameters somewhere in this thread before, is I believe your issue for any users who might be interested in some other cli frontend integration....
Uhm, indeed. We will support all software toolz using aud-x on our web site and provide necessary information related to them (e.g. how to use CLI, etc).
Cheers, 3d.

Rockaria

22nd May 2006, 08:28

Good, that sounds reasonable.

..but only in the ancillary data. But it still does not answer where itself is located.
Moreover I have found out that people think that this is matrix (like DPLII) encoding. This is incorrect, because information describing other channels (beside the ones that are audible when played with standard mp3 player) is also present there.
AFAIK, I didn't start mentioning the DPL II as part of the solution to the crosstalks, fortunately.;)

Hum! but again, I conclude it deserves 'not disclosable', indeed.. (http://www.answers.com/indeed)

sillKotscha

22nd May 2006, 16:06

Hi 3dsnar,

just a little bugreport concerning the Aud-X DirectShow decoder...

if you choose "send surround sound through spdif" and you choose 'movie mode' as output then it is not remembered until the next start... it switches automatically to 'music mode'

and a little question... is it possible to implement some kind of sound detection? because if you choose "send surround sound through spdif" your sound will get upmixed to dpl II even if you have standard stereo - and that sounds not always good :)

maybe just like in ffdshow: you have the mixer option 'same as input' - so, if dpl II streams are available they will get decoded via dpl II if not only standard stereo decoding is used

thynk you - Sill

3dsnar

22nd May 2006, 16:22

Rockaria

22nd May 2006, 17:02

<the audx.exe cli encoder>
. perfectly 'no help', in a hope to integrate to foobar2k cli encoder front end.

Uhm, indeed. We will support all software toolz using aud-x on our web site and provide necessary information related to them (e.g. how to use CLI, etc).
Cheers, 3d.Sorry, but it's not that I wanted to hear!
Imbeding the help in the encoder is the widely accepted simple way.

3dsnar

22nd May 2006, 17:08

Sorry, but it's not that I wanted to hear!
Imbeding the help in the encoder is the widely accepted simple way.
Oh, I see.
U are right
----------
Dimzon, are you there?
Could you please add some help option in the STDIN CLI
that you've prepared?
Please let me know.
----------
We will add help to audx.exe (used by our frontend) as well.
Thanks, 3d

Rockaria

22nd May 2006, 17:22

Hi Sill,
If you chose: "Send surround sound through spdif"
all streams will always be converted to 5.1 (i.e. mono or stereo).
When you want to play a stereo sound, the 'Pseudosurround' mode
should work fine (instead of DPL2 upmix)

However, if you do not want to do any channel conversion,
you should select a standard stereo output, and force your soundcard to send the sound (as PCM) through spdif output.

:thanks: 4 the bug report.
If I am not incorrect, he made two inquires :
1. can it save the last setting to be reused for the next launches ?
2. can the output mode set automatic depening on the system speaker setting detection ?

Another sorry but I believe none of them answered?