wish list for HeadAC3he

[DarkAvenger]

Beside
-SSRC downsampling
-automatic delay compensation
-enhanced buffering/faster i/o
-more log output
-canceling process (just being too lazy to get things sorted out..)
which will def make it in near future,
what would you like to have on the wish list for future versions of HeadAC3he.

Current items (in no particular order):

-playback of AC3
-AC3-WAV->AC3
-input of other formats (WAV,OGG,MP2,MP3)
-output to other formats (MP2, well waiting for LGPL tooLame)
-multiple input files for batch processing, optional usage of common peak for 2pass
-splitting of output into preselected sizes/pieces
-splitting/selectioning in PAL/NTSC frames

Please give a *short* description of what you would like to see and why. Esp the latter points would be interesting to me, as once Smartripper gives out the d2v project without bugs, making SVCDs would be very speedy... (Now I decode complete ac3 to wav and select the range with TMPGENC and encode it using tooLame, while making the video with CCE, feeded with mpeg2dec.dll, with the range gained from TMPGEnc.)

[tangent]

Automatically normalize the mp3 after encoding using mp3gain at the end.

[user]

BeSweet / HeadAC3he with azid - AC3 / DD 2.0 sort problem max gain


dts files encoding would be useful.
Sometimes the dts track is better than the ac3 5.1 track.

Then users could take the dts 5.1 sound if they want and decode it to wav, downmix to stereo surround like with ac3 now....



DA answered:

dts decoding can cause trouble, not recommended, so not included.





Something different:

is there a program to add wav header to dts track ?
Like you can add wav header to ac3 with DA's HeadAChe.

Then you can listen to some dts soundtracks via CDplayer.

Perhaps new function for HeadAC3he ?






azid:

It is not really a bug.
Some time ago I asked and reported about following solution:

problem:
sort of gain / overflows of eg. AC3 2.0 streams.
With usual DD 5.1 the max value is on top of sort text file.
Then usual destination of gain etc. is no problem.
No problem for BeSweet either.

But 2.0 DD streams:

Often max gain is like 14.2 db.
As result you would calculate: 20.0 - 14.2 = 5.8 db.

But : sorting in text file:
as "highest" value is written:
9.9 db decreasing to 2.0 db.
Then in the middle/end of text:
14.2 db decreasing to 1.0, 0.0 db.

You understand what's the problem ?

If you automatically look out for highest gain value on top of text file list, then you get wrong.
You would choose 9.9 db instead of correct 14.2 db.

Result: gain too high, clipping, very worse.

In one of the last BeSweet versions (about 1 month ago) I watched this behaviuor. (as I did month ago, when I used manually command line azid, but then I found out, that I have to look in the middle of text file behind 2.0 db to find really highest gain...)

I have not tried HeadAC3he with this aim, but you can try yourself easily with a Dolby Digital/AC3 2.0 stream that you decode to stereo wav with maximizing gain as usual.

Either azids sort behaviour gets changed or BeSweet/HeadAC3he's functionality gets changed, so that they "scan" the text-file to find really highest gain, not just taking first value on top, 9.9 db, if highest gain (bigger than 9.9, eg. 14 db) is in middle of text-file.



DA answered:


HeadAC3he uses own way of analyzing gain. So no problem as described above regarding BeSweet-experience with HeadAC3he !

Use HeadAC3he with DD 2.0 / 5.1 !!

[user]

Hi,

useful:

if you have several ac3's,

1.ac3 , 2.ac3, 3.ac3,

adding them together to

all.ac3 = summe.ac3


It works with DOS :



copy 1.ac3/b + 2.ac3/b + 3.ac3/b + 4.ac3/b + 5.ac3/b + 6.ac3/b + 7.ac3/b summe.ac3/b

[DarkAvenger]

We have no love 4 x-posting.:mad:

reagrding 2nd post: Ok, I'll put it onto the list: Appending files.

[user]

selecting either 44,1 or 48 kHz wav header for ac3 file

We discussed it.

[user]

Hey,

I know....

But first I read and answered a thread with Azid-list.



Then I looked and searched for previous messages, on page 3 or 4 I found your wish-list for HeadAC3he.

So now I am collecting some wishes.

It is then at one thread for you...

[DarkAvenger]

Nevertheless, if you read my answer in the other thread, you know, you can del your post (copy) here.

[user]

-AC3-WAV->AC3

I guess, you have included this wish already in your list at the top.

[user]

Is there already a program to add wav header to dts track ?
Like you can add wav header to ac3 with DA's HeadAChe.

Then you can listen to some dts soundtracks via CDplayer.

Perhaps new function for HeadAC3he ?



Hi,
I copied a written text out of my previous.

But to this possibilty nobody answered so far, so excuse me.
It could be forgotten in previous messages with all the other text around.

I think , it is an interesting feature in future :)


P.S.:
I edited my x-post above with your answers, so it is clear.

[DarkAvenger]

About that dts thing: It is quite hard to impossible, since there are no docs on DTS.

[ChristianHJW]

My wish list for HeadAC3he :

- direct Vorbis 5.1 transcoding ( done ;) )
- direct Vorbis multi-channel transcoding with selectable channel coupling ( master, slave channels )
- pre-defined channel mapping for multi-channel transcoding, like
2 language Stereo ( 4 channels, 2 masters, 2 slaves )
3 language Stereo ( 6 channels, 2 masters, 4 slaves )
5.1 one language ( 6 channels, 3 masters, 3 slaves )
7.1 one language ( 8 channels, 3 masters, 5 slaves )
5.1 two languages ( 9 channels, 3 masters, 6 slaves )
including possibility to use different encoding qualities for different channels ( smaller bitrate )
- DTS decoding ( not downmixing .. for mc transcoding this is )

:D ;)

[DSPguru]

Originally posted by user
Hi,

useful:

if you have several ac3's,

1.ac3 , 2.ac3, 3.ac3,

adding them together to

all.ac3 = summe.ac3
create a list with those files and simply use BeSweet's payload mode.

about "overflow errors in BeSweet".
there were never erroRs of max gain in BeSweet, there was a rare semi-bug, when azid use to report +0db overflow, and this has been fixed.
i believe you know what 0db means...

btw,
BeSweet's ota( -g max ) is a smart normalizer that achieves higher gain values than -azid( -g max ) without harming the quality of the track.

[Koepi]

I _wish_ I could get my hands on a copy of headache0.21pre instead of real headaches trying to find a binary :)=

If I understand it correctly, it already is capable of SSRC downsampling? I'd love to use that on the soundtrack to the mvie I encode ATM ;)

Best regards,
Koepi

P.S.: tomorrow tripod should be up and running again! Yippi!

[DarkAvenger]

Well, you are welcome to be a betatester (though alpha tester would be more suitable). Yes, SSRC works like a charm in 2 pass dumb mode (similar to BeSweet's way).
Today I fixed a whole bunch of bugs. I am wondering why it worked so well on my machine...

[Koepi]

can you please email me a copy of it (de_koepi@lycos.de)? I am really keen on finishing that movie :)

EDIT: usually the stuff that you develop works way better on the developer machine as elsewhere. Might be related to more installed dll's, debug dll's that catch some exceptions etc.... That's the faith of a developer ;)

Thanks a million,
regards,
Koepi

[MaTTeR]

I can't possibly think of anything else I'd like to see added;)

The 2ch and 5ch Vorbis just simply rocks. Thanks for all your effort on the Vorbis issues! I know it hasn't been easy:)

[DarkAvenger]

-CPU priority selection
-possibility of more than one instance of HEadAC3he running

There are always things to be added. ;)

[Ripe73]

Yes a processing thread priority like "idle",i dont know if it works with sound:)