BeLight 0.22 beta 8 released [Archive] - Page 28

tebasuna51

1st October 2006, 12:56

so, should it be used for ac3 to aac encoding (source: dvd movie tracks...)???
Not at all, speech have a limited bandwidth and don't be used with movie tracks.

Use speech for contents like training, learning languages, news, ...

shon3i

1st October 2006, 13:21

Did, somebody try PNS. I tink can incrase quality @ 24kbs.

Rockaria

1st October 2006, 14:04

--speech - tune Encoder Engine for Speech
--pns - Enable Perceptual Noise Subsitution

Both options are said for speech enhancement, in different processes, (might be) usable together.
Considering these are configurable given the same aac conding condition(bit rate, mode,,,), applying to movie clips to get clearer dialog sounds also reasonable.

[edit] more->also,,, 4min

Kurtnoise

1st October 2006, 14:08

Not at all, speech have a limited bandwidth and don't be used with movie tracks...
It depends of movie tracks content I'd say. Directors or Actors comments as example...

Rockaria

1st October 2006, 15:56

shon3i

1st October 2006, 20:29

@shon3i,

Are there any guidelines(i.e. suggested bitrate boundaries) for the -pns and -speech?
I don't find any restrictions in the source.

A demo AAC pns coding in a slugish pdf(http://sound.media.mit.edu/mpeg4/audio/general/aes106_1-GeneralAudio.pdf) on Suzan Vega
is @ 32 kbit/s, fs=48 kHz, bandwidth 14 kHz .
There is no any restrictions, and both can be used on any bitrate. But winamp devs recommend pns only for LC-AAC encoding, so i have no idea where is best to use, which bitrate and other. IMHO i don't see any differents when is PNS on, i tryed with 64kbps LC-AAC, but i can't find noticable differents becuse have 14khz bandwith

Kurtnoise

1st October 2006, 23:16

PNS is targeted to improve efficiency for the lower bitrates...just like SBR tool.

Rockaria

2nd October 2006, 03:34

Thanks, it sounds to me the SBR is taking the high-bands while PNS is taking certain components images by the perception models, to regenerate the component/bands images when decoding. Some components area might be overlapping the SBR band, but not allways or completely.

• Perceptual Noise Substitution (PNS):
Exploiting noise-like components in the signal
• Long Term Prediction (LTP):
Taking advantage of very stationary / tonal signals

The -speech option seems to be the LTP in the linked pdf(last page), presumed mostly effective on monologues.

Kurtnoise

2nd October 2006, 10:06

Yep...note also that PNS is only compatible with MPEG-4 AAC not MPEG-2 AAC streams.

soundino

2nd October 2006, 12:15

BeLight 20061002 does not code in the following cases:
ac3(wav-pcm) -> Nero Digital (CLI) HEv2 *.m4a;
ac3(wav-pcm) -> Nero Digital (CLI) HE *.m4a;
wav-pcm -> 3GP Reference "EnhAacPlusEnc.exe is missing in BeSweet folder", though this file lays in a folder with BeSweet;
wav-pcm -> Speex, hangs also occurs nothing;
Why?
I use bsn_20060930 and WA_aac-2.4(20060930)
Where to take the information on flake?

Kurtnoise

2nd October 2006, 12:41

BeLight 20061002 does not code in the following cases:
ac3(wav-pcm) -> Nero Digital (CLI) HEv2 *.m4a;
ac3(wav-pcm) -> Nero Digital (CLI) HE *.m4a;

works fine here...

wav-pcm -> 3GP Reference "EnhAacPlusEnc.exe is missing in BeSweet folder", though this file lays in a folder with BeSweet;
wav-pcm -> Speex, hangs also occurs nothing;
Why?
For Speex, it works *only* with input files with 8/16/32 KHz as Sampling Rate. For the 3GPP reference, this will be fixed in the next build.

Where to take the information on flake?
On the website : http://flake-enc.sourceforge.net/ but what kind of infos would you like ?

shon3i

2nd October 2006, 18:53

Thanks, it sounds to me the SBR is taking the high-bands while PNS is taking certain components images by the perception models, to regenerate the component/bands images when decoding. Some components area might be overlapping the SBR band, but not allways or completely.

The -speech option seems to be the LTP in the linked pdf(last page), presumed mostly effective on monologues.
PNS scans noise and try to see is that noise is real noise or isn't, if yes then instead encode that noise, only signal that spectrum in bitstream. When decoding decoder generate random noise and put it desired signaled places. PNS must detect right places of noise or big artifacts will be noticable.

Rockaria

2nd October 2006, 22:15

Yes, I also expect 'right place' in the timeline & channel, when reproduce, with kinda component-instance-serialized-descriptions.
When seperating the Exploiting noise-like component instances in the signal, those meta info will be saved in a most precise/brief format.

We may require some more quotes from the related doc(s)..

guada 2

2nd October 2006, 23:33

Kurtnoise13,

Great job, update is OK.

A+ ;)

Bentso

17th October 2006, 14:12

BeLight 20061002 does not code in the following cases:
ac3(wav-pcm) -> Nero Digital (CLI) HEv2 *.m4a;
ac3(wav-pcm) -> Nero Digital (CLI) HE *.m4a;

I can't get HEv2 work either. HE works ok though.

MrPiercer

20th October 2006, 15:52

Is there a time frame for file list processing? I have a couple of hundred files in about 50 different dirs that I'd like to convert...

Kurtnoise

20th October 2006, 17:55

No...

Kurtnoise

20th October 2006, 17:57

I can't get HEv2 work either.
For Encoding or Playback ?

What kind of target bitrate or quality level did you set ?

Bentso

21st October 2006, 11:15

For encoding.

I tried to encode 5.1 AC3 to 5.1 AAC HEv2 Q = 0,2 and it didn't work. BeLight doesn't give any error. When I start the encoding it just stops after the PreGain part and the output size of the encoded file is 0 bytes.

I just tried stereo HEv2 file and it encoded properly.

Maybe the 5.1 isn't supported in HEv2.

tebasuna51

21st October 2006, 12:02

I just tried stereo HEv2 file and it encoded properly.

Maybe the 5.1 isn't supported in HEv2.
AAC HE v2 = AAC HE + PS = AAC + SBR + PS

SBR = Spectral Band Replication

PS = Parametric Stereo
"The Parametric Stereo encoder extracts a parametric representation of the stereo image of an audio signal, whereas only a monaural representation of the original signal is encoded in a conventional fashion. The stereo image information is represented as a small amount of high quality parametric stereo information and transmitted along with the monaural signal in the bit stream. Based on the parametric stereo information, the decoder is capable of regenerating the stereo image."

shon3i

21st October 2006, 12:29

AAC HE v2 = AAC HE + PS = AAC + SBR + PS

SBR = Spectral Band Replication

PS = Parametric Stereo
"The Parametric Stereo encoder extracts a parametric representation of the stereo image of an audio signal, whereas only a monaural representation of the original signal is encoded in a conventional fashion. The stereo image information is represented as a small amount of high quality parametric stereo information and transmitted along with the monaural signal in the bit stream. Based on the parametric stereo information, the decoder is capable of regenerating the stereo image."
Visualy

http://www.codingtechnologies.com/products/assets/PS_level1.gif

soundino

23rd October 2006, 02:51

I have understood so.

At coding 5.1wav-pcm a file on a parameter of quality:
- At 0.2 <Q <0.3 it is received AAC HE;
- At Q> 0.3 it is received AAC LC;
- At Q <0.2 for 5.1wav-pcm coding does not occur and probably it is connected to installations of the coder.
AAC HEv2 it turns out for stereo-wav a file at 0 <Q <0.1. At 0.1 <Q <0.4 it is received AAC HE, and at Q> 0.4 it is received AAC LC.

By the way, WinAmp (CT) at me for some reason refused to code in AAC LC.

logfile:

Using Shibatch.dll v0.25 by Naoki Shibata & DSPguru (shibatch.sourceforge.net).
Using WA_aac.dll v2.4 by Kurtnoise (kurtnoise@free.fr)

C:\Audio tools\BeLight-0.22beta9_20060930\BeSweet.exe -core( -input E:\Music\prob.wav -output E:\Music\prob.aac -logfile E:\Music\prob.log ) -dimzon( -dllname WA_aac.dll -cbr 320 -mode 2 -lc )
[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : E:\Music\prob.wav
[00:00:00:000] | Output: E:\Music\prob.aac
[00:00:00:000] | Floating-Point Process: No
Error 84: error configuring bsn!
Quiting...
[00:00:00:000] Conversion Completed !

tebasuna51

23rd October 2006, 08:55

@soundino
1) stereo wav
-dimzon( -dllname WA_aac.dll -cbr 320 -mode 2 -lc )
OK

2) wav 6 chan
-dimzon( -dllname WA_aac.dll -cbr 320 -lc -6chnew )
OK

3) wav 6 chan
-dimzon( -dllname WA_aac.dll -cbr 320 -mode 2 -lc )
Error 84: error configuring bsn!
Quiting...

Is your prob.wav 6 chan?

Kurtnoise

23rd October 2006, 09:03

Note that with BeLight, -6chnew is setted automatically...Otherwise your bsn.dll is too old.

tebasuna51

23rd October 2006, 11:29

Note that with BeLight, -6chnew is setted automatically...Otherwise your bsn.dll is too old.
If last setting is Stereo and I open a wav6, I need set manually 5.1 otherwise I get the "Error 84: error configuring bsn!" with BeLight.

bsn.dll 49152 bytes 30-09-2006
BeSweet v1.5b31 by DSPguru.
--------------------------
Using Shibatch.dll v0.25 by Naoki Shibata & DSPguru (shibatch.sourceforge.net).
Using WA_aac.dll v2.4 by Kurtnoise (kurtnoise@free.fr)

Logging start : 10/23/06 , 09:45:13.

D:\BeSweet.exe -core( -input D:\6chan.wav -output D:\outp.aac )
-dimzon( -dllname WA_aac.dll -cbr 320 -mode 2 -lc )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : D:\6chan.wav
[00:00:00:000] | Output: D:\outp.aac
[00:00:00:000] | Floating-Point Process: No
Error 84: error configuring bsn!
Quiting...
[00:00:00:000] Conversion Completed !

Logging ends : 10/23/06 , 09:45:13.

Kurtnoise

23rd October 2006, 12:28

yes, just like for all 5.1 output :). You have to check "5.1" for output channel because is stereo by default...

by automatically, I meant "click on 5.1 and -6chnew will be added in the command line"...

tebasuna51

23rd October 2006, 13:33

Sorry Kurtnoise, I misunderstand your previous comment.

@soundino
If your prob.wav is 6 channels you must check "5.1" button.

soundino

24th October 2006, 09:39

I use bsn.dll 49152 bytes 30-09-2006.
At coding probe6ch.wav with set "5.1" button.
Logfile:

BeSweet v1.5b31 by DSPguru.
--------------------------
Using Shibatch.dll v0.25 by Naoki Shibata & DSPguru (shibatch.sourceforge.net).
Using WA_aac.dll v2.4 by Kurtnoise (kurtnoise@free.fr)

C:\Audio tools\BeLight-0.22beta9_20060930\BeSweet.exe -core( -input C:\temp\prob\1\probe6ch.wav -output C:\temp\prob\1\probe6ch_1.aac -logfile C:\temp\prob\1\probe6ch_1.log ) -dimzon( -dllname WA_aac.dll -cbr 320 -lc -6chnew )

[00:00:00:000] +------- BeSweet -----
[00:00:00:000] | Input : C:\temp\prob\1\probe6ch.wav
[00:00:00:000] | Output: C:\temp\prob\1\probe6ch_1.aac
[00:00:00:000] | Floating-Point Process: No
Error 84: error configuring bsn!
Quiting...
[00:00:00:000] Conversion Completed !

In general, I do not understand, that occurs!:(

P.S. At coding probe6ch.wav in HE AAC all is normal.

Kurtnoise

24th October 2006, 12:33

That's strange...can't reproduce this here. Which enc_aacplus.dll did you use ? (at least, check the date in Properties of this library.)

soundino

24th October 2006, 13:29

Thanks Kurtnoise13.
Really, at me was old enc_aacplus.dll a file.
Once again all thanks!:)

sillKotscha

26th October 2006, 21:07

Hi Kurtnoise13,

I have a question about the BeLight Log-file...

it states that two encoder are missing but AFAIK I should have all off them in my BeLight-Folder...

does BeLight doesn't recognize them - therefor the error message - or do I have wrong versions (= too old version?)

thanks in advance, Sill

Log:

http://img55.imageshack.us/img55/2483/logqr9.jpg (http://imageshack.us)

my "installed" *.exe files:

http://img55.imageshack.us/img55/5291/exenc6.jpg (http://imageshack.us)

my "installed" *.dll files:

http://img523.imageshack.us/img523/4250/dllwf3.jpg (http://imageshack.us)

EDIT:

btw, do I have obsolete files in my BeLight directory? And it is impossible to reach http://www.corecodec.org/frs/?group_id=45&release_id=292#r292 at the moment... is there any other place to download dimzon's latest files?

slicksta

26th October 2006, 22:25

Im trying transcode a whole heap of mp3s to mp4 or AAC+ to use in my W850 (sony/ericsson phone), but despite going over this forum, I cant seem to find a solution to getting more than 1 input file to process at a time. For some reason it only processes the last file in the list. I have a few thousand to do, so 1 at a time would take forever. Any help/advice appreciated. Im using 0.22beta9.

Bluedan

26th October 2006, 22:49

@slicksta Your question has already been answered here. (http://forum.doom9.org/showthread.php?p=881619#post881619)

Don't use drag'n drop, but open file(s) dialogue.

Kurtnoise

27th October 2006, 12:58

I have a question about the BeLight Log-file...

it states that two encoder are missing but AFAIK I should have all off them in my BeLight-Folder...

does BeLight doesn't recognize them - therefor the error message - or do I have wrong versions (= too old version?)

This is not errors but just some warnings...:) Errors are reported with an other picture (the white cross with red background). Besides, BeLight works fine even these warnings.

I know there is a bug with enAacplusEnc. It's already fixed but I can't upload a new build at the moment as you can see...CC.org is down for me too. Maybe tonight...

About Aud-X encoder, this is not the right one. You just have to install Aud-X installer (BeLight uses the Aud-X encoder from Program Files\Aud-X\...) and that should be ok. Note that there is none obligation to install this compressor (only needed for mp3 surround).

About obsolete files : if you use NeroAACEnc.exe then aac.dll & aacenc32.dll are useless. Same thing with bsn.dll : if you use this last one then all bse_xxxxx.dll are not needed anymore.

sillKotscha

27th October 2006, 17:25

Hi Kurtnoise13,

thank you for the answer.

TheManuel

28th October 2006, 16:51

Is there a daily build of BeLight that supports both .avi files as input and the Aften encoder?

Thanks.

Kurtnoise

28th October 2006, 16:59

Nope...but avi support will be included soon.

TheManuel

28th October 2006, 17:58

Great! I look forward to that.
Thanks for sharing your work.

Audionut

29th October 2006, 01:43

Hi, I'm trying to mux 6 mono wav files to a 5.1 wave file. But all i'm getting is noise in return.

The transcode finishes completely and the file is created.

slicksta

2nd November 2006, 04:33

"@slicksta Your question has already been answered here.

Don't use drag'n drop, but open file(s) dialogue"

Like I said Ive been over this entire forum. This was the first thing I tried. For some reason its still not working. Only transcodes one file from the list. Am I missing something?
I tried different encoders as well as previous Belight versions, & still no go. Any suggestions/solutions?

m0rtal

2nd November 2006, 07:23

Bluedan

2nd November 2006, 09:14

Indeed. Slicksta, sorry, you were right. I confirm.

m0rtal

2nd November 2006, 10:44

Indeed. Slicksta, sorry, you were right. I confirm.
so it's a common problem? :-\

slicksta

4th November 2006, 07:06

Indeed. Slicksta, sorry, you were right. I confirm.

Thanx for confirming Bluedan

Any suggestions Kurtnoise?

Kurtnoise

5th November 2006, 10:38

Any suggestions Kurtnoise?
Wait a new build...or use fb2k.

MacAddict

5th November 2006, 16:45

The AC3 384kbps preset enables 'hybrid' gain. Is this correct or should I use 'pre-gain'? I'm using the Aften mode so should I have anything else turned on within the 'advanced' window such as DNR?

Any info is much appreciated. Thx again for such a great app!