07-05-2008, 10:57 AM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Hi,
I haven't been doing kstuff for a while, been changing my computer and I feel rusted, can you help me ?
I have two dvd-documentaries that I'd like to put on one dvd.
Some details :
- the dvds are 4/3 (720x576)
- the first one is interlaced (obvious combing effect), I think the second one is too, but it's less obvious because there's almost no movements.
- for the sound, the first dvd is LPCM, 48 khz, 16bits, 2 ch; the second one is LPCM, 48khz, 16bits, 1ch (!).
For both of theses dvd, there's no sound on my standalone dvd player (thomson), but it works on my PC.
So far, I have ripped the two dvd and made a d2v project + demux the sound with dgmpgdec150.
- Do I need to change the wav to ac3 with besweet ?
- what are the settings for fitcd ? (anamorphic or not, IUT-br ticked or not ?)
- what are the settings for HCEncoder ?
- what script may I use ?
Thank you.
- Do I need to change
|
Someday, 12:01 PM
|
|
Site Staff / Ad Manager
|
|
Join Date: Dec 2002
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
|
|
|
07-09-2008, 10:25 AM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Is everybody at the beach ?
Well, it summertime, I can wait.
|
07-09-2008, 12:05 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Weird I did not see your message yesterday when I came
Quote:
Originally Posted by jason
- the dvds are 4/3 (720x576)
- the first one is interlaced (obvious combing effect), I think the second one is too, but it's less obvious because there's almost no movements.
|
(TV ?) documentary are almost always interlaced (true interlacing).
Quote:
- Do I need to change the wav to ac3 with besweet ?
|
yes. or mp2. These are the only two formats for audio on a DVD.
Quote:
- what are the settings for fitcd ? (anamorphic or not, IUT-br ticked or not ?)
|
you say that they are 4:3 so not anamorphic. IUT ticked.
Quote:
- what are the settings for HCEncoder ?
|
Use HCMatic
Quote:
- what script may I use ?
|
Read there :
http://www.kvcd.net/forum/viewtopic.php?t=13768
|
07-09-2008, 01:07 PM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Thanks.
It's not "tv documentary", actually. They were given at differents exhibition. Both of them are burnt dvd, not pressed dvd.
Quote:
Yes, for evaluating the size, I'll do. But what the settings like "dc prec", "smp"; "scene changes"; "max bitrate"...
I have forgotten a little all theses settings.
In HC 0.23, there are 3 settings menu (settings 1, settings 2, settings 3), I don't even remember if "constant quantization" was checked or not, for instance.
Thanks for the interlacing guide.
May I use this script of yours with the interlacing filters ?
Code:
DGDecode_Mpeg2Source("Path/Filename.d2v",cpu=4,idct=7)
ColorMatrix(d2v="Path/Filename.d2v")
#TDeint(tryweave=true)
LanczosResize(...)
LimitedSharpenFaster()
LRemoveDust_YV12(17,1)
#MergeLuma(Blur(0.2)) # pour gagner de la compressibilité
#MergeChroma(Blur(1)) # sinon, non
AddBorders(...)
|
07-09-2008, 02:13 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jason
Yes, for evaluating the size, I'll do. But what the settings like "dc prec", "smp"; "scene changes"; "max bitrate"...
|
HCMatic takes care of all settings except gop length (use 12 for PAL standard, 24 is you want more compression but your DVD won't be in the standards) and max bitrate (you can let 9800 as HCEnc will reduce the max if needed).
Code:
DGDecode_Mpeg2Source("Path/Filename.d2v",cpu=4,idct=7)
ColorMatrix(d2v="Path/Filename.d2v")
#TDeint(tryweave=true)
LanczosResize(...)
LimitedSharpenFaster()
LRemoveDust_YV12(17,1)
#MergeLuma(Blur(0.2)) # pour gagner de la compressibilité
#MergeChroma(Blur(1)) # sinon, non
AddBorders(...)
[/quote]
Actually you can even use the script like that, just remove the # in front of TDeint. That line de-interlace the source pretty well, even if it introduce some slight blending.
Note: in HCMatic don't check "interlaced' as the source is progressive after "TDeint".
|
07-10-2008, 11:37 AM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Thanks,
I've been doing some quick tests.
At the end, HCMATIC said something like "can't determine accurately", is it normal ?
The two sample.ini have set the cq-maxbitrate to 5.
Can I set the GOP to 15 instead of 12 ?
|
07-10-2008, 02:18 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jason
Can I set the GOP to 15 instead of 12 ?
|
15 is normally max for NTSC, not PAL. But I don't think your player will refuse to read that.
Concerning your error, I'm very surprised that you can have something about 5 ! I never encode above 4 as the result is too bad over that value. Better to stay arround 3. Yous hould rework you script !
Nevertheless the error you have means that the ideal Q value is above the Qmax (or under the Qmin) that you set in HCMatic. The only way to solve that is to raise the QMax (or lower QMin). But as said above, 5 is already too bad !
|
07-10-2008, 02:35 PM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by Dialhot
Quote:
Originally Posted by jason
Can I set the GOP to 15 instead of 12 ?
|
15 is normally max for NTSC, not PAL. But I don't think your player will refuse to read that.
|
Really ??
I thought it was 15 for Pal and 18 for NTSC.
Quote:
Concerning your error, I'm very surprised that you can have something about 5 ! I never encode above 4 as the result is too bad over that value. Better to stay arround 3. Yous hould rework you script !
Nevertheless the error you have means that the ideal Q value is above the Qmax (or under the Qmin) that you set in HCMatic. The only way to solve that is to raise the QMax (or lower QMin). But as said above, 5 is already too bad !
|
Yep, I'm surprised, too. One of the DVD is 4 go (1h05), and the other 2 go (52 minutes), I thought it would be easy to put all that on only one dvd.
I guess the 4/3 takes much place...
I will retry your script with mergeluma and mergechroma, see you.
|
07-10-2008, 03:33 PM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Test with gop at 15; mergechroma/luma added in the script; q range between 2 and 7 :
average quantizer : 5.51
I've noticed the minimum bitrate was between 600 and 1600, that's pretty high, no ?
|
07-10-2008, 05:38 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
I think you are right for GOP length.
Yes 600 is very hight. I don't know what are these doc, but they're surely not a chess game filmed in real time .
Try this script :
Code:
DGDecode_Mpeg2Source("Path/Filename.d2v",cpu=4,idct=7)
ColorMatrix(d2v="Path/Filename.d2v")
#TDeint(tryweave=true)
LanczosResize(...)
LimitedSharpenFaster()
LRemoveDust_YV12(17,1)
ScriptClip(" nf = YDifferenceToNext()" +chr(13)+ "unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ).TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 1) ")
AddBorders(...)
|
07-11-2008, 10:59 AM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Hello,
I've tried your script. The clip plays but it says "there is no function named fmin".
I've been looking on this forum and elsewhere (doom9... ) and find no answer. I've added the plugin unfilter.dll, but the error is still there.
|
07-11-2008, 12:47 PM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jason
I've tried your script. The clip plays but it says "there is no function named fmin".
|
I've uninstall/reinstall avisynth 2.57, no changes.
Look, I may encode only one dvd. For the other, I may use the m2v demuxed by dgmpgedc and the mp2 created in besweet, it will make 1700 + 72 mo=1772 mo, maybe a little more when it's authorized, let's say 2000 mo top.
It gives me about 2500 mo, right ?
I've just tried hcmatic with only one dvd, and the normal script (with no scriptclip) : source 4.12 go, target 2.5 go, it gives me a q of 4.90.
It's not glorious, but it's the best q I've got so far.
If I'm able to make the scriptclip function working and put the dc prec on 9 or 10, I may gain a few q points.
And it's just documentaries after all, as long as it's watcheable, I don't need a perfect picture.
Anyway, thanks for your help.
|
07-11-2008, 01:01 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by jason
Quote:
Originally Posted by jason
I've tried your script. The clip plays but it says "there is no function named fmin".
|
I've uninstall/reinstall avisynth 2.57, no changes.
|
This is because it's a function, not a built-in command. I forgot I have it in a separate mail.
(note: avisynth 2.58 has now a built-in min/max).
Add these lines at the end of the script :
Quote:
function fmin( int f1, int f2) {
return ( f1<f2 ) ? f1 : f2
}
|
Remember also that mono sound at 64kbs is allowed on DVD too. May be you don't need to have a stereo, even at 112.
|
07-11-2008, 01:07 PM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Cool, no error.
I'm doing a test right now.
I will think about the mono sound, too (which would be logical since one of the dvd is 1 ch).
Thanks.
|
07-11-2008, 01:54 PM
|
Free Member
|
|
Join Date: Oct 2003
Posts: 130
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Q=4.27
It's better, isn't it ?
I'll do it like that.
Thanks again.
|
07-11-2008, 06:49 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
You're welcome .
|
09-06-2008, 09:56 PM
|
Free Member
|
|
Join Date: Mar 2008
Location: Tierra del Fuego, Chile
Posts: 121
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Quote:
Originally Posted by Dialhot
HCMatic takes care of all settings except gop length (use 12 for PAL standard, 24 is you want more compression but your DVD won't be in the standards) and max bitrate (you can let 9800 as HCEnc will reduce the max if needed).
|
How about a pulldowned NTSC material? How far can I go with gop?
also DC should always be set on 8 for low bitrates?
thanks
|
09-07-2008, 07:22 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Standard for NTSC is 15.
For a reason I can't explain HCenc seems to always produce smaller files with DC=9, not 8 or 10.
|
09-13-2008, 05:22 PM
|
Free Member
|
|
Join Date: Mar 2008
Location: Tierra del Fuego, Chile
Posts: 121
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
thanks, so i guess i can go as far as my standalone supports with gop, last thing, should i use open gop or closed for better compression?
saluts
|
09-13-2008, 06:35 PM
|
Free Member
|
|
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
|
|
Use closed GOP, it has several benefits and no real drawback (the impact on compression is not very big).
|
All times are GMT -5. The time now is 01:29 PM — vBulletin © Jelsoft Enterprises Ltd
|