Q increase problem using d2sroba since upgrading all components - help needed

[homerjay]

i will now explain

having recently updated all my programs / plugins to latest versions i started a new encoding burst and noticed the two films i wanted to convert came up with a huge Q value greater than 80 in both cases for films of less than 100 minutes :confused:

so i spoke to a friend who has reported similar increases in Q with his 1cd encodes.

i upgraded all of the following ensuring that any changes to the default scripts were reintroduced after upgrading

CCE version
ECLCCE
D2SROBA
FACAR

i decided to run a reconversion test on a given film which i had converted previously which came out the time with a Q of only 24 for a 112 minute film - see original log below

========================================================
D2SRoBa v2.7.0 RC2 - DVD2SVCD plugin, by tylo
OS Version: WIN_XP
--------------------------------------------------------
- DVD2SVCD Shutdown
- 2003-10-25 12:58:35
--------------------------------------------------------
Settings:
- Encoding mode : OPV RoBa (Recover=No)
- CD size : 800
- Number of CD's : 1
- Auto Q. (worst) : ---
- Maximum Q. : 128
- Sample percentage : 2.0
- Adjust Q. : -0.8
- Sample GOPs : 1
- Sizing pass : on=0 (-2.0 .. 1.0)%
- Audio after video : on=0 (Fixed bitrate, 10%, min 128 kbps)
- Adjust max video br : on=0 (total 2754 kbps)
--------------------------------------------------------
- Movie length : 01:52:09 (168229 frames, 25 fps)
- D2S audio settings : (128 kbps + 0) kbps
- D2S video estimate : 832 kbps, 1 CDs
- SelectRangeEvery : every 750, select 15 frames
- Num. sample frames : 3375
--------------------------------------------------------
Calculations:
- 1 CD: Video BR=832 Audio 1 BR=128 Video SZ=699832640.
- 2 CD: Video BR=1808 Audio 1 BR=128 Video SZ=1520790160.
- 3 CD: Video BR=2784 Audio 1 BR=128 Video SZ=2341747680.
- 4 CD: Video BR=3760 Audio 1 BR=128 Video SZ=3162705200.
--------------------------------------------------------
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Number of CDs : 1
- Target mpv size : 699832640. (832 / max 2530) kbps
--------------------------------------------------------
- Binary search for Q:
- Estimated mpv size : 436070703. (Q=64, 518 kbps, 8748424 sample sz)
- Estimated mpv size : 589464647. (Q=32, 700 kbps, 11825804 sample sz)
- Estimated mpv size : 828901718. (Q=16, 985 kbps, 16629376 sample sz)
- Estimated mpv size : 679458538. (Q=24, 807 kbps, 13631256 sample sz)
- Computed mpv size : 754180128. (Q=20, 896 kbps, 15130316.=(Q24+Q16)/2)
- Computed mpv size : 716819333. (Q=22, 852 kbps, 14380786.=(Q24+Q20)/2)
- Estimated mpv size : 693830030. (Q=23, 824 kbps, 13919576 sample sz)
--------------------------------------------------------
- Q Weight : 1.0 - (699832640. / 49.8456296296296 - 13919576) / (14380786. - 13919576)
- Adjust Q : (23 + (0.738895513974091) - (-0.) = 24.5388955139741
- Determined Q. : 24
--------------------------------------------------------
- Encoding Movie
- 2003-10-25 13:18:05
- Actual mpv file size:
25/10/2003 15:31 650827624 Encoded_Video_CCE_PAL.mpv
- OPV pass result: 93.% on target ( 650827624 / 699832640.)
--------------------------------------------------------
- Recover from pulldown/muxing
- 2003-10-25 15:31:25

the new conversion came out with a Q of 28 - not as high as i was expecting but still higher than the original

--------------------------------------------------------
D2SRoBa v3.50, by Tylo
--------------------------------------------------------
- WIN_XP - AutoIt 3.0.101.0
- 2004-05-02 18:43:05
--------------------------------------------------------
- Output type : SVCD
- CD size : 800
- Number of CD's : 1 (max 32)
- Sample percent : 2.0
- Adjust % : 0.8
- Cond. sizing pass : No
- Cut last image : If opv sz > 0.4% oversize
- Audio after video : Yes (128 kbps)
- Adjust max br : No
- Clean previous : No
- Run mode : Normal
--------------------------------------------------------
- Waiting for CCE window ...
- Detected a CCE encode window
- Detected the movie encode window - shutting down
--------------------------------------------------------
- 2004-05-02 18:53:32
- Project dir: E:\DVD2SVCD\Movie\
- Config file: E:\DVD2SVCD\Movie\DVD2SVCD.INI
- Avisynth ini: E:\DVD2SVCD\AVISYNTH.INI
- Avisynth setup: [FACAR], cond. filter: None
- D2S ver. 1.2.1.3, CCE ver. 2.67.0.27
--------------------------------------------------------
- Movie length : 02:01:40 (182523 frames, 25 fps)
- D2S audio select : 128 + 0 kbps
- D2S video calc. : 1655 kbps, 2 CDs
- Select Ranges : every 750, select 15 frames
- Sample frames : 3660
--------------------------------------------------------
Calculations:
- 1 CD: video_br=756 (756) audio_br=128+0 video_sz=689936940 fill=100% cbr=0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Using num CDs : 1
- Target mpv BR : 756 (max 2530) kbps, size=689936940 bytes
--------------------------------------------------------
Search for Q:
- Sample encode : Q=66: 508 kbps, err=-32.8%, size=463852582, sample sz=9301296
- Sample encode : Q=31: 719 kbps, err=-4.9%, size=657067041, sample sz=13175684
- Sample encode : Q=25: 803 kbps, err=6.2%, size=733576294, sample sz=14709868
- Switch to bin search
- Sample encode : Q=28: 758 kbps, err=0.3%, size=691830591, sample sz=13872772
- Determined Q : 28 = Round(28 + (0.3 - 0.8)/2)
--------------------------------------------------------
- 2004-05-02 19:18:06
- Start movie OPV encoding (Q 28)
- OPV pass result: 98.5% on target (679685932 / 689936940) Speed factor: 0.64
--------------------------------------------------------
- 2004-05-02 22:29:39
- Executing: E:\DVD2SVCD\BeSweet\BeSweet.exe -core( -input "E:\DVD2SVCD\Movie\Extracted_audio_1.ac3" -output "E:\DVD2SVCD\Movie\Encoded_audio_1.mp2" -logfile "E:\DVD2SVCD\Movie\Encoded_audio_1.log" ) -azid( -L -3db -C 0db -S 0db -c normal -s surround -g max ) -shibatch( --rate 44100 ) -2lame( -e -b 128 -m s )
--------------------------------------------------------
- 2004-05-02 22:46:25
- Recover DVD2SVCD
--------------------------------------------------------
- 2004-05-02 23:39:35
- D2SRoBa successfully finished
--------------------------------------------------------



my question is this - has anybody else noticed significant increases in Q on recent conversions and if so have you been able to identify what part of the process has changed

thanks

[tylo]

Hi, homerjay.
D2SRoBa does not influence the relation between Q and bitrate. However, there are many differences between the two runs that may affect it.

1. Different target bitrate. This is because D2SRoBa now uses D2S's calculations for bitrate.

2. Possible differences in FACAR settings /defaults.

3. Different CCE version - may change Q / bitrate relation.

4. Different CCE template / D2S usage of template. (Your old run may have used CCE filters - cuts down Q, but does not always look better).

Brief answer, but hope it helps.
Cheers.

[homerjay]

thanks for clarifying things a bit tylo :)i had already noticed the different target bitrate (1) and that would explain a one or two point change in Q maybe:)

as for (2) the FACAR script had definately changed so it could be a factor

(3) only way to test is to go back a version or two and retry i guess

my filters havent changed between encodes for (4) although as you say their usage may have

this is a little disapointing if there is nothing that can be done but go back to older versions as i am a firm believer in keeping your s/w up to date so anybody elses experiences would be appreciated

ps love the polar bear icon for d2sroba :D

[Manngo]

@HomerJay

Maybe you are interested in this. (http://forum.doom9.org/showthread.php?s=&threadid=73954) This thing coused me even 40-50 difference in the found Q value. I use different matrices for different length movies, to get a Q value between 10-30.

[homerjay]

thanks for your input manngo - i would definately like to try implementing the second matrix to see what results come out :D

how or where should this matrix be entered as it doesnt take the form of the existing matrices in the .ini file that have intra and not intra

also if you are reading DDogg i have seen you metion a bach1 matrix you have used - i found it in a thread but when i tried to use it it came up with an error ( i dont have the error currently )

would you mind posting its content here please just in case i copied it incorrectly

thanks :)

[Pixar2]

Maybe a related problem to what Tylo said: "1. Different target bitrate. This is because D2SRoBa now uses D2S's calculations for bitrate."

When I try to encode shortish movies, i.e. TV shows, I run into problems with bitrate calculation. This is what d2sroba tells me after being fed with a 55.49 min.sec NTSC avi:
--------------------------------------------------------
D2SRoBa v3.50, by Tylo
--------------------------------------------------------
- WIN_2000 - AutoIt 3.0.101.0
- 2004-05-04 17:46:19
--------------------------------------------------------
- Output type : SVCD
- CD size : 800
- Number of CD's : 1 (max 32)
- Sample percent : 1.0
- Adjust % : 1.2
- Cond. sizing pass : Yes (Transcode if opv sz > 0.3% oversize, and Q <= 40)
(Encode if opv sz < 2.5% undersize, or Q > 40)
- Audio after video : Yes (9% of video, min 160 kbps)
- Adjust max br : Yes (total: 2754, max avg offset: 300 kbps)
- Clean previous : Yes
- Run mode : Normal
--------------------------------------------------------
- Waiting for CCE window ...
- Detected a CCE encode window
- Detected the movie encode window - shutting down
--------------------------------------------------------
- 2004-05-04 17:50:37
- Project dir: D:\Mina Filmer\Under arbete\Source\
- Config file: D:\Mina Filmer\Under arbete\Source\SVCD.INI
- Avisynth ini: D:\Mina Filmer\Under arbete\Source\AVISYNTH.INI
- Avisynth setup: [FitCD], cond. filter: None
- D2S ver. 1.2.2.1, CCE ver. 2.67.0.23
- Setting low process priority
--------------------------------------------------------
- Movie length : (80381 frames, 0 fps)
- D2S audio select : 160 + 0 kbps
- D2S video calc. : kbps, 0 CDs
- Select Ranges : every 1500, select 15 frames
- Sample frames : 810
--------------------------------------------------------
Calculations:
- 1 CD: video_br=2530 (-1.#IND) audio_br=224+0 video_sz=-9.22337203685478e+018 fill=-1.#IND% cbr=1
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Using num CDs : 1
- Target mpv BR : 2530 (max 2530) kbps, size=-9.22337203685478e+018 bytes
--------------------------------------------------------
--------------------------------------------------------
- 2004-05-04 17:50:39
- Start movie CBR encoding (2530 kbps)

Using all the latest; 1.2.2 B1 clean install. No funny scripts, resize given by AutoFitCD only. Bitrate should land somewhere around the 1700's range instead?

Humbly noted,
Pixar2

/ADD
Hei Tylo!

Norwegian, right? I'm Swedish, so bear with me :cool:

Summarizing my findings after testing some different scenarios:

No changes made in D2SRoBa setting between tries. In DVD2SVCD I only load source and run AutoFitCD to get the right crop/resize. I have two similar clips around 50 mins, one slightly longer than the other, the behaviour is exactly the same.

1 Ran it in a clean directory, and now it worked fine! (WTF?)
2 Ran it in the same directory, not deleting any files, same error appeared as before
3 Deleted all files, no error
4 No deleting of files, same error appears again
5 No deleting of files, using a longer movie of ab 90 min, no error :confused:
6 No deleting of files, back to my 50 min clip, the error appears again
7 Deleting only .d2s file, error remains
8 Now it gets interesting, using an empty directory and a shorter clip, ab 40 mins, it tries again to run CBR max bitrate, real bitrate should be just over 2500 says FitCD. Also the same problem with the .d2s file is there - no bitrate given and no of CDs is zero.

This beats me. It seems to be connected to wether the directory is empty or not but also to the length of the clip. Anybody experienced anything like this with shorter clips? Can it be recreated by someone else? The result if I let it run to finish is that I get one CD but without the end of the clip.

This is the projectfile (d2s) I get, most info missing like you suspected Tylo. In the successful run the project position is 32 BTW, not 31 as below. And the movie section is filled in naturally.

[General]
Version=AVI2SVCD ver. 1.2.2 build 1

[Filenames]
DVDAudioFileName0=D:\Mina Filmer\Test short\Audio\Extracted_audio_1.mpa
MP2FileName0=D:\Mina Filmer\Test short\Audio\Encoded_audio_1.mp2
DVDAudioFileName1=D:\Mina Filmer\Test short\Audio\Extracted_audio_2
MP2FileName1=D:\Mina Filmer\Test short\Audio\Encoded_audio_2.mp2
IFOFilename=
MPVFileName=D:\Mina Filmer\Test short\Video\Encoded_Video_CCE_NTSC.mpv
FinalMPVName=D:\Mina Filmer\Test short\Video\Encoded_Video_CCE_NTSC.mpv
TitlePictureFileName=D:\Mina Filmer\Test short\Video\TitlePicture.bmp
ChangeCDPictureFileName=D:\Mina Filmer\Test short\Video\ChangeCDPicture.bmp
LastCDPictureFileName=D:\Mina Filmer\Test short\Video\LastCDPicture.bmp
TargetFolder=\
FinalbbMPEGFileName=D:\Mina Filmer\Test short\Mux\bbMPEG_Muxed_File
DVD2SVCD INI File=D:\Mina Filmer\Test short\Source\DVD2SVCD.INI

[Settings]
Project Position=31
ForceFilm=0
FileSource=2
PGCISelect=0
bbMPEGCDSize=800
ReduceCDSize=0
Frames=0
VideoPTS=0
PAL=0
NTSC2PAL=0
TitlePictureFormat=1
ChangeCDPictureFormat=1
viHorizontalSize=640
viVerticalSize=368
viAspectRatio=2
viFrameRate=1
vifps=23976
SelectedAspectRatio=2
CDLabel=
MovieTitle=
Number of images=0
Volumes=0
fps=0
SubYPos=0
Deinterlace=0
SmartDeinterlace Mode=2
SmartDeinterlace Threshold=10
SmartDeinterlace Denoise=1
SmartDeinterlace Blend=0
SmartDeinterlace Fast=1
CCEProgressive=0
Use Topfield First=1
AudioPTS0=0
Delay0=0
AudioSource0=1
AudioLangCode0=--
AudioLanguage0=Unspecified
AudioStream0=0
AudioIndex0=0
AudioChannels0=0
AudioBitrate0=160
wFormatTag0=0
nSamplesPerSec0=0
AudioPTS1=0
Delay1=0
AudioSource1=0
AudioLangCode1=
AudioLanguage1=
AudioStream1=-1
AudioIndex1=0
AudioChannels1=0
AudioBitrate1=0
wFormatTag1=0
nSamplesPerSec1=0
NumOfAudioStreams=1

[VOBFiles]
VobName_0=D:\Mina Filmer\Test\non.copyrighted.film.avi

[MovieInfo]
SourceType=AVI (xvid)
MovieOffset=
CDOverlapSeconds=
StartupDelays=
VideoFormat=NTSC
Length=
Images=
RealMovieTitle=
UserRating=
DirectedBy=
WritingCredits=
Genre=
Tagline=
PlotOutline=
AdditionalInfo=
Encoder=
EncodingMode=
AvgBitrate=
MaxBitrate=
MinBitrate=
QFactor=
Passes=
ImageQualityPriority=
Bias=
AntinoiseFilter=
LinearQuantizerScale=
ZigzagScanningOrder=
UpperFieldFirst=
ProgressiveFrames=
CQValue=
MotionSearchPrecision=
DCComponentPrecision=
OutputYUVData=
FloatingPointDCT=
NoMotionSearch=
SoftenBlockNoise=
iDCTAlgorithm=
AvisynthScript=
Matrix=
NTSCFieldOperation=
Deinterlace=None
Pulldown=No
Width=0
Height=0
Audiostreams=1
AudioLanguage0=Unspecified
AudioBitrate0=160
AudioSampleFrequency0=48000
AudioMode0=Joint stereo
AudioLanguage1=
AudioBitrate1=
AudioSampleFrequency1=
AudioMode1=
Subtitlestreams=0
Type=
SubtitleInfo0=
SubtitleInfo1=
SubtitleInfo2=
SubtitleInfo3=
Authoring=
UseEntrysvd=


Maybe time to create a separate topic? Thankful for any input.

[tylo]

Hei, Pixar2

- Movie length : (80381 frames, 0 fps)
- D2S video calc. : kbps, 0 CDs

Don't know what's going on here, but length, framerate, calculated bitrate and num CDs (D2S's) are all zero. Those figures are taken from the .d2s project file. If you can reproduce it (run again clean), please post the .d2s file.

/Add: it seems to me that those keys in your .d2s file exists, but they are empty:
[Settings]: fps, Number of images
[MovieInfo]: Length, AvgBitrate

[MichiganFocker]

Ok I am the other person homerjay mentioned having the problem with a high Q probem. I for the life of me cannot get set back up. I keep getting a 2nd CD for the most part just 1 min or credits. Plus a high Q.


Just ripped predator which was approx 102 minutes or so. Got a Q of 74. Using all the latest DVD2SVCD_122_B1, FACAR_1.06.7, D2SRoBa350, ECLCCE_1.81, CCE 2.67.0.27.

I set back up the way I had it with the older versions and for the life of me cannot get back to normal. I get a even bigger 2nd CD that way 180mb and 260mb. I was getting a Q most of the time around 35-38 all the time this way before I upgraded proggys. Never once got a 2nd CD.

Here is the log for predator I just did with all the latest proggys.

--------------------------------------------------------
D2SRoBa v3.50, by Tylo
--------------------------------------------------------
- WIN_XP - AutoIt 3.0.101.0
- 2004-05-05 10:15:56
--------------------------------------------------------
- Output type : SVCD
- CD size : 800
- Number of CD's : 1 (max 32)
- Sample percent : 2.0
- Adjust % : 1.2
- Cond. sizing pass : No
- Cut last image : If opv sz > 0.3% oversize
- Audio after video : Yes (9% of video, min 128 kbps)
- Adjust max br : Yes (total: 2754, max avg offset: 400 kbps)
- Clean previous : Yes
- Run mode : Normal
--------------------------------------------------------
- Waiting for CCE window ...
- Detected a CCE encode window
- Detected the movie encode window - shutting down
--------------------------------------------------------
- 2004-05-05 10:36:51
- Project dir: C:\Program Files\DVD2SVCD\Movie\
- Config file: C:\Program Files\DVD2SVCD\Movie\SVCD.INI
- Avisynth ini: C:\Program Files\DVD2SVCD\Movie\AVISYNTH.INI
- Avisynth setup: [FACAR], cond. filter: None
- D2S ver. 1.2.2.1, CCE ver. 2.67.0.27
--------------------------------------------------------
- Movie length : 01:46:32 (153273 frames, 23.976 fps)
- D2S audio select : 128 + 0 kbps
- D2S video calc. : 884 kbps, 1 CDs
- Select Ranges : every 750, select 15 frames
- Sample frames : 3075
--------------------------------------------------------
Calculations:
- 1 CD: video_br=884 (884) audio_br=128+0 video_sz=706400838 fill=100% cbr=0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Using num CDs : 1
- Target mpv BR : 884 (max 2626) kbps, size=706400838 bytes
--------------------------------------------------------
Search for Q:
- Sample encode : Q=56: 1039 kbps, err=17.5%, size=831055477, sample sz=16672836
- Sample encode : Q=72: 906 kbps, err=2.5%, size=724242888, sample sz=14529936
- Sample encode : Q=75: 887 kbps, err=0.3%, size=708834839, sample sz=14220816
- Determined Q : 74 = Round(75 + (0.3 - 1.2)/0.7)
--------------------------------------------------------
- 2004-05-05 10:44:03
- Start movie OPV encoding (Q 74)
- OPV pass result: 102% on target (720225928 / 706400838) Speed factor: 0.94
- Oversized: Cutting last CD image off by 2%
--------------------------------------------------------
- 2004-05-05 12:37:32
- Executing: C:\Program Files\DVD2SVCD\BeSweet\BeSweet.exe -core( -input "C:\Program Files\DVD2SVCD\Movie\Extracted_audio_1.ac3" -output "C:\PROGRA~1\DVD2SVCD\Movie\Encoded_audio_1.mp2" -logfile "C:\PROGRA~1\DVD2SVCD\Movie\Encoded_audio_1.log" ) -azid( -L -3db -C 0db -S 0db -c normal -s surround -g max ) -shibatch( --rate 44100 ) -2lame( -e -b 128 -m s )
--------------------------------------------------------
- 2004-05-05 13:15:08
- Recover DVD2SVCD
--------------------------------------------------------
- 2004-05-05 13:38:19
- D2SRoBa successfully finished
--------------------------------------------------------


Hope this gets figured out, I'm not gonna have any hair left soon...:D ;)

[r6d2]

Using FACAR, there are three easy ways to reduce Q:

1. Use a good filter, like UnDot()+Deen() (KISS) on the preresize script

2. Use a smaller frame size (ResizeTo=2)

3. Use Bilinear resizer.

Also remember to set FACAR overscan parameters to crop what you cannot see.

Other than that, I don't think that other thing has changed to increase your Q. Anyway, 110 minutes is a long movie for 1 CD, even if it is a 16:9.

Cheers.

[MichiganFocker]

@r6d2 Thanks for the reply...

Well I do use alot of avisynth filters in a script and undot and deen are some of them. Like I said I have only upgraded my programs. Still using the same filters when the movies were coming out awesome.

Been making SVCD's with up to 2 hours and 10 minutes on 1 CD and coming out great (only have done a hand full over 2hrs though).

Have done tons so far around the same time as predator...110 mins. and everyone has come out awesome. Someone actually gave me a AVISYNTH.INI and matrices.ini to replace the one in DVCD2SVCD. I use Notch=KVCD for the matrice in DVD2SVCD.

What is the max minutes you will put on 1CD?

Well if you want me to show you the avisynth script I will be glad to. Or if you want to see anything elese let me know. :D

Gonna try to consult with my friend who set me up this way in the frist place. Hopefully can get this sorted. Thanks again for the reply. :D

[r6d2]

Originally posted by MichiganFocker
What is the max minutes you will put on 1CD?As many as they fit with at least the quality I expect. I use D2Sroba Auto mode with Auto Q about 30.

You may want to see the guide on my sig for details.

[homerjay]

i have read & re-read yours and ddoggs posts and guides on the subject and searched all thru for a reason for this increase in base Q for films

i already use the KISS method in my preresize.avs ( have since i saw it here )

and as michiganfocker has already stated all we updated were the apps

now i know you guys are purists when it comes to quality but believe us the 1cd films are very watchable up to a tested 135 minutes at which point the end of the source is cut off/ignored

i am fully prepared to try anything suggested to resolve the situation and am going to start by a complete uninstall and reload of everything unless someone can suggest anything at all to try/check first

please help us get our 1cd svcd's back :scared: we dont want to have to switch to the dark side - kvcd :eek: :D

[tylo]

@MichiganFocker:

You have disabled the Sizing pass. If the mpv is oversized, there are two choices: either get a too big image, or cut away the oversized part. In the former case, you may be able to overburn the image, but often it's too big. In the latter case, you may only loose some credit text, so you can ignore the second (small) image.

- Cut last image : If opv sz > 0.3% oversize

You can control when to cut, and when not to, by editing D2SRoBa.ini and change cut_per_over. Setting it to 0.0 will always cut when you oversize.

To avoid all these worries, use Sizing pass when oversizing (with ReJig transcoder). If you don't want CCE resizing when the mpv is undersized, simply set a big number for the lower limit (e.g. 5), and 0 for the upper limit.

As r6d2 said, D2SRoBa is not responsible for the higher Qs you get. Usage of proper resizing (bicubic), AutoFitCD or FACAR cropping, and filtering will pull down Q.

/Add: homerjay: Also make sure to use proper matrices (e.g. Ultra low br for 1CD) and RB's templates in D2S's CCE advanced settings.

/Add2: I believe it is wise also to untick 'Adjust max video br' in D2SRoBa when doing 1CD encodes. Matrices suited for low br also demands a lower max br, so you must set that in D2S bitrate tab. Maybe anyone can elaborate on the effect of that?

[r6d2]

Originally posted by tylo
Also make sure to use proper matrices (e.g. Ultra low br for 1CD) and RB's templates in D2S's CCE advanced settings.I was thinking that is you reinstalled CCE, maybe your templates got removed or something? As Tylo said, make sure RB's templates are in place.
I believe it is wise also to untick 'Adjust max video br' in D2SRoBa when doing 1CD encodes. Matrices suited for low br also demands a lower max br, so you must set that in D2S bitrate tab.@Tylo, but you can also set a lower limit in D2Sroba and it will be adjusted in D2S, right?
Maybe anyone can elaborate on the effect of that? When you're using such a low BR (and hence a high Q) it actually does not matter very much if you lower the Max. And if you do to an extreme, you may get a lower Q but it would be a "false" lower Q, since the frames where Max is surpassed will have a higher quantization than expected.

This is the reason why BR(Q) tends to be the same with low values of Q when Max is set too low for the intended Q. If you leave Max at a high value (i.e. 9000), BR(Q) is much more regular, like a "pure" negative exponential. (I hope I did not get too Babylonian again.) :)

Also, never forget that some movies compress very differently. Notorious cases: SPR, BadBoys II, Scarface, and others, are difficult to compress so a high Q is something you have to live with at times, particularly if you are a 1 CD fan. :)

[homerjay]

thanks again tylo and r6d2 - your input is always helpful ;)

i have followed your advise during reinstalling my last known working ( as i want it to ) configuration
d2s121b3 / d2sroba315 / cce2.66 / eclcce 1.6 / facar 1.06.6

i reencoded at a fixed Q of 35 a two hour plus film i knew resulted in an undersized file using d2sroba previously

after encoding 20% i cancelled as it was already up to 200mb

i then applied the settings that i have used in the past to the qmf15b1.avs for detecting more high motion frames and set the resize method to 5 as i have had it in the past in the FACAR script, imported the registry settings in DDoggs ccetemplates.zip and selected progressive from the dropdown and repeated the process using the KVCD-NOTCH matrix again it reached 200mb around the 20% mark.

then i played back the Encoded_Video_CCE_PAL.mpv just to look at the quality which was what i expected but i finally noticed it was much darker - in fact i had to increase the brightness in bsplayer to 20 before i would say it was comparable to the original source.

so my new question is what would suddenly make my dvds play so much darker is it a codec problem that has somehow been introduced and what if anything can i do to resolve it :o

[homerjay]

ok im continuing to work on this to try and find a resolution

working with my previously stated setup i reconverted another film overnight which had a Q of 35 previously and came out with a final Q of 92.

a friend fastchef suggested adding the following into the FACAR script to try and deal with the brightness issue
LumaFilter(+10,1.1)
but it only served at the end to increase the filesize not decrease it

so i have now upgraded all component parts back to the latest versions and retried with similar high Q results

so please if anyone has seen anything similar either with your source material playing back darker or with your Q increasing since upgrading a particular part of the dvd2svcd process reply to this thread

thankyou

[homerjay]

ok im still chipping away at this - have got a little further having uninstalled my codecs and reinstalled the latest from http://www.cole2k.net/?display=CP-A my source seems brighter but still not quite there

i have now performed a few test as detailed below speeds included just for reference

source big wednesday - originally encoded at fixed Q35 - quality excellent

all latest versions of available s/w and filters installed

all using bach1 matrix for extra compressability ( originally kvcd notch )

all using FACAR which will be tested with various filters and parameters as below so if anyone has any suggestions please pass them on

thanks

-------------------------
test one - speed 1.72 - determined Q82

parameters changed in FACAR.avs

resizemethod = 1

-------------------------
test two - speed 1.70 - determined Q82

parameters changed in FACAR.avs

resizemethod = 5

unhashed 11=Import(!qmf15b1.avs)
unhashed 12=Import(!QMF_functions.avs)
unhashed 14=Import(!FACAR.avs)
--------------------------------------
test three - speed 0.92 - determined Q76

parameters changed in FACAR.avs

unhashed 13=Import(!preresize.avs)

parameters changed in preresize.avs

unhashed UnDot()
unhashed Deen()
---------------------------------------

now the third is a lot better than when i ran this before but is still just over double the Q of that i originally managed plus it looks crap

if someone is interested i can clip something from the original coversion and the same section from the new conversion for comparison and make them available...

below are latest d2sroba and dvd2svcd log files


--------------------------------------------------------
D2SRoBa v3.50, by Tylo
--------------------------------------------------------
- WIN_XP - AutoIt 3.0.101.0
- 2004-05-08 18:00:25
--------------------------------------------------------
- Output type : SVCD
- CD size : 800
- Number of CD's : 1 (max 32)
- Sample percent : 1.0
- Adjust % : 1.2
- Cond. sizing pass : No
- Cut last image : If opv sz > 0.3% oversize
- Audio after video : Yes (128 kbps)
- Adjust max br : No
- Clean previous : Yes
- Run mode : Normal
--------------------------------------------------------
- Waiting for CCE window ...
- Detected a CCE encode window
- Detected the movie encode window - shutting down
--------------------------------------------------------
- 2004-05-08 18:11:14
- Project dir: E:\DVD2SVCD\Movie\
- Config file: E:\DVD2SVCD\Movie\SVCD.INI
- Avisynth ini: E:\DVD2SVCD\Movie\AVISYNTH.INI
- Avisynth setup: [FACAR], cond. filter: None
- D2S ver. 1.2.2.1, CCE ver. 2.67.0.27
--------------------------------------------------------
- Movie length : 02:12:33 (198845 frames, 25 fps)
- D2S audio select : 128 + 0 kbps
- D2S video calc. : 1503 kbps, 2 CDs
- Select Ranges : every 1200, select 12 frames
- Sample frames : 1992
--------------------------------------------------------
Calculations:
- 1 CD: video_br=680 (680) audio_br=128+0 video_sz=676073000 fill=100% cbr=0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Using num CDs : 1
- Target mpv BR : 680 (max 2530) kbps, size=676073000 bytes
--------------------------------------------------------
Search for Q:
- Sample encode : Q=73: 698 kbps, err=2.6%, size=694664608, sample sz=6959048
- Sample encode : Q=76: 687 kbps, err=1%, size=683685809, sample sz=6849064
- Sample encode : Q=78: 680 kbps, err=0%, size=676645178, sample sz=6778532
- Determined Q : 76 = Round(78 + (0 - 1.2)/0.5)
--------------------------------------------------------
- 2004-05-08 18:16:38
- Start movie OPV encoding (Q 76)



WNASPI32.DLL 0.0.0.0
--------------------------------------------------------
- 08/05/2004 18:01:12
- DVD to SVCD Conversion
- DVD2SVCD ver. 1.2.2 build 1
--------------------------------------------------------
Initializing
- E:\BigWed\VTS_01_0.IFO
- E:\BigWed\VTS_01_1.vob
- E:\BigWed\VTS_01_2.vob
- E:\BigWed\VTS_01_3.vob
- E:\BigWed\VTS_01_4.vob
- E:\BigWed\VTS_01_5.vob
- E:\BigWed\VTS_01_6.vob
- E:\BigWed\VTS_01_7.vob
Initializing finished.

--------------------------------------------------------
- 08/05/2004 18:01:16
- DVD2AVI
--------------------------------------------------------
Creating DVD2AVI INI file:
- E:\DVD2SVCD\DVD2AVI\DVD2AVI.INI

Variable settings:
iDCT_Algorithm: 32-bit SSE MMX

Executing DVD2AVI.
Executing DVD2AVI. Commandline:
"E:\DVD2SVCD\DVD2AVI\DVD2AVI.exe" -CS=2 -YR=1 -OM=0 -EXIT -OF=[E:\DVD2SVCD\Movie\DVD2AVI_Project_file] -IF=[E:\BigWed\VTS_01_1.vob,E:\BigWed\VTS_01_2.vob,E:\BigWed\VTS_01_3.vob,E:\BigWed\VTS_01_4.vob,E:\BigWed\VTS_01_5.vob,E:\BigWed\VTS_01_6.vob,E:\BigWed\VTS_01_7.vob]
Framerate: 25000
DVD2AVI processing done.

--------------------------------------------------------
- 08/05/2004 18:07:05
- Free on drive E: 5861.13 mb
- AUDIO Extraction
--------------------------------------------------------
Found AC3 stream id: 0x80
Filename: E:\DVD2SVCD\Movie\Extracted_audio_1.ac3
Audio1 delay: -88 ms
Audio extraction finished.


Encoding Audio. Filename: E:\DVD2SVCD\Movie\Extracted_audio_1.ac3
Executing BeSweet. Commandline:
"E:\DVD2SVCD\Tylo\BeSweetLater.bat" -core( -input "E:\DVD2SVCD\Movie\Extracted_audio_1.ac3" -output "E:\DVD2SVCD\Movie\Encoded_audio_1.mp2" -logfile "E:\DVD2SVCD\Movie\Encoded_audio_1.log" ) -azid( -L -3db -l 0db -c normal -s surround -g max ) -ota( -d -88 ) -shibatch( --rate 44100 ) -2lame( -e -b 128 -m s )
Audio conversion of E:\DVD2SVCD\Movie\Extracted_audio_1.ac3 finished.

Audio conversion finished.

--------------------------------------------------------
- 08/05/2004 18:10:16
- Free on drive E: 5703.28 mb
- Video Encoding using Cinema Craft
--------------------------------------------------------

Executing Cinema Craft Encoder.
StreamSectors: 1668571576
AudioSectors: 129095876
VideoPAPO: 23847300
ScanOffsetBytes: 298268
SeqAligningBytes: 18956557
DVDBytes: 0
VideoEndHeader: 8
SubtitleSectors: 0
EmptySectors: 238.00
PictureSectors: 0.00
PureMPEGStream: 1496373567.83
Seconds: 7960.80
CalcMPEGStream: 1496373567.83
Frames: 198845
CDSize: 800.00
Number of CDs: 2
Cut point 796.00
Variable Settings:
Frames: 198845
Anti Noise Filter: Off
Passes: 2
Image Quality: 17
VAF file creation: On
Video Encoding Mode: One Pass VBR
Q. Factor: 35
Min. bitrate: 0
Max. bitrate: 1503
Aspect Ratio: 16:9 (borders added, encoded as 4:3)

---AVS Begin---
LoadPlugin("E:\DVD2SVCD\AVISYN~1.5PL\Mpeg2dec\MPEG2D~1.DLL")
MPEG2Source("E:\DVD2SVCD\Movie\DVD2AV~1.D2V")
ResizeTo=0 global ResizeMethod=1 DebugMode=0
#LumaFilter(+10,1.1)
GammaCorrection=1.0 Hue=0 Saturation=1.0 Brightness=0 Contrast=1.0
DetectedTop=72 DetectedHeight=434
OverscanH=0.05 OverscanV=0.05 BlockOptimization=0 AsoTV=1
global Bicubic_b_Value=0 global Bicubic_c_Value=0.6
DestinationWidth=480
DestinationHeight=432
DestinationTop=72
DestinationBottom=72
LoadPlugin("E:\DVD2SVCD\Avisynth2.5 Plugins\SimpleResize\SimpleResize.dll")
Import("E:\DVD2SVCD\Avisynth2.5 Plugins\FACAR\qmf15b1.avs")
Import("E:\DVD2SVCD\Avisynth2.5 Plugins\FACAR\QMF_functions.avs")
Import("E:\DVD2SVCD\Avisynth2.5 Plugins\FACAR\PreResize.avs")
Import("E:\DVD2SVCD\Avisynth2.5 Plugins\FACAR\FACAR.avs")
Import("E:\DVD2SVCD\Avisynth2.5 Plugins\FACAR\PostResize.avs")
Import("E:\DVD2SVCD\Movie\ResampleAudio.avs")
ResampleAudio(44100)
ConvertToYUY2()
----AVS End----

[MichiganFocker]

@tylo

Thanks for the reply. I went ahead and formated my puter and re-installed my OS. Long overdue for it.

I have went ahead and took your advice on checking transcoding and also unticked 'Adjust max video br' in D2SRoBa to let D2S do the bitrate. Going to also experiment with a few different matrices.

I actually always had transcoding checked before til I asked about why CCE does a 2 pass when I was encoding and homerjay said it was because I had transcoding checked and to untick it. Come to think about it after that is when I started getting the 2nd CD...duhhhh :D

Will post results after done later today. Thanks again :D

[DDogg]

homerjay, could you do me a big favor and try HybridFuPP? I think your test might show something positive and Fupp has worked damn hard on this. It needs to be evaluated by compression sensitive folks..maybe I should say compression challenged? :). What I did was rename it to HybridFuPP.avsi and put it in my plugin Dir along with the dlls it needs included in the zip (see avisynth forum). Then it becomes simple to do a HybridFuPP(preset="medium"). I used it on Gothica and it did beat kiss a fair bit. I'm more curious what you think about quality. It certainly has an interesting premise. Oh, yeah, forgot to say, it is a little, er, slower than kiss.

[homerjay]

thanks ddogg

but im getting an invalid floating point operation when doing a preview

i installed HybridFuPP under my avisynth plugins folder

then i unzipped all files in that folder directly into my avisynth plugins folder

then i added the command to my preresize.avs ( replacing deen )
HybridFuPP(preset="medium")

is there anything i missed ?

09.43 arse im getting that even after removing that line from the preresize.avs :( any suggestions ...

10.49 ok deleted my avisynth filters completely and reinstalled and no floating point error
going out for an hour or so - when i get back will unzip the HybridFuPP files one filter at a time and try to pinpoint where the error occurred ;)

12.57 after a little trial and error it would appear to be a problem with the motion.dll
the strangest part being that if i leave preresize.avs as it was and just unzip the motion.dll into the avisynth plugin folder it throws up this error
this was confirmed by just deleting it at which point the preview worked straight away :confused:

[DDogg]

homerjay, sorry man, I have not even tried to integrate HybridFuPP with FACAR. Perhaps just using it in a simple script might be easier? An example in the readme is:

Mpeg2Source("c:\test\test.d2v")
GripCrop(480, 576, overscan = 2, source_anamorphic = true, dest_anamorphic = false)
HybridFuPP(preset = "very high", width = GripFit_resize_width, height = GripFit_resize_height)
GripBorders()

[MichiganFocker]

Well happy day for me....all sorted on my end. After a re-install of my OS and all my programs it now works like a charm.

My latest movie which is The Hunt For Red October and 135 minutes. Came out awesome. I wasnt sure at all that it would being so long. Here's my log on it.

--------------------------------------------------------
D2SRoBa v3.50, by Tylo
--------------------------------------------------------
- WIN_XP - AutoIt 3.0.101.0
- 2004-05-10 23:53:26
--------------------------------------------------------
- Output type : SVCD
- CD size : 800
- Number of CD's : 1 (max 32)
- Sample percent : 2.0
- Adjust % : 0.8
- Cond. sizing pass : Yes (Transcode if opv sz > 0.1% oversize, and Q <= 40)
(Encode if opv sz < 2.5% undersize, or Q > 40)
- Audio after video : Yes (9% of video, min 128 kbps)
- Adjust max br : Yes (total: 2754, max avg offset: 300 kbps)
- Clean previous : Yes
- Run mode : Normal
--------------------------------------------------------
- Waiting for CCE window ...
- Detected a CCE encode window
- Detected a CCE encode window
- Detected the movie encode window - shutting down
--------------------------------------------------------
- 2004-05-11 00:11:43
- Project dir: C:\Program Files\DVD2SVCD\Movie\
- Config file: C:\Program Files\DVD2SVCD\Movie\SVCD.INI
- Avisynth ini: C:\Program Files\DVD2SVCD\Movie\AVISYNTH.INI
- Avisynth setup: [FACAR], cond. filter: None
- D2S ver. 1.2.2.1, CCE ver. 2.67.0.9
--------------------------------------------------------
- Movie length : 02:17:33 (197874 frames, 23.976 fps)
- D2S audio select : 128 + 0 kbps
- D2S video calc. : 651 kbps, 1 CDs
- Select Ranges : every 750, select 15 frames
- Sample frames : 3960
--------------------------------------------------------
Calculations:
- 1 CD: video_br=651 (651) audio_br=128+0 video_sz=671588119 fill=100% cbr=0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Using num CDs : 1
- Target mpv BR : 651 (max 2626) kbps, size=671588119 bytes
--------------------------------------------------------
Search for Q:
- Sample encode : Q=76: 497 kbps, err=-23.7%, size=513417871, sample sz=10274896
- Sample encode : Q=47: 643 kbps, err=-1.2%, size=664002784, sample sz=13288512
- Sample encode : Q=45: 661 kbps, err=1.5%, size=682001922, sample sz=13648724
- Sample compute : Q=46: 652 kbps, err=0.2%, using (br(Q45) + br(Q47))/2
- Determined Q : 46 = Round(46 + (0.2 - 0.8)/1.4)
--------------------------------------------------------
- 2004-05-11 00:21:38
- Start movie OPV encoding (Q 46)
- OPV pass result: 96.9% on target (650710760 / 671588119) Speed factor: 1.24
--------------------------------------------------------
- 2004-05-11 02:12:25
- Start VBR sizing pass (651 kbps)
- Sizing pass result: 99.8% on target (670534608 / 671588119) Speed factor: 1.3
--------------------------------------------------------
- 2004-05-11 03:57:55
- Executing: C:\Program Files\DVD2SVCD\BeSweet\BeSweet.exe -core( -input "C:\Program Files\DVD2SVCD\Movie\Extracted_audio_1.ac3" -output "C:\PROGRA~1\DVD2SVCD\Movie\Encoded_audio_1.mp2" -logfile "C:\PROGRA~1\DVD2SVCD\Movie\Encoded_audio_1.log" ) -azid( -L -3db -C 0db -S 0db -c normal -s surround -g max ) -ota( -d -67 ) -shibatch( --rate 44100 ) -2lame( -e -b 128 -m s )
--------------------------------------------------------
- 2004-05-11 04:34:58
- Recover DVD2SVCD
--------------------------------------------------------
- 2004-05-11 10:01:44
- D2SRoBa successfully finished
--------------------------------------------------------

[homerjay]

@mich lucky git :devil:

@ddogg i think its a bit more serious than trying to run it from within FACAR - sorry i misunderstood in the first case :rolleyes:

after the initial gung ho everything where it should be i went back to a working avisynth plugin setup and made sure i could do a preview then all i unzipped was the motion.dll and i got the error

i have since just as a test only unzipped the HybridFuPP file into my plugins folder and straight away without even referencing it i got the invalid floating point operation on preview :confused:

it is from what mich said begining to look like something somewhere is holding on even after uninstalling and reloading everything as one dll file shouldnt cause it to crash - should it ?

18.48 following on from that thought i have uninstalled all again and this time ran regsupreme to clean up the registry and reinstalled only the latest versions of all and at the moment i am encoding my test film at Q56 which is a marked improvement :D not 35 but definately an improvement

once its completed i will give HybridFuPP another go ddogg ;)

[DDogg]

one thing to check quickly. make sure warpsharp, smartdecimate, avisynth_c.dll and loadpluginex are NOT in your plugin dir. ALso, rename plugin dir and just put the stuff for hybridfupp in it to test. There are several nasty dlls floating around that cause strange and weird things to happen. Oh, plus make sure you are running the latest avisynth (http://cultact-server.novi.dk/kpo/avisynth/avs_cvs.html) from sh0dan's sig. I think some of the latest advanced scripts need it to function. It will not hurt d2s.

[homerjay]

a couple of those would have been in there yes
but not anymore - i havent installed any additional filters as yet - just what comes with DVD2SVCD & FACAR

and im on 2.54 but if you say 2.55 is safe i will get it and use it for the test :D

thanks for taking an interest will report results back later ;)

[FuPP]

Waiting for the results. ;)

Regards,
FuPP

[DDogg]

Howdy, FuPP, nice to see you here. Welcome.

[homerjay]

sorry guys - quick question where do i get gripfit from ? cant find it anywhere ? thanks

[DDogg]

attached. But hopefully we can get it working with FACAR now that you cleaned house :)

This is what I had in my plugin dir. Hopefully it is the correct one. Fupp would probably know for sure.

[homerjay]

your a star - ok will get on with the test now - thanks :)

[homerjay]

ok here we go main film is at 2% now so here are my findings before i fall asleep on my keyboard.

D2SROBA + FACAR with QMF = Q52 and just about passable quality
only problem i saw with preview and final .mpg is there is now a black border on LHS & RHS which im sure i never used to have
im sure this will be costing me Q will need to read the FAQ in the morning

D2SROBA + the following script
GripCrop(480, 576, overscan = 2, source_anamorphic = true, dest_anamorphic = false)
HybridFuPP(preset = "very high", width = GripFit_resize_width, height = GripFit_resize_height)
GripBorders()

= determined Q of 54 and it is actually a little faster than the previous method

will report back on final results in the am - g'night all :)

[DDogg]

Darn, wake up sleepyhead. Use low or medium. Very high does not compress as well. :)

[Holomatrix]

@MichiganFocker - Just wondering what script you used? (FACAR w/QMF?) That really is a low Q :) for that length. Have you tried doing 'Stuck On You' or 'Beyond Borders'? Q I got with thoes was around 54 for approx 110 min, thoes ones might just have not been to compressable.

I think I should take ddoggs advice and clean my Avisynth plug-in folder as well :) and do some testing. You guys got me going again on a low Q frenzi with all this talk on different matrices, clean plug-in folder and all, to drop Q :)

[homerjay]

@DDOGG (YAWN) sorry i was out of it last night still i thought 56 was good :)

will try it again shortly at MEDIUM just as soon as i find out why my pc was shut down and the conversion was incomplete last night - when i tried to recover i noticed muxing was missing from the list and yet the mpg was roughly the size i expected :confused:

if its not one thing its another :p

[homerjay]

mixed results with our test encode

--------------------------------------------------------
D2SRoBa v3.50, by Tylo
--------------------------------------------------------
- WIN_XP - AutoIt 3.0.101.0
- 2004-05-12 10:12:10
--------------------------------------------------------
- Output type : SVCD
- CD size : 800
- Number of CD's : 1 (max 32)
- Sample percent : 1.0
- Adjust % : 0.8
- Cond. sizing pass : No
- Cut last image : If opv sz > 0.3% oversize
- Audio after video : Yes (128 kbps)
- Adjust max br : No
- Clean previous : No
- Run mode : Crash recovery
--------------------------------------------------------
- Waiting for CCE window ...
- Detected a CCE encode window
- Detected the movie encode window - shutting down
--------------------------------------------------------
- 2004-05-12 10:13:22
- Project dir: E:\DVD2SVCD\Movie\
- Config file: e:\DVD2SVCD\Movie\SVCD.INI
- Avisynth ini: e:\DVD2SVCD\Movie\AVISYNTH.INI
- Avisynth setup: [HybridFuPP - Simple], cond. filter: None
- D2S ver. 1.2.2.1, CCE ver. 2.67.0.27
--------------------------------------------------------
- Movie length : 02:12:33 (198845 frames, 25 fps)
- D2S audio select : 128 + 0 kbps
- D2S video calc. : 1507 kbps, 2 CDs
- Select Ranges : every 1500, select 15 frames
- Sample frames : 1995
--------------------------------------------------------
Calculations:
- 1 CD: video_br=682 (682) audio_br=128+0 video_sz=678061450 fill=100% cbr=0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Using num CDs : 1
- Target mpv BR : 682 (max 2530) kbps, size=678061450 bytes
--------------------------------------------------------
Search for Q:
- Sample encode : Q=73: 578 kbps, err=-15.2%, size=575226391, sample sz=5771212
- Sample encode : Q=55: 656 kbps, err=-3.8%, size=652348748, sample sz=6544976
- Sample encode : Q=49: 694 kbps, err=1.8%, size=690784540, sample sz=6930600
- Sample encode : Q=51: 680 kbps, err=-0.3%, size=676966856, sample sz=6791968
- Determined Q : 50 = Round(51 + (-0.3 - 0.8)/1)
--------------------------------------------------------
- 2004-05-12 10:21:19
- Start movie OPV encoding (Q 50)
- OPV pass result: 105.1% on target (712511056 / 678061450) Speed factor: 0.88
- Oversized: Cutting last CD image off by 5.1%
--------------------------------------------------------
- 2004-05-12 12:51:47
- Executing: e:\DVD2SVCD\BeSweet\BeSweet.exe -core( -input "e:\DVD2SVCD\Movie\Extracted_audio_1.ac3" -output "e:\DVD2SVCD\Movie\Encoded_audio_1.mp2" -logfile "e:\DVD2SVCD\Movie\Encoded_audio_1.log" ) -azid( -L -3db -l 0db -c normal -s surround -g max ) -ota( -d -88 ) -2lame( -e -b 128 -m s )
--------------------------------------------------------
- 2004-05-12 12:58:32
- Recover DVD2SVCD
--------------------------------------------------------
- 2004-05-12 13:25:41
- D2SRoBa successfully finished
--------------------------------------------------------

but i would have to says that the quality is not as good as the FACAR+QMF encode which came out at Q56

also a prob with bbmpeg cutting which i will post separately

[Holomatrix]

Cleaned out plugin dir and tried this matrix rather than KVCD(usual).
Will try bach1 next but only dropped a couple Q points with this one.

Jawor's 1CD Matrix

8,16,20,23,27,30,32,34
16,16,23,24,29,31,34,38
20,23,27,28,30,35,38,40
23,24,28,29,35,39,40,44
27,29,30,35,39,42,45,50
30,31,35,39,42,46,54,59
32,34,38,40,45,54,62,72
34,38,40,44,50,59,72,84

17,19,21,23,25,27,29,31
19,21,23,25,27,28,29,33
21,23,25,27,29,30,32,37
23,25,27,30,31,34,40,45
25,27,29,31,38,46,54,60
27,28,30,34,46,58,72,74
29,29,32,40,54,72,90,100
31,33,37,45,60,74,100,124

[DDogg]

btw, check mrbass's site (www.mrbass.org) for CQME. It is a utility that will allow you to open a matrix file and 'inject' it into the CCE registry settings or TMPG ini. Yeah, no more typing in little weird numbers, plus it does a lot of other things matrix wise. The original thread is here (http://forum.doom9.org/showthread.php?s=&threadid=73666).

[FuPP]

quality is not as good as the FACAR+QMF encode which came out at Q56

Why ?

1/ HybridFuPP/Medium is less sharp ?
2/ is more blocky ?

If 1/, you can try with preset = "high" (better compressibility than "very high", and sharper than "medium")

Choosing the right preset depends a lot on the source.

Cheers,
FuPP

[homerjay]

it was definately more blocky around heads/objects + in the surf

but if it helps i will give it a go on high later

[DDogg]

lol, another run just using simple undot().Deen() as a benchmark (without qmf) would be great :D Poor homerjay, I can see your eyes rolling up in your head :)

[FuPP]

Originally posted by homerjay
it was definately more blocky around heads/objects + in the surf


Blocky or ringing (ie around edges)? if ringing, you can try HybridFuPP(preset = "low", dering=true).

or try even "very low". Using higher preset won't help in that case :)

FuPP

[homerjay]

ok if you can give me the full command line i will start it running shortly ;)is the one below correct ??

HybridFuPP(preset = "low", dering=true, width = GripFit_resize_width, height = GripFit_resize_height)

also is there anything i can do to stop narrow black borders from appearing to the left and right hand sides.

im quite happy to try anything you suggest if it gets me a better quality 1cd svcd, conversion time is not really an issue :)

[FuPP]

If you want to use gripfit + HybridFupp, you could use the following command line :

GripCrop(480, 576, overscan = 0, source_anamorphic=true, dest_anamorphic=true)
HybridFuPP(preset = "low", dering=true, width = GripFit_resize_width, height = GripFit_resize_height)
GripBorders()

This one won't create left and right borders. Btw, you must be aware that these borders won't be visible on a TV set, but enable to save some precious bits (if you want them back, use parameter overscan = 2 with GripCrop)

FuPP

[homerjay]

ok underway - thanks for explaining the black borders - i will keep them then LOL

[homerjay]

i think we have a winner :)

picture is a lot better and the overall Q has been reduced to 46

you can put me down for testing any improvements FuPP youve created a nice tool here - thankyou ;D

will try tomorrow with another film and post results in your original thread ;)

i would just like to thank everyone for their help and hope some of the points that have come out of this can be of use to others



--------------------------------------------------------
D2SRoBa v3.50, by Tylo
--------------------------------------------------------
- WIN_XP - AutoIt 3.0.101.0
- 2004-05-12 21:04:37
--------------------------------------------------------
- Output type : SVCD
- CD size : 800
- Number of CD's : 1 (max 32)
- Sample percent : 1.0
- Adjust % : 0.8
- Cond. sizing pass : No
- Cut last image : If opv sz > 0.3% oversize
- Audio after video : Yes (128 kbps)
- Adjust max br : No
- Clean previous : No
- Run mode : Normal
--------------------------------------------------------
- Waiting for CCE window ...
- Detected a CCE encode window
- Detected the movie encode window - shutting down
--------------------------------------------------------
- 2004-05-12 21:12:09
- Project dir: e:\DVD2SVCD\Movie\
- Config file: e:\DVD2SVCD\Movie\SVCD.INI
- Avisynth ini: e:\DVD2SVCD\Movie\AVISYNTH.INI
- Avisynth setup: [HybridFuPP - Simple], cond. filter: None
- D2S ver. 1.2.2.1, CCE ver. 2.67.0.27
--------------------------------------------------------
- Movie length : 02:12:33 (198845 frames, 25 fps)
- D2S audio select : 128 + 0 kbps
- D2S video calc. : 1507 kbps, 2 CDs
- Select Ranges : every 1500, select 15 frames
- Sample frames : 1995
--------------------------------------------------------
Calculations:
- 1 CD: video_br=682 (682) audio_br=128+0 video_sz=678061450 fill=100% cbr=0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Using num CDs : 1
- Target mpv BR : 682 (max 2530) kbps, size=678061450 bytes
--------------------------------------------------------
Search for Q:
- Sample encode : Q=73: 559 kbps, err=-18%, size=555968228, sample sz=5577996
- Sample encode : Q=52: 646 kbps, err=-5.3%, size=642340914, sample sz=6444568
- Sample encode : Q=43: 706 kbps, err=3.5%, size=702741553, sample sz=7050564
- Sample encode : Q=47: 677 kbps, err=-0.7%, size=673430504, sample sz=6756488
- Sample encode : Q=46: Skipped
- Determined Q : 46 = Round(47 + (-0.7 - 0.8)/1.1)
--------------------------------------------------------
- 2004-05-12 21:20:05
- Start movie OPV encoding (Q 46)
- OPV pass result: 104.4% on target (708230876 / 678061450) Speed factor: 0.85
- Oversized: Cutting last CD image off by 4.4%
--------------------------------------------------------
- 2004-05-12 23:56:09
- Executing: e:\DVD2SVCD\BeSweet\BeSweet.exe -core( -input "e:\DVD2SVCD\Movie\Extracted_audio_1.ac3" -output "e:\DVD2SVCD\Movie\Encoded_audio_1.mp2" -logfile "e:\DVD2SVCD\Movie\Encoded_audio_1.log" ) -azid( -L -3db -l 0db -c normal -s surround -g max ) -ota( -d -88 ) -2lame( -e -b 128 -m s )
--------------------------------------------------------
- 2004-05-13 00:03:27
- Recover DVD2SVCD
--------------------------------------------------------
- 2004-05-13 00:08:54
- D2SRoBa successfully finished
--------------------------------------------------------

WNASPI32.DLL 0.0.0.0
--------------------------------------------------------
- 12/05/2004 21:05:11
- DVD to SVCD Conversion
- DVD2SVCD ver. 1.2.2 build 1
--------------------------------------------------------
Initializing
- E:\BigWed\VTS_01_0.IFO
- E:\BigWed\VTS_01_1.vob
- E:\BigWed\VTS_01_2.vob
- E:\BigWed\VTS_01_3.vob
- E:\BigWed\VTS_01_4.vob
- E:\BigWed\VTS_01_5.vob
- E:\BigWed\VTS_01_6.vob
- E:\BigWed\VTS_01_7.vob
Initializing finished.

--------------------------------------------------------
- 12/05/2004 21:05:14
- DVD2AVI
--------------------------------------------------------
Creating DVD2AVI INI file:
- e:\DVD2SVCD\DVD2AVI\DVD2AVI.INI

Variable settings:
iDCT_Algorithm: 32-bit SSE MMX

Executing DVD2AVI.
Executing DVD2AVI. Commandline:
"e:\DVD2SVCD\DVD2AVI\DVD2AVI.exe" -CS=2 -YR=1 -OM=0 -EXIT -OF=[e:\DVD2SVCD\Movie\DVD2AVI_Project_file] -IF=[E:\BigWed\VTS_01_1.vob,E:\BigWed\VTS_01_2.vob,E:\BigWed\VTS_01_3.vob,E:\BigWed\VTS_01_4.vob,E:\BigWed\VTS_01_5.vob,E:\BigWed\VTS_01_6.vob,E:\BigWed\VTS_01_7.vob]
Framerate: 25000
DVD2AVI processing done.

--------------------------------------------------------
- 12/05/2004 21:08:09
- Free on drive e: 6498.61 mb
- AUDIO Extraction
--------------------------------------------------------
Found AC3 stream id: 0x80
Filename: e:\DVD2SVCD\Movie\Extracted_audio_1.ac3
Audio1 delay: -88 ms
Audio extraction finished.


Encoding Audio. Filename: e:\DVD2SVCD\Movie\Extracted_audio_1.ac3
Executing BeSweet. Commandline:
"E:\DVD2SVCD\Tylo\BeSweetLater.bat" -core( -input "e:\DVD2SVCD\Movie\Extracted_audio_1.ac3" -output "e:\DVD2SVCD\Movie\Encoded_audio_1.mp2" -logfile "e:\DVD2SVCD\Movie\Encoded_audio_1.log" ) -azid( -L -3db -l 0db -c normal -s surround -g max ) -ota( -d -88 ) -2lame( -e -b 128 -m s )
Audio conversion of e:\DVD2SVCD\Movie\Extracted_audio_1.ac3 finished.

Audio conversion finished.

--------------------------------------------------------
- 12/05/2004 21:11:11
- Free on drive e: 6340.77 mb
- Video Encoding using Cinema Craft
--------------------------------------------------------

Executing Cinema Craft Encoder.
StreamSectors: 1668571576
AudioSectors: 129112144
VideoPAPO: 23847048
ScanOffsetBytes: 238614
SeqAligningBytes: 15165245
DVDBytes: 0
VideoEndHeader: 8
SubtitleSectors: 0
EmptySectors: 238.00
PictureSectors: 0.00
PureMPEGStream: 1500208516.67
Seconds: 7961.80
CalcMPEGStream: 1500208516.67
Frames: 198845
CDSize: 800.00
Number of CDs: 2
Cut point 796.00
Variable Settings:
Frames: 198845
Anti Noise Filter: Off
Passes: 4
Image Quality: 17
VAF file creation: On
Video Encoding Mode: One Pass VBR
Q. Factor: 5
Min. bitrate: 0
Max. bitrate: 1507
Aspect Ratio: 16:9 (borders added, encoded as 4:3)

---AVS Begin---
LoadPlugin("e:\DVD2SVCD\AVISYN~1.5PL\Mpeg2dec\MPEG2D~1.DLL")
MPEG2Source("e:\DVD2SVCD\Movie\DVD2AV~1.D2V")
GripCrop(480, 576, overscan = 2, source_anamorphic = true, dest_anamorphic = false)
HybridFuPP(preset = "low", dering=true, width = GripFit_resize_width, height = GripFit_resize_height)
GripBorders()
Import("e:\DVD2SVCD\Movie\ResampleAudio.avs")
ResampleAudio(44100)
ConvertToYUY2()
----AVS End----

--------------------------------------------------------
- 13/05/2004 00:03:32
- DVD to SVCD Conversion
- DVD2SVCD ver. 1.2.2 build 1
--------------------------------------------------------
Initializing
-
Initializing finished.

--------------------------------------------------------
- 13/05/2004 00:03:33
- Free on drive e: 5486.67 mb
- Restarting from: Muxing and Cutting
--------------------------------------------------------
- 13/05/2004 00:03:35
- Free on drive e: 5486.67 mb
- Converting Pictures from ES to PS
--------------------------------------------------------
Saving bbMPEG settings: e:\DVD2SVCD\bbMPEG\default.ini
--------------------------------------------------------
- 13/05/2004 00:03:37
- Free on drive e: 5486.67 mb
- Converting Pictures from ES to PS
--------------------------------------------------------
Saving bbMPEG settings: e:\DVD2SVCD\bbMPEG\default.ini
--------------------------------------------------------
- 13/05/2004 00:03:39
- Free on drive e: 5486.67 mb
- Converting Pictures from ES to PS
--------------------------------------------------------
Saving bbMPEG settings: e:\DVD2SVCD\bbMPEG\default.ini
--------------------------------------------------------
- 13/05/2004 00:03:42
- Free on drive e: 5486.67 mb
- Multiplexing and cutting
--------------------------------------------------------
Saving bbMPEG settings: e:\DVD2SVCD\bbMPEG\default.ini
Offset in Seconds: 1
Executing bbMPEG.
Variable Settings:
Movie offset: 1 seconds
Cut point: 796 mb

Executing RunbbMPEG. Commandline:
"e:\DVD2SVCD\bbMPEG\RunbbMPEG.exe" "e:\DVD2SVCD\Movie\bbMPEG_Muxed_File.mpg"
Length of e:\DVD2SVCD\Movie\bbMPEG_Muxed_File00.mpg
6891 Seconds
Offset in Seconds: 6892
Executing bbMPEG.
Variable Settings:
Movie offset: 1 seconds
Cut point: 796 mb

Executing RunbbMPEG. Commandline:
"e:\DVD2SVCD\bbMPEG\RunbbMPEG.exe" "e:\DVD2SVCD\Movie\bbMPEG_Muxed_File.mpg"
Multiplexing and cutting finished.

Image creation finished.

--------------------------------------------------------
- 13/05/2004 00:08:51
- Free on drive e: 4598.53 mb
- Conversion done!
--------------------------------------------------------


had the problem with bbmpeg again at the point where it went over 814mb on the credits but by clicking suspend and then no it created the second mpeg - but thats another issue ;)

[Holomatrix]

Trying this script:

GripCrop(480, 480, overscan = 2, source_anamorphic=true, dest_anamorphic=true)
HybridFuPP(preset = "low", dering=true, width = GripFit_resize_width, height = GripFit_resize_height)
GripBorders()

but the video is streched vertically. How can I use HybridFuPP by it self to get the true/actual video size? HybridFuPP by it's self stretches the video to. Here is the actual size when useing auto detect borders;
Left 0, Top 56, Right 719, Bottom 419, Width 720, Height 364.
Thanks

[homerjay]

just a guess at this stage but is it

source_anamorphic=true, dest_anamorphic=true should be false ?

[Holomatrix]

The movie is a 2.35:1 Anamorphic Widescreen and not a 16:9 1.85:1 widescreen which I think is what the plugin is converting it to. So that's why I put 'source_anamorphic=true, dest_anamorphic=true' is that wrong? Don't have much experience with Gripfit.
Thanks

[FuPP]

posting "dest_anamorphic=true", I guessed the destination was anamorphic (don't ask me why :p ). For 4:3 aspect ratio, use "dest_anamorphic=false".

Be aware that croping and adding black borders is not a part of HybridFuPP. Moreover, the destination size is normally input manually.

Though, the use of GripFit (ie GripCrop + GripBorders) enables the use of global variables GripFit_resize_width and GripFit_resize_height to get automatic resizing calculations.

[Holomatrix]

sorry my mistake, I was thinking there for a sec that anamorphic was the really wide/thicker top and bottom black bars movie which I'm wrong. The movie is non-anamorphic (really skinny) :)

EDIT: so , GripFit_resize_width and GripFit_resize_height takes the place of the gripcrop + gripborders and uses DVD2SVCD's values? (SVCD 480X480)