Modification and optimization of robshot method and its possible use with DVD2SVCD

[Holomatrix]

Ok, thanks, I just wasn't sure what to use with your program. I know in the readme it says you can use 1pass VBR but didn't notice if it mentioned to use the vaf or not.
That is all :)

[tylo]

Originally posted by Holomatrix
Ok, thanks, I just wasn't sure what to use with your program. I know in the readme it says you can use 1pass VBR but didn't notice if it mentioned to use the vaf or not.
That is all :)
No problem,
And you're right, its not mentioned anywhere.

@ MictXP: Found what causes your problems. At present, d2sroba doesn't support use of title picture, change cd pic. etc. which causes multiple avs entries in the ecl files! Use "No PCB" or turn these off. I'm working on a fix for this.

have a nice eastern
tylo

[arlsair]

Originally posted by tylo
@ MictXP: Found what causes your problems. At present, d2sroba doesn't support use of title picture, change cd pic. etc. which causes multiple avs entries in the ecl files! Use "No PCB" or turn these off. I'm working on a fix for this.
You can also .mpg files as title pictures. Just use VCDEasy to create them. In DVD2SVCD just select them as you select your bitmaps

[MictXP]

Ok, it's working great now. Now I'm going to test in on a full movie :D
I still have one question, though: earlier, you wrote that "Ticking off "Create vaf file" will only obscure more I think," meaning that you should leave Create vaf file on, right? Then later you wrote that "with One-pass vbr, there is no need to create vaf," meaning that, using your program, one wouldn't need to create the vaf file, right? So, now I'm just confused.
I'm testing a movie without the vaf file right now, but should I instead let CCE create a vaf file? Thank you!

Update: movie testing finished. Created Shrek, a 1 CD copy, with a Q setting of 30. It actually looks pretty decent, and the whole process took 2 hours. But I have to ask: what Q setting do other people use?

[Crackhead]

hi
I've tested ur prog v1.2 and it worked fine(i've got some other problems in teh last few days so I couldn't test it any further)
I've encoded Three Kings, it created 3CDs.
My normal calculation in D2S would create 2CDs with an avg bitrate of 1900. So I would say there was no need to use ur prog cause there was no quality difference for the eye between the 2 encoded films.
I've used Q setting of 32!
So now my question: Can someone tell me which min avg bitrate is according to which Q setting?:confused:
I think 32 is about min 2000kbps, but I don't know the steps to 1 point higher or lower!
thx for ur help!
PS. Excuse my bad english, i still learn!:D

[arlsair]

First, sorry for my poor english.

I think the Q-factor has nothing directly to do with the bitrate. The Q stands for quantisation and so the Q-factor is equal to a special quantisation. And the quantisation is dependent to the compressablitity of the movie.
Of course, when you have more bitrate, the quantisation is lower (better).

[tylo]

New release - D2Sroba 1.7

- support creation of Title pic., ChangeCD pic., and so forth.
- minor tweaks

tylo

http://home.no.net/tylo/ - DVD2SVCD plugins: D2Sroba + vStrip for DVDDecrypter

[JoLander]

Hi all,
I just got aware of this tool and try now to figure out how it can be helpful for me.
Because I use a very slow machine for my encodings (PIII, 600 MHz) it take sometimes veeeery long to encode a movie in good quality. I always want to have as less CD's as possible with best quality (usually 2 CD's for a 'normal' movie).
With this in mind, I switched off the function of CD calculation and used the DVD2SVCS settings instead. Everything worked fine, D2SRoBa made some nice test-conversions and found some values different from the defaults. In the D2SRoBa Log I found a message that the original values now replaced by the calculated ones. The following Recovery and the encoding has been started well, but instead of using the replaced values calculated by D2SRoBa the original values specified in DVD2SVCD are used.
In the project folder I found a file CCE_Project_File.ecl.new which really contains the new values, but there was also a file CCE_Project_File.ecl which has been created during the Crash recovery with the old values.
Earlier in this thread there was a statement, that the calculated values are replaced in the DVD2SVCD.INI file. I have not found this replaced information. The DVD2SVCD.INI was untouched. Is it possible that the replacement only works if D2SRoBa is setup to calculate the number of CD's??

I cancelled the encoding and used the CCE_Project_File.ecl.new as input and the result was fine.

I'm using DVD2SVCD 1.1.3B2, D2SRoBa 1.8 and CCE 2.66
Greetz,
Joe

[tylo]

From your description, I can tell that D2Sroba most likely closed, or something else went wrong just after the recover phase started. The .ecl.new file should have been replaced with the .ecl file (and the .ecl.new file deleted).

Anyway, D2SRoBa v2.0.1 - http://home.no.net/tylo/ does not have this particular problem, and is hopefully more robust. Upgrade! This version does the CCE encoding itself, and the recover will start from muxing (or pulldown if NTSC). The modified project file is now named CCE_Project_file_MOVIE.ecl (the movie only).

You may want to try the automatic Bitrate tweaking feature in this release. (Only for multipass VBR, CCE 2.66+ and EclCCE).

Please, report problems with this version (or how great it is) here.

tylo

[JoLander]

Hi Tylo,
thanks for the quick response and for the even quicker new version.
I will download and test the version immediately.
It seems that you are right with the assumption that D2Sroba has been stopped working after the recovery has started, because the same problem appear when I let D2Sroba calculate the number of CD's. I also tested it on 2 Machines (Win2000 and WinXP) with the same result.
The only odd thing is that when the recovery is started the D2Sroba icon still stays in the System Tray.

Nevertheless, even with some manual work your tool saved me a lot of time because I received very good results even with 1-pass vbr encoding. I'm sure I will continue using this tool in the future.
I will post my experience with the new version asap.
Thanks for your work.
Joe

[telemike]

Any chance of making a standalone program that can calculate the Q factor and bitrate needed to fill "x" number of cd's without having to run thru the whole dvd2svcd program?

[DDogg]

telemike, you took the words out of my mouth. While I very much appreciate the work done on this plugin, I find it aggravating to use. That may well be because I misunderstand its use, but I do not find it or the docs intuitive and it is just easier to do it the way I have always done it.

What I would really like is a 1-pass VBR diagnostic tool that would tell me:

If you use 1 cds you will have an approximate average Q of X
If you use 2 cds you will have an approximate average Q of X
If you use 3 cds you will have an approximate average Q of X
If you use 4 cds you will have an approximate average Q of X

I want to make the decision how many CDs will be used but would still like the option to use 1 pass VBR.

[tylo]

@DDogg
Ok, I admit that the documentation is really bad, and probably confusing, partly because it is not updated. :o I have planned for some time to replace it altogether. Onto your requests:

a) When you unselect "Compute num CD's", it let DVD2SVCD choose the number of CD's in the regular way, based on time intervals only. (It would be very easy for me to add an input box to set num of CD's instead, however). D2Sroba will now compute a Q that will fill those CD's. So, yes - you can decide how many CDs will be used and still have the option to use 1 pass VBR.

b) If you want to compute the Q for 1, 2, 3, and 4 CD's, it will require a binary search (Bach's proposal) for each, resulting in 4*6 = 24 encodings of 2% of the source, which would take more than half the time of the vaf-creation in a multipass encoding. I guess I could reduce the time (reuse encodings for Q values) to, say 30% of the full vaf-creation pass -- but would it be worth it then? Another thing: When would you use the the four Q values? They cannot be computed before ripping and DVD2AVI has done their job, so you would need to break the job in two.

@telemike
That is a good idea. But it requires ripping and DVD2AVI + creation of the exact same avs file that you would use in you final encoding -- otherwise the estimations would be worthless. However if you thougth of using it for AVI / DV convertion it makes a lot of sense. And, yes for that purpose, there is a good chance I will make such a program. (I actually done quite a bit on it) :)

PS: I've just upgraded my system to an XP2600+, and D2Sroba doesn't seem to work correct for me anymore. I will make an update in the coming week, and post on the progress of the new tool. Your inputs are invaluable to make it useful.

[DDogg]

tylo, don't take me wrong. As said I really do appreciate all your effort very sincerely. I think the legwork you have done will end up being a building block for yourself and others. Yep, doc suck :D, but saying that means I should say, "Need any help?" . A person can't really beat ya up unless they are willing to help solve the thing. Maybe I actually can help on that down the road.

As to you other, actually a person should only need two choices, 1 or 2, 2 or 3, 3 or 4 so that just cut the prediction time a bunch. I do hope you would consider doing this as I don't want to start a 1-pass vbr encode unless I have at least a general idea of what the Avg Q will be when I finish. That would really make the project start falling together IMHO and make a little more sense.

[tylo]

No offence taken - I appreciate any feedback. People shouldn't be so afraid of giving critisism to developers. For sure, we never get it right the first time. :)

Here's what I may do:
1. Add an option to compute the Q.factor for 1-2, 2-3, or 3-4 CD's.
2. Present the results of the computations, and let you choose number of CD's. (maybe with a timeout...)
These two steps are straightforward to implement.

I have an unpublished command line app:
CCEAut <file.avi | @avifiles.lst> <profilename>
The profilename referes to a user defined section in the ini-file which may look something like this:

[Profile SVCD]
audioin_format=wav
audioout_format=mp2
besweet_args=-core( -input $input -output $output ) -2lame( -e -b 192 )
cce_template=SVCD
avisynth_script=LoadPlugin("C:\Program Files\CCEAut\Avisynth\MPEG2Dec.dll") LoadPlugin("C:\Program Files\CCEAut\Avisynth\SimpleResize.dll") AVISource($input,False) ConvertToYUY2() SimpleResize(480,576) ResampleAudio(44100)
mplex_args=-f 4 -V -o $output $inputvideo $inputaudio

Basically, it let you encode avi-files, and select which cce template they should use, and run it though the avs-file that will be created from the avisynth_script key. Optionally it can extract audio(with Avi2wav) reencode (besweet) and multiplex (mplex). Incidently, it works with CCE 2.66+ even without EclCCE, because it uses the template functionality instead of the ecl file format.

Should I make this the basis for a power command line app that can include OPV (Bach's method), and optionally do ripping and dvd2avi?

@DDogg - Thanks for offering help - I'll take you up on that.

[tylo]

Sorry for my lengthy replies, but...

DDogg wrote: I don't want to start a 1-pass vbr encode unless I have at least a general idea of what the Avg Q will be when I finish.

Well, you *do* have a general idea when using D2Sroba: The Q factor will always be equal or better (lower) than the Q that you specified - given that you also tick on the "Compute num CD's".

The question is: Do you want a fixed minimum quality or a fixed number of CD's? You can't have both! (Actually DVD2SVCD tries to do the former - it varies the number of CD's based on length to keep the quality about constant for any movie, but that is very inaccurate).

Example. If you were given:
2CDs: Q=38, avg brate=1650
3CDs: Q=25, avg brate=2250

What would you choose? If you select 2CD's you should set e.g. Q=40 in D2Sroba, otherwise set Q < 38. You would never need to change that, unless you change your quality standards! :D

[DDogg]

Here's what I may do:
1. Add an option to compute the Q.factor for 1-2, 2-3, or 3-4 CD's.
2. Present the results of the computations, and let you choose number of CD's. (maybe with a timeout...)
These two steps are straightforward to implement.
That would be great!

Example. If you were given:
2CDs: Q=38, avg brate=1650
3CDs: Q=25, avg brate=2250
I would then like to just pick '3' and not worry about much else :)

[DDogg]

deleted due to massive confusion on my part :)

[tylo]

Okee, new release of D2Sroba 2.1.3 is out!

Features
- Predictable one-pass VBR encoding with CCE - fills last CD to the edge
- Automatic setting of optimal number of CD's, based on prefered Q. factor
- Input of exact number of CD's (overrides DVD2SVCD)
- Input of a range of number of CD's (see below).
- Automatic credits bitrate tweaking - available through RB's EclCCE.
- Status window - tells what's going on
- Works with CCE 2.50 and CCE 2.66+.

What's New
- Added setting of Num CDs + ranges
- Added Status Window
- Improved GUI
- Improved DOC

Comments are welcome!

Get it at http://home.no.net/tylo/

[arlsair]

You didn`t update the website, but you can download it directly.
http://home.no.net/tylo/setup_D2Sroba213.exe

[arlsair]

I get an error:
D2Sroba terminate DVD2SVCD, start the CCE and start to encode, I get the following error message:
SYNTAX: Radiobox <WindowTitel> <WindowText> <Text> <"Item1;Item2.ItemX"> <Select>

P.S.: The Titel "WindowTitel" sounds German not English. Is that correct so ?

[DDogg]

:D:D
tylo, you outdid yourself this time. The user interface makes so much more sense now (at least to me). The plugin is working flawlessly on a 2 disk 1 pass vbr encode although I have not given it too much of a workout yet. I may very well never use anything else now. Couple of clarification questions for you:

1> Can you expand on the q tweak function. I got a slightly small second disk. WHat action would you suggest?
2> Expand on why/what/methodology of the multipass vbr and how it differs from the standard multipass in d2s.
3> If the titles bitrate tweak is checked the one pass is unchecked. Does this mean you can not use the two together?

Note: I think D2S has a bug in that it is passing its predicted bitrate to CCE as the max value. I sure thought a 1pass encode CCE encode should be passed the MAX like 2520, or in my case with 160 audio 2584 as max and 300 as min. Tylo, are you looking at this?

I don't worry about it now because I set up D2S for a locked 300-2584 bitrate, but I wonder if you should not have min and max in your plugin? Just throwing that out I have not really thought it through.

[tylo]

@arlsair: You found a combo I hadn't intended: I guess you selected a range of CDs (2-3?) and then selected 'Bitrate Tweak' or 'Multipass vbr'. For now, only use ranges together with '1-pass vbr RoBa' - I'll make that a legal combo later.

@DDogg: Thanks! Yepp, I am quite pleased myself.

1> You can actually type inn values outside the ranges in all the combo boxes, and they're even sticky! (except Num CDs). E.g. you can put in 1-4 in the Num CDs box. :cool:. About 'Adjust Q': You will notice that the size estimation will vary quite a bit, so I wouln't go too far. One Q value may represent 20-40MB.. (Edit): You may increase the Sample % instead. That should increase the prediction accuracy.

2> OK. It is actually NO different than the standard d2s vbr. Use this selection if you only want to take advantage of the Num CDs=m-n/auto functions. The auto num CDs is actually much more accurate and convenient than setting time intervals in the d2s' bitrate tab: Set the 'Prefered Q' and you'll get minimum num CDs that gives at least this Q. BTW, peek on the tooltips.

3> Correct. Bitrate Tweaking works only with multipass VBR - Check out RB's EclCCE documentation.

D2Sroba already fixes that d2s-opv-max-bug: I replace the opv_max value with the proper one, as you describe. Darn, where is that SAINT Smile :p

[arlsair]

Originally posted by tylo
@arlsair: You found a combo I hadn't intended: I guess you selected a range of CDs (2-3?) and then selected 'Bitrate Tweak' or 'Multipass vbr'. For now, only use ranges together with '1-pass vbr RoBa' - I'll make that a legal combo later.

You`re right, 2-3 CDs with Multipass VBR. I will keep in mind. Thanks.

P.S.: What do you think is the best way to test which q-factor fits the personal mininum quality standard ?
When I use Roba I can`t set a fixed q-factor, it will use all space on CD. Have I to tweak the quantisation directly in the CCE ?

[tylo]

I've thought about adding a new mode where you can rip one chapter (using internal ripper, or my DVDDecrypter plugin), and determine this Q. We'll see.

Until then, here is a way to find your prefered Q manually:
1. Use DVD2SVCD. Set encoder mode to OPV. Q=16
2. Rip only a chapter. Break D2S after encoding has started
3. Edit the .ecl file: set opv_max_brate to 2530
4. Run the .ecl file through CCE.
5. Watch the .mpv in media player. If your not happy with the result anylonger -> you're done: prefered q = current q - 2
6. Edit the .ecl file: increase opv_q_factor(I think), by 4
7. repeat from step 4

[arlsair]

Yeah, that should work. Thanks.

[tylo]

Update D2Sroba v2.1.5
- Fixed the error reported by arlsair + some cosmetic and doc. updates
http://home.no.net/tylo/

[telemike]

Tylo- Does your program take the baseline "Q" value and adjust up or down or both? If I start out a "Q" of 15, will the programincreae it to 20 if needed or down to 5 if needed? THanks.

[tylo]

The "baseline Q" is, I guess, what I called "Prefered Q" (hmm, maybe your name is more descriptive?). Anyway, it takes it down, or keeps it when you use Num CDs=auto. Otherwise, it may go both ways. In fact, the "Prefered Q" is only relevant when using "Num CDs=auto", Otherwise the algorithm always start out with Q=32. Note that 'Adjust Q' also take it down, even though you put in a positive number. Hope that helps.

[DDogg]

I would suggest anybody using this plugin use either a specific number of Cd's if they have a good 'feel' for where they want it or use the "2-3" as a requester will come up with the information you need to make a final decision.

The one pass vbr section of the plugin really works well although I am finding it is difficult to get a completely full second cd. I have been increasing the Q adjust upward to actually force a slight oversize on a two cd encode to cause a third partial cd to be created in the muxing step. That small third cue/bin is normally just a few minutes of titles which I can discard thus allowing the second cd to be full. As tylo suggested, perhaps increasing the sample rate to 3% or maybe even 4% might be a more proper way to go to accomplish this although it would take a few minutes in additional time to do the prediction.

Anyway, I have been using tylo's plugin and just doing 1pass vbr encodes with excellent results and I would highly recommend it, especially to those folks with slower machines, a lack of time, or the need to use heavy filtering for problem source. My old 1.4 TBird will do a completely finished burn ready encode on a 2 hour source in about 3 and a half hours total even with light filters. BTW, using undot() and Deen() causes the Q to be about 3 points lower than a unfiltered encode with only a small speed hit.

Although I have not tried it yet, tylo mention support for batch files (how?). Conceivably this would allow me to do three full encodes in a overnight night cycle assuming I got it started a few hours before midnight.

[tylo]

Thank you, DDogg, I'm glad it works so well for you.

I would rather use auto num CDs. It will always select the number you would have choosen manually (if you have set the right 'Prefered Q', that is). Remember, it uses enough CDs to guarantee a certain quality (Q factor), but no more!
I think it's time to go back one year, to the "golden era" : this post from Bach (http://forum.doom9.org/showthread.php?s=&postid=158971#post158971) and this from "the one" (http://forum.doom9.org/showthread.php?s=&postid=159384#post159384), 4 posts later. Read them, and then read them again.

The batch feature is a little limited at the moment. You may only use the CDs=auto, and all the other settings will be equal for all the convertions. I will add additional command line support to eliminate this problem. Anyway. In the .bat file created in the dvd2svcd home dir, add the line:
"c:\Program files\DVD2SVCD\Tylo\D2Sroba.exe" -batch
Before each batch run of DVD2SVCD. I will later add command line args like:
-CDs <n> -sample <%> , etc.

Cheers.

[Holomatrix]

I'm finding the plugin very usefull by doing a 2CD encode I use 3% sample @ 1.1 Q adjust and I get pretty close to 2 full CD's but using the same settings set for 1 CD I will get like 1 1/2 CD's for the same movie. Just wondering why the Q guesstimation be way off for 1 CD encodes?
Thanks

[tylo]

@Holomatrix: It's because you have reached the max Q(=63) that this algorithm computes. That is crappy quality already. To stay within only 1CD, you would need a much higher Q (80?), so there you are.

@DDogg: Sorry, just checked the batch feature. Hadn't tested it for this release, so I think it's broken. Will be there in the next release. I was planning to use that myself, you see. :)

[Holomatrix]

Allright then, that explains it :)

[DDogg]

You know the quality freak that I am. Do you think I would suggest anything not close to "excellent" Quite seriously that is the only reason I took the time to figure this plugin out. Now that tylo has added the new features I think it should appeal to everybody and it is really a no brainer to use. I would still advocate the use of the 2-3 setting for most instead of auto as the left brained, non control-oriented tylo suggested :D (tylo, you do know I am having a bit of fun with you I hope)

BTW, Bach, your mailbox is full, or was a day ago. Heh, perhaps you like it like that :p

[Kandissa]

Hi tylo!

I have downloaded version 2.1.5 of D2SROBA. I tried to use it with DVD2SVCD 1.1.3 b2. After filling in the suggested fields in the D2SROBA window and clicking go, the DVD2SVCD window comes up and nothing happens.

If I manually start Rip & convert in DVD2SVCD it runs till the video encoding step where the CCE window comes up & stays. Then nothing happens.

I am using P4 and Windows ME.

Thanks in advance for your help

[tylo]

Originally posted by DDogg
I would still advocate the use of the 2-3 setting for most instead of auto as the left brained, non control-oriented tylo suggested :D (tylo, you do know I am having a bit of fun with you I hope)Hehe, I guess there's some thuth in that. I have removed the possibility to set ranges in the new release, so the subject is now out of the way.

Just kidding. But, yes, I've got a new release for you. This time with full batch functionality, and with the easiest user interface imaginable for it. You're getting spoiled.:p

I think now the plugin is as perfect as I wanted to have it, so no further developement is planned from my part, in the nearest future at least. Small bug fixes will surely be done, if needed.:D

Download D2Sroba 2.2.0: http://home.no.net/tylo/


@Kandissa: (we posted at the same time)
Read the doc carefully. Make sure D2Sroba is running in the background (Status win + tray icon). CCE+DVD2SVCD should break when the CCE starts to encode. I have not tested with WinME.

[Holomatrix]

@tylo - thanks for all your work on your program but I was wonderng if D2Sroba still works the way it's supposed to if I start it up again but select to do a recover within DVD2SVCD to re-do the video encoding with some different options set? so I don't have to go through the audio extract/conversion process again. D2Sroba should still take over once the recover/continue @ video encoding starts again, right?

[tylo]

Yes, but if you change the video enc. mode, you'll have to do that in D2S after loading the project file when recovering (i think you can, at least).

Recover works because D2Sroba does not do/set anything before it sees the CCE encoding window, and that happens just after the D2S 'Video encoding' recover stage, so D2Sroba will jump in at the right place.

(edit) Ok, I misread that. It was just a quiestion whether Recover worked or not. So, yes it does - I've done it quite a few times myself.

[DDogg]

Yeah, I am definitely getting spoiled now! This is just getting better and better.

It was just a question whether Recover worked or not. So, yes it does - I've done it quite a few times myself. Tylo, I have noticed if you have previously completed the prediction phase and then for some reason such as testing, you redo the encode, the plugin will pick up its last successfully completed prediction data and start the long full encode without redoing the prediction. This is great if you have had a crash or such, but requires a manual deletion of files if you wish to force a re-prediction. Perhaps a requester asking to continue previous session or new might be something to consider?

I noticed a small cosmetic anomaly in that after a 2-3 prediction run I chose 2 but the bitrate shown in the status window was from the 3 choice. Trivial, but I thought you would want to know.

Add: CCE run by plugin should have same priority as set in Dvd2svcd? Or a checkbox for low? Just wondering.

[Kandissa]

@tylo

I followed the new instructions that come with version 2.2.0. I have switched over to WinXP. I am using CCE 2.66 & eclCCE.

The CCE SP loads AviSynth_Script_file.avs, encodes & stops. DVD2SVCD also stops but both windows remain on screen. Then nothing happens. I have waited for around 25 mins.

What can I do to make D2SROBBA work ? Or is it just stupid me?

dvd2svcd_log.txt
================

WNASPI32.DLL 4.6.0.1021
--------------------------------------------------------
- 14/06/2003 07:16:37
- DVD to SVCD Conversion
- DVD2SVCD ver. 1.1.3 build 2
--------------------------------------------------------
Initializing
Initializing finished.

--------------------------------------------------------
- 14/06/2003 07:16:40
- Free on drive F: 16393.48 mb
- Internal rip
--------------------------------------------------------
Preparing vobfiles to be ripped:
- G:\VIDEO_TS\VTS_01_1.vob
- G:\VIDEO_TS\VTS_01_2.vob
- G:\VIDEO_TS\VTS_01_3.vob
- G:\VIDEO_TS\VTS_01_4.vob

Ripping: G:\VIDEO_TS\VTS_01_2.vob
Ripping finished. Speed: 7537

--------------------------------------------------------
- 14/06/2003 07:16:50
- DVD2AVI
--------------------------------------------------------
Creating DVD2AVI INI file:
- E:\Program Files\DVD2SVCD\DVD2AVI\DVD2AVI.INI

Variable settings:
iDCT_Algorithm: 32-bit SSE MMX

Executing DVD2AVI.
Executing DVD2AVI. Commandline:
"E:\Program Files\DVD2SVCD\DVD2AVI\DVD2AVI.exe" -CS=2 -YR=1 -EXIT -OF=[F:\DVD2SVCD\DVD2AVI_Project_file] -IF=[F:\DVD2SVCD\VTS_01_1.vob]
Analyzing DVD2AVI Project file
Force Film NOT activated!
Framerate: 29970
DVD2AVI processing done.

--------------------------------------------------------
- 14/06/2003 07:17:00
- Free on drive F: 16662.84 mb
- AUDIO Extraction
--------------------------------------------------------
Found MPA stream id: 0xC0
Filename: F:\DVD2SVCD\AUDIO\Extracted_audio_1.mpa
Audio1 delay: -161 ms
Audio extraction finished.

--------------------------------------------------------
- 14/06/2003 07:17:04
- Free on drive F: 16660.20 mb
- AUDIO conversion
--------------------------------------------------------

Encoding Audio. Filename: F:\DVD2SVCD\AUDIO\Extracted_audio_1.mpa
Executing BeSweet. Commandline:
"E:\Program Files\DVD2SVCD\BeSweet\BeSweet.exe" -core( -input "F:\DVD2SVCD\AUDIO\Extracted_audio_1.mpa" -output "F:\DVD2SVCD\AUDIO\Encoded_audio_1.mp2" -logfile "F:\DVD2SVCD\AUDIO\Encoded_audio_1.log" ) -ota( -g max -d -161 ) -2lame( -e -b 224 -m s )
Audio conversion of F:\DVD2SVCD\AUDIO\Extracted_audio_1.mpa finished.

Audio conversion finished.

--------------------------------------------------------
- 14/06/2003 07:17:25
- Free on drive F: 16657.56 mb
- Video Encoding using Cinema Craft
--------------------------------------------------------

Executing Cinema Craft Encoder.
StreamSectors: 771186864
AudioSectors: 427616
VideoPAPO: 11939472
ScanOffsetBytes: 288
SeqAligningBytes: 18304
DVDBytes: 0
VideoEndHeader: 4
SubtitleSectors: 0
EmptySectors: 238.00
PictureSectors: 1.00
PureMPEGStream: 758801180.00
Seconds: 15.00
CalcMPEGStream: 758801180.00
Frames: 240
CDSize: 740.00
Number of CDs: 1
Cut point 735.00
Variable Settings:
Frames: 240
Anti Noise Filter: Off
Passes: 4
Image Quality: 17
VAF file creation: On
Video Encoding Mode: One Pass VBR
Q. Factor: 5
Min. bitrate: 300
Max. bitrate: 2230
Aspect Ratio: 4:3 (No borders, encoded as 4:3)

---AVS Begin---
Import("F:\DVD2SVCD\VIDEO\RESAMP~1.AVS")
LoadPlugin("E:\PROGRA~1\DVD2SVCD\AVISYN~1.5PL\Mpeg2dec\MPEG2D~1.DLL")
mpeg2source("F:\DVD2SVCD\DVD2AV~1.D2V")
ConvertToYUY2()
SmartDeinterlace(2,10,True,False,True)
BicubicResize(480,480,0.0,0.6)
ResampleAudio(44100)
----AVS End----

[tylo]

Perhaps a requester asking to continue previous session or new might be something to consider?
That would require a user intervention at a late stage. I could make an opition to clean beforehand, but I just think you'll have to delete the files manually. I do. After all, re-prediction is not normal way you would run the plugin.
Add: CCE run by plugin should have same priority as set in Dvd2svcd? Or a checkbox for low?
I don't know how to do that right now, but there is a way if you use CCE 2.66+ (I thought you did). RB has added permanent priority settings in the EclCCE tab :)
I noticed a small cosmetic anomaly in that after a 2-3 prediction run I chose 2 but the bitrate shown in the status window was from the 3 choice.
Thanks for notifying me. I have fixed that. Actually, I uploaded a new release with much improved (and enjoyable) output in the status window. However, I have also changed the range m-n option. You may want to :devil: me for that, but check it out, and tell me what you think.

@Kandissa: I still don't have a clue what the problem is. You say the CCE window breaks and closes, but the DVD2SVCD window will not close? It should! Clean the desktop, and dont touch the PC (before the main encode runs, at least).

[Holomatrix]

Plus tylo can you tell me how to add more adjust Q values. I've done three movies so far with this new program and even at 3% 1.6 the second cd is to small.
Thanks

[tylo]

Just type in the value you want. It will even be sticky (default next time you run). I've also experienced that most movies overestimate (real size are smaller). I think it's because the samples created have very short scenes. It kind of gets more action than the real movie. However, on heavy action movies, I have noticed that the differences are smaller, i.e the estimates are more accurate (which makes sense).

It would be interresting to hear if other have seen the same thing.

[DDogg]

I could make an option to clean beforehand Probably not a big deal as you suggest, but it will cause confusion for people trying to learn this plugin as those people will be the ones running it multiple times. Actually an option to delete the triggering file AFTER a successful encode might be better if you decide to mess with it or a never re-use checkmark?permanent priority settings in the EclCCE tab I.m a dummy, sorry.:o You may want to (flame) me for that Never! :) However, having an exclude option for 1cd and 4 cd prediction would not upset me at all. It just seems such a waste of time to see it do three or four runs on 1 cd prediction using a silly Q of 48 or something that nobody with a brain would ever select. Frustrating really, I don't really understand why you changed it, but I assume you had your reasons. <cough, sputter> I did not realize it was broke and needed fixing as it seemed PERFECT <more sputter> :)

Edit, yeah, I have now been here 30 minutes and it is still on the 1cd prediction and up to Q68.....Where is my shotgun?

Edit 2: I lied, where is the big assed nuclear flamethrower?
Calculations:
++++++++++++++++++++++++++++++++++++++++++++++++++++++++
- Number of CDs : 1
- Target mpv size : 671246420. (783 kbps)
--------------------------------------------------------
- Binary search for Q:
- Estimated mpv size : 1440810044. (Q=32, 1682 kbps)
- Estimated mpv size : 1146671824. (Q=48, 1338 kbps)
- Estimated mpv size : 1045837197. (Q=56, 1220 kbps)
- Estimated mpv size : 1002937049. (Q=60, 1170 kbps)
- Estimated mpv size : 983078214. (Q=62, 1147 kbps)
- Estimated mpv size : 973550092. (Q=63, 1136 kbps)
--------------------------------------------------------
- Determined Q. : 63

Well, this means I never will never use the ? and so the change has cost me the ability to make that fine decision between 2 and 3 cds which is what we started our original comversation about. It was just great before this. Iz you punishing an old dogg for not reading the funky readme? (sniff!) :D

[DDogg]

I fear it is broken anyway as it only gives a choice between one or two cds, both using a projected Q above the minimum specified. Yeah, I think you said you don't use the minimum for the range work, but hey, do you know the frame count? Assuming you do, or know the length, why bother with a full 147 step (seems like):) 1 cd prediction on a 2 hour movie if the Q is set at 20? Howz about getting a little AI going on?

I am kidding and picking only a little when I ask, "and how were those mushrooms" (which you must have been eating when you decided to do these changes). lol

Oh, got a link for the old version? This one is unusable for me.

[Holomatrix]

Originally posted by tylo
Just type in the value you want. It will even be sticky (default next time you run). I've also experienced that most movies overestimate (real size are smaller). I think it's because the samples created have very short scenes. It kind of gets more action than the real movie. However, on heavy action movies, I have noticed that the differences are smaller, i.e the estimates are more accurate (which makes sense).

It would be interresting to hear if other have seen the same thing.

I didn't even think about typing it in manually, duh :) the movie's I did were Tears of the Sun and Die another day, seem to have a good amount of action and gave me small second CD size for both. Oh, well.
When I say small second CD size I'm only talking about 40 meg shy of 800 meg so not really that big of deal but now that I can type in a bigger value, I'm happy :)

[tylo]

Hmm. Tough one. First, let me explain why I did it. I though it was frustrating to choose from a range of num CDs, which one of them easily could fall out of the range for usable bitrates to fill up the CDs. You can always do a guess of 2-3, which in 90% will have at least one usable number - which is what you did with the 2-3 option.

The heart of the matter is: you just never want to have avg. bitrates below 1100, and above 2400 or so (SVCD spec. limited). If you set the Min and Max average bitrate in DVD2SVCD (check "Use" Min avg.), The ? option calculates Q values for those num CDs that fall between these bitrates, and thus are usable for you.

So, when you did your so-called "fine decisions" between 2 and 3 CDs, I guess that most of the time, one of them where out of the question. It would either have too high or low bitrate, and/or too high or low Q value.

With v2.2.4, when there actually is a "fine decision" to be made (i.e. two usable num CDs - there will never be more), you can be sure that the plugin will present them as a choices for you - and it will only show and calculate on those choices that indeed are usable. That could well be 1-2 or 3-4.

As an ad nota:
If you also wanted to select on num CDs that gives Avg bitrates above 2500, (outside SVCD spec). you could do that by increasing the Max avg. value, but it would be unwise: It would lead to OPV encoding with Q=1, and pressed to limit bitrates to 2500 - in practice a CBR 2500 encode. I believe that then it would be much better to do a true CBR 2500 encode (experts, please comment).
1 cd prediction on a 2 hour movie if the Q is set at 20? Howz about getting a little AI going on?
The whole discussion above is about putting AI into the app. For example to rule out completly useless choices of CDs to begin with.

This is what is going on with the ? option for Num CDs:
for nCDs=1 to 4 {
-- Compute bitrate for nCDs:
-- if bitrate lower than dvd2svcd Min avg. -> forget it
-- else if bitrate higher than dvd2svcd Max avg. -> forget it
-- else estimate Q. that produce this bitrate, and add as choice for user.
}
If more than one choice: RadioBox(choices)
else let the plugin use the one usable choice.

(the auto option filters on Q values instead of bitrates, securing that you don't get those rediculous high Q values in your estimations)

Here is what is going on in the old 2-3 option:
for nCDs=2 to 3 {
-- Compute bitrate for nCDs:
-- No matter what stupid bitrate:
-- estimate Q that produce this bitrate and add as choice for user.
}
RadioBox(choices) - one often useless, so no real choice

It will cost far less to put the old code back in, than it did to write this post, so don't worry, :D it'll be there for people like you... :rolleyes:

(edit) The pitfall with this version was that I didn't warn you about having to set the Min.avg bitrate in DVD2SVCD. It took it as 0, so you got the 1CD estimation. It ought not be necesarry to do any special settings in DVD2SVCD for D2Sroba to work. Sorry. :) Had I just hardcoded 1000 - 2500 things would have worked flawlessly.

@Holomatrix: Great to hear that. 40 megs is as close you can expect to get. If you adjust more, you can easily fall over the edge (normally just loosing some credits text).

[DDogg]

The pitfall with this version was that I didn't warn you about having to set the Min.avg bitrate in DVD2SVCD. Well that would certainly have been very handy to know as it is the single deciding factor in whether it works or not. :rolleyes: (one roll eyes back at ya :) ) If the min average has to be checked that will cause a lot of problems for casual users, IMO. The hard coding you mentioned might be better.This is what is going on with the ? option for Num CDs:
for nCDs=1 to 4 {
-- Compute bitrate for nCDs:
-- if bitrate lower than dvd2svcd Min avg. -> forget it
-- else if bitrate higher than dvd2svcd Max avg. -> forget it
-- else estimate Q. that produce this bitrate, and add as choice for user.
This sounds great now that we know the secret. Assuming you can use a hard code NOT dependant on a min average entry then it sure sounds like the smart way to go. I do understand what you are saying and yes, some users might allow a bitrate higher than svcd max under the old 2-3 system [add] although I wonder if a simple "out of standard range" warning might not have accomplished the same protection from user stupidity.

[Holomatrix]

I think you guys just lost me :( when would I need to set/check Min avg. 1100? If I always set the amount of CD's I want, can I ignore this step?