AviSynth rendered subtitles to BluRay SUP/PGS and BDN XML (v2.08)

[Hyper Shinchan]

@Hyper Shinchan: Those programs don't support split lines, like those produced by -b1. It only works with Scenarist, I think.
Actually other splitted subtitles didn't seem to be a problem for those programs, I can post more samples if it's needed. Anyway if it's a problem with those two programs (that surely doesn't support all the functions of Scenarist) it could be worth mentioning it either in the printed help or in your home page.

[ps auxw]

Actually other splitted subtitles didn't seem to be a problem for those programs, I can post more samples if it's needed. Anyway if it's a problem with those two programs (that surely doesn't support all the functions of Scenarist) it could be worth mentioning it either in the printed help or in your home page.
Peculiar. As is, I can't even look at the sample you posted above, since it isn't approved yet. Also, good point about mentioning -b compatibility issues. I put a note in the first post.

[hamletiii]

I dunno if anyone else had similar problems, but apparently both BDSup2Sub and BDSupEdit can't deal correctly with some subtitles splitted by the -b1 switch, the second line isn't displayed; is it a problem of those programs?
I've included the single line of the ass script that gave me problems (the avs script was created as: Masksub("filename",1920,1080,23.976,frame number)) if you want to test it by yourself.

Not sure about BDSupEdit, but BDSup2Sub does not support -b1. This has been talked about several times in this thread and the BDS2S thread. The author 0xdeadbeef also confirms (http://forum.doom9.org/showpost.php?p=1389088&postcount=129) that -b1 is not supported and he has no intention to add such support. This means subtitles created by avs2bdnxml and many Japanese BDs with similar features are not supported.

-b1 (need avs2bdnmxl ver 1.13, previous versions have bugs) works extremely well with Scenarist BD. Scenarist BD reads the region size from the tags, but the window size is based on the overall size of these pictures after Scenarist BD did the Epoch combining step.

[hamletiii]

Actually other splitted subtitles didn't seem to be a problem for those programs, I can post more samples if it's needed. Anyway if it's a problem with those two programs (that surely doesn't support all the functions of Scenarist) it could be worth mentioning it either in the printed help or in your home page.

Are you sure those are dual window splitted like what -b1 does? I would be curious to see it. Can you upload the sample on a file sharing site such as megaupload or mediafire?

[Hyper Shinchan]

Are you sure those are dual window splitted like what -b1 does? I would be curious to see it. Can you upload the sample on a file sharing site such as megaupload or mediafire?

I'm sorry, I've jumped to conclusion without checking the properly, I simply didn't notice that they were missing while watching the movie on the PS3 (and after searching for splitted subs I noticed that my scripts had other than the one I've posted that I've recognized as missing while watching the movie), now I've checked them and they're not processed correctly by BDSup2Sub, I'm sorry.

The author 0xdeadbeef also confirms that -b1 is not supported and he has no intention to add such support. This means subtitles created by avs2bdnxml and many Japanese BDs with similar features are not supported.
You say that many Japanese BDs use them, have you mentioned that to him?

Also, good point about mentioning -b compatibility issues. I put a note in the first post.
Thank you very much, some people (like me) could simply start using your program without checking all the discussions in doom9 and they could encounter similar issues, that note will save them some problem.

[hamletiii]

You say that many Japanese BDs use them, have you mentioned that to him?

Thank you very much, some people (like me) could simply start using your program without checking all the discussions in doom9 and they could encounter similar issues, that note will save them some problem.

Please read pages 5-7 in this thread, all your questions should be answered already. I believe reading the threads is the best way to learn things if you encounter problems.

So basically, BDS2S is a subtitle CONVERSION tool instead of a rendering tool. The author is not bothered to extend his tool to cover all BD Specs. Your best bet is to wait for ps auxw's direct sup support or someone else to come up a fully spec compliant xml+png->sup program.

For now if you still want to use BDS2S, my suggestion is to keep all your subtitle within half of your vertical resolution (540 pixels for 1080) given any time. This way you can get the correct results without using -b1. I know it sucks that you can't put subtitles anytime anywhere you want, but this is just the way it is for now, unless you are using Scenarist BD...

[ps auxw]

Once again I have to apologize for a long-ish period without updates. Still, I bring some nice news. I just uploaded avs2bdnxml 2.01 (http://ps-auxw.de/avs2bdnxml/avs2bdnxml-2.01.tar.bz2) with experimental SUP output support. Please note, that I do not currently have access to BluRay playback hardware or software, so I could not do any real world tests, but the output files look reasonable, compared to reference files. If you experience any problems, please report them. If it works, I'd be happy to hear about that too. ;)

I have also added the options "--count" and "--seek" to support partial processing of input files.

[hamletiii]

Thanks for your hard work. This software just gets better and better.
I did a brief test with a pretty complex *.ass (dual window consecutive multiple composition objects in one epoch), but there's some problem with the "compositionposition structure" as reported by Scenarist BD.

The generated sup does not work in TMT3, only the first subtitle shows up, then never goes away. If you drag the timeline around, there's no subtitle showing up.
So then I demuxed the project with BD Reauthor Pro, got the *.pes file, then drag the *.pes file into Scenarist BD, got the following error:


Error : Out of range - there are out of range value(s), in the CompositionPosition structure.
Error : PG Decode error. K:\ALL MUXING\PES testing2\00000_1_01.pes offset=0x70120 (first byte of invalid segment).
Error : FilePath : K:\ALL MUXING\PES testing2\00000_1_01.pes,
Decode failed.



HERE (http://www.mediafire.com/?61wcwf1zfha53xx) is the sample, the invalid one is using avs2bdnxml 2.01, the valid one is xml+png from avs2bdnxml 1.13 muxed with Scenarist BD then demuxed with eac3to.

Also can you make avs2bdnxml to generate *.sup AND xml+png at the same time?

The pgsparse.exe in the debug folder looks interesting, but I don't see it automatically generates reports?

[ps auxw]

Thank you for the bug report. I'll look into it today.

Also can you make avs2bdnxml to generate *.sup AND xml+png at the same time?
I'll see what I can do. Edit: Implemented. In the next version, you can use "-o" twice, with different output formats.

The pgsparse.exe in the debug folder looks interesting, but I don't see it automatically generates reports?
It writes the information to standard output. You can generate a report like this:
pgsparse.exe input.sup > report.txt

Edit: Were identical settings used for avs2bdnxml in both cases? If so, which? The content of those SUP files seems to be a bit different. Actually, the valid file looks a bit munged (e.g. the dts field is always set to 0). Could you possibly send me the .ass and .avs files too?

Edit 2: I just released avs2bdnxml 2.02 (http://ps-auxw.de/avs2bdnxml/avs2bdnxml-2.02.tar.bz2) with the aforementioned double-output and some fixes to SUP output. I am not sure, whether any of them could possibly help with the CompositionPosition structure problem. If not, having the .pes file might be useful so I can look at the data near 0x70120 for clues.

[hamletiii]

The double output works, but sup output is still the same result. In addition, I tested in Scenarist QC this time, the subtitle formatting is messed up, only composition object 2 is showing up in window 1, composition object 1 is supposed to be showing up in window 2 but is showing a wrong positioned composition object 2 instead, and the first subtitle shows up and goes away repeatedly...

The command I used is:


avs2bdnxml -t testing -l zho -v 1080p -f 23.976 -a1 -p1 -b1 -m3 -u0 -o test.sup -o testing.xml xx.avs



I've included the *.pes file generated by Scenarist BD, *.ass, *.avs this time HERE (http://www.mediafire.com/?vup6na3d48nv717), let me know if you need anything else.

[ps auxw]

Thank you very much for the thorough information.

The double output works, but sup output is still the same result. In addition, I tested in Scenarist QC this time, the subtitle formatting is messed up, only composition object 2 is showing up in window 1, composition object 1 is supposed to be showing up in window 2 but is showing a wrong positioned composition object 2 instead, and the first subtitle shows up and goes away repeatedly...
That sounds very peculiar. It will be interesting to see what the problem is. Although, I have some pretty good ideas about the cause. In any case, the next update might take a few days.

I've included the *.pes file generated by Scenarist BD, *.ass, *.avs this time HERE (http://www.mediafire.com/?vup6na3d48nv717), let me know if you need anything else.
Unless I am missing something, the .avs and .ass files are not included in the file.

Edit: testing_invalid_tsmuxer.sup and testing_valid_tsmuxer.sup are identical files. Is that intended?

[hamletiii]

Lol, I made so many mistakes this morning... THIS (http://www.mediafire.com/?mm0mdy7578p4x1v) is an updated package.

Testing_invalid_tsmuxer.sup and testing_valid_tsmuxer.sup are supposed to be different, sorry for the confusion...
The valid one is xml+png generated by avs2bdnxml 2.02, import into Scenarist BD, which produces a *.pes, then muxing the project, and lastly the *.m2ts is demuxed by tsmuxer.
The invalid one is *.sup produced by avs2bdnxml 2.02, then using tsmuxer to output a Blu-ray Disc structure with audio/video, and the *.m2ts is demuxed by tsmuxer.

Testing_invalid_tsmuxer.sup should be identical to test_org.sup which is the original sup generated by avs2bdnxml 2.02, it suggests that tsmuxer doesn't alter the sup in any way while the sup extracted by eac3to resets all DTS timestamp to 0.

Upon examining the reports from pgsparse you provided, I can see that the valid sup is likely to be what I see in Scenarist BD's PES Editor, while the invalid sup is quite different from what the correct implementation should be.

[ps auxw]

Thanks for the updated package. It will make testing much easier. And yes, there still are some pretty big differences. I'll try to get rid of them. ;)

[fransky]

excellent work here. Are there any more inline effects supported other than fadein/fadeout in the xml, which Scenarist BD also supports?

[ps auxw]

So, slightly longer delay again, sorry. I hope I'll get to work more on this in the next days.

excellent work here. Are there any more inline effects supported other than fadein/fadeout in the xml, which Scenarist BD also supports?
Take a look at this thread (http://www.urusoft.net/forum/viewtopic.php?id=5446) with the XML Schema. There seem to be at least Crop, Position and Color effects too. No idea how they work or if Scenarist supports them, though.

[hamletiii]

The inline fading effect was generated by Lemony Pro, which works on Scenarist BD.
The fading also can be written as a global function, and called in subsequent block instead of individual function in the example I gave earlier:



<?xml version="1.0" encoding="UTF-8"?>
<BDN Version="1.00" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="BD-03-006-0092b BDN File Format.xsd">
<Description>
<Name Title="Title" />
<Language Code="eng" />
<Format VideoFormat="1080p" FrameRate="23.976" DropFrame="false" />
<Events NumberofEvents="2" Type="Graphic" />
</Description>
<Effects>
<Effect EffectID="FadeIn1" Type="Fade" Duration="00:00:00:12" Anchor="Start">
<Fade FadeType="FadeIn" />
</Effect>
<Effect EffectID="FadeOut1" Type="Fade" Duration="00:00:00:12" Anchor="End">
<Fade FadeType="FadeOut" />
</Effect>
</Effects>
<Events>
<Event InTC="00:00:05:14" OutTC="00:00:08:14">
<Graphic>ue_Track_0001.png</Graphic>
<Effect EffectID="FadeIn1" />
<Effect EffectID="FadeOut1" />
</Event>
<Event InTC="00:00:11:23" OutTC="00:00:14:23">
<Graphic>ue_Track_0002.png</Graphic>
<Effect EffectID="FadeIn1" />
<Effect EffectID="FadeOut1" />
</Event>
</Events>
</BDN>



There's some difficulties in applying fading from the graphics generated in avs2bdnxml though, because it tends to use too much color to index each picture. Pure thin black boarder with white text should be fine, but other fancier colors will almost always having problem, this is especially true when you need cross fading.

I'm not sure if crop, position, color effects are supported in subtitles, I can't seem to make those happen in Lemony. But they are definitely supported in Interactive Graphics - IGEffect Editor, where you can archieve some slide in/out effects using a combination of cropping and moving. You could also specify an initial color palette and an ending palette, Scenarist BD will fill in the missing ones. The cropping and moving can also be applied upto two individual windows, but if the windows are too far apart, you'll run into timing issue which the animation won't be smooth... Anyway, it's pretty neat stuff without using animated button.

[fransky]

thanks for explaining, ps auxw and hamletiii.
fading shouldn't be any problem for me now. I'm now more concerned about moving subtitles. Let's say a line moving from left to right of the screen. The way avs2bdnxml does is producing an png image for each frame. It often causes buffer overflow, especially if another line of subtitle appears during the motion. It would be very nice if avs2bdnxml could output xml+png in a way that let Scenarist BD or other tools understand it's a moving line. So it won't need multiple images for one moving line. Maybe I'm asking too much. But some idea about possibility is appreciated.

[Daiz]

I think the easiest way to do that would be to read the .ass file along with the .avs file and look for lines with the \move command, then just render the first frame and use the \move-command parameters as help to produce the BDN-XML movement... but I don't know how easy that would be to pull off. I guess it should be easier than trying to implement some sort of motion tracker or whatever else would be required to do it otherwise.

[hamletiii]

I think the easiest way to do that would be to read the .ass file along with the .avs file and look for lines with the \move command, then just render the first frame and use the \move-command parameters as help to produce the BDN-XML movement... but I don't know how easy that would be to pull off. I guess it should be easier than trying to implement some sort of motion tracker or whatever else would be required to do it otherwise.

I believe we are already doing that for fading. But we don't know the exact move command (if there is) like the fade for Scenarist BD to implement.

[hamletiii]

DVDLogic has a new tool called EasyBD (http://www.dvd-logic.com/easybd/features.html) which can take xml+png and create pes+mui. I assume it works the same way as Scenarist BD to allow dual window multiple composition objects configuration.

[ps auxw]

Interesting. If it works well, I will put a link to it in the first post. Could anybody test?

(Yes, I'm still alive, just too busy. :()

[Hyper Shinchan]

Interesting. If it works well, I will put a link to it in the first post. Could anybody test?
I've tested it and it seems that it works fine; unlike last time with BDSup2Sub I'm pretty sure that it works, I've checked that both objects appear correctly.
Anyway I only made a fast test with a pair of subtitles, more testing is needed; also Scenarist gives you a warning if there's something wrong with your XML, I wonder if EasyBD gives similar warnings.
BTW I've used it only to mux a dummy AVC stream and the subtitles, I've used tsmuxer to remux the subs in AVCHD with the video and audio (and I've tested it with the PS3).
Did you test it by yourself hamletiii?

[hamletiii]

I've tested it and it seems that it works fine; unlike last time with BDSup2Sub I'm pretty sure that it works, I've checked that both objects appear correctly.
Anyway I only made a fast test with a pair of subtitles, more testing is needed; also Scenarist gives you a warning if there's something wrong with your XML, I wonder if EasyBD gives similar warnings.
BTW I've used it only to mux a dummy AVC stream and the subtitles, I've used tsmuxer to remux the subs in AVCHD with the video and audio (and I've tested it with the PS3).
Did you test it by yourself hamletiii?

No, I haven't tested it. But DVD-Logic has a good record at making software that's compliant to specification. And they are working with Sonic closely with the EDGe program. So I think it's probably going to work. I'll test it and compare the result with Scenarist BD when I got time.

Muxing is not the main problem with subtitle get screwed up here, as long as you are using the same frame rate and resolution. But before muxing, the subtitles have to be in sup or pes format so that these muxers can mux. This is where things get wrong. At this time, it seems nobody except Scenarist BD can fully support the specification of BD subtitles. It seems most programmers understand only part of it, but never fully understand what's going on. Wirepole has wrote some good discovery here (http://wirepole.spaces.live.com/default.aspx), but it's in Chinese. I think it could be useful to ps auxw.

What gets wrong with BDSup2Sub is that it deliberately puts gaps between subtitles with timings that are very close so that it can write just one composition object per epoch. This is NOT the right way to handle the situation. Plus it cannot handle dual window configuration which is an even bigger problem.

[ps auxw]

Ok, I have now added a note about EasyBD to the first post.

At this time, it seems nobody except Scenarist BD can fully support the specification of BD subtitles. It seems most programmers understand only part of it, but never fully understand what's going on. Wirepole has wrote some good discovery here (http://wirepole.spaces.live.com/default.aspx), but it's in Chinese. I think it could be useful to ps auxw.
Thanks. While I do not know Chinese, maybe I'll get some useful information from there. Now, if I just got around to actually working on this...

[hamletiii]

I've tested it and it seems that it works fine; unlike last time with BDSup2Sub I'm pretty sure that it works, I've checked that both objects appear correctly.
Anyway I only made a fast test with a pair of subtitles, more testing is needed; also Scenarist gives you a warning if there's something wrong with your XML, I wonder if EasyBD gives similar warnings.
BTW I've used it only to mux a dummy AVC stream and the subtitles, I've used tsmuxer to remux the subs in AVCHD with the video and audio (and I've tested it with the PS3).
Did you test it by yourself hamletiii?

OK, I did some small tests. I'm rather disappointed with the result. I used the same files in my post#162 (http://forum.doom9.org/showpost.php?p=1454471&postcount=162) as a stress test, but unfortunately the program crashes. Then I tested some simple scripts with no image splitting (dual window), and it compiles fine. So I can pretty much say that this program doesn't support dual window configuration. I didn't test fading tag support, but it'll probably NOT work. I'm not impressed with the GUI as well...

[hamletiii]

Hi, ps auxw, just in case if you could start working on this program again, could you implement this:
In the event a really big picture is generated, it would be best to split them no matter how ugly they are. I think if the vertical pixels go over 540 (half of 1080), you need split.

Here is an example why this is needed:

http://i44.photobucket.com/albums/f18/DVDmaster/5c13ade7.png

This subtitle needs some busy on screen translation. So the timings are really close. Therefore Scenarist BD tries to group these seven lines together by putting them into one epoch. The problem came after this. Now since one of the picture is so large (avs2bdnxml didn't split them probably the lines are not too far apart?), Scenarist BD is unable to put a dual window structure for the entire epoch! This increases buffer usage significantly. So there are two solutions after this:
1. Reduce the text on each picture, which means you can't put as many text as you want. Reducing pixel counts help reducing buffer usage.
2. Splitting the pictures so that Scenarist BD can put them into two windows.

And here's what it looks like after I manually split the picture:

http://i44.photobucket.com/albums/f18/DVDmaster/9019757b.png

Scenarist BD picks up the dual window here. And it easily saves 8% of buffer!

So I think if a picture has more than 540 pixels, it needs split. I think it is somewhat similar to what you proposed here (http://forum.doom9.org/showpost.php?p=1416039&postcount=144). But in this case, I just want it split no matter how ugly the picture is, as long as the final result looks seamless.

[ps auxw]

I will add something like that as soon as I get around to working on this, hopefully in the not too far future. Also, at least in that case a non-ugly split should've been found. If that is not the case, well. I have an idea for a better splitting algorithm that might help. :)

[hamletiii]

I will add something like that as soon as I get around to working on this, hopefully in the not too far future. Also, at least in that case a non-ugly split should've been found. If that is not the case, well. I have an idea for a better splitting algorithm that might help. :)

In my example, it shouldn't be too hard to find the right splitting coordinates, but avs2bdnxml didn't pick up. I was actually doing something more complex at first like this:

http://i44.photobucket.com/albums/f18/DVDmaster/02f75cb3.png

I think it would be much harder to distinguish the pixels in this situation?

Well, it's good to hear you have a better idea.

[Oleg Rode]

I will add something like that as soon as I get around to working on this, hopefully in the not too far future. Also, at least in that case a non-ugly split should've been found. If that is not the case, well. I have an idea for a better splitting algorithm that might help. :)
Hi again and thanks for your work. The .sup output option works very well. Your subs splitting engine produces very good results. The problem exists only with very complex subs, but these subs are not used very often.
Best regards and, again, sorry for my English.

[hamletiii]

Well, I wouldn't call it work "very well". The sup output definitely needs some more work, dual window splitting is quite off from the correct implementation. But still, it's a big step forward towards the right direction since so far none of the other free and some commercial subtitle software I tried can do this, I bet some author didn't even know you can have multiple composition objects under one epoch, others aren't willing to add such support.

[hamletiii]

Hi, ps auxw. Since the wirepole blog I posted earlier has been down, here is the article (http://www.mediafire.com/?qd44uin3p729p1w) that's related to the ies, pes, sup format I had saved. I took some time to translate it as well, hopefully it could be useful for you to figure out how to output sup correctly.

[ps auxw]

Hi, ps auxw. Since the wirepole blog I posted earlier has been down, here is the article (http://www.mediafire.com/?qd44uin3p729p1w) that's related to the ies, pes, sup format I had saved. I took some time to translate it as well, hopefully it could be useful for you to figure out how to output sup correctly.

Thank you very much for the translation, this should be quite helpful.

Edit: Just adding a note to say it might take another month, possibly a bit longer, before I can get back to work on this.

[hamletiii]

With the help from someone on another forum, we were able to finally figure out ALL four inline effect tags: fade, crop, position, and color.
Again, it would be cool if avs2bdnxml could support converting these tags from *.ass tags without having to manually adding these afterwards.
The difficulty in doing this is when -b1 splitting is specified, it would be very hard to apply the tags.

I've collected all tag usages in one article which is a bit long, it's available HERE (http://www.mediafire.com/?ge19y4pwswiiwdg) for anyone who's interested.
Again, these tags only works on Scenarist BD, different version may behave somewhat differently.

[twazerty]

When I run it via Wine under linux (ubuntu) I get some errors. I noticed a makefile for linux. Is it possible to run this application native on Ubuntu?

[ps auxw]

General update: I'm finally getting around to working on this again. The SUP writer is much improved, but some new bugs were introduced, which I have to fix before I can release it. Well, at least there's progress after long last.

When I run it via Wine under linux (ubuntu) I get some errors. I noticed a makefile for linux. Is it possible to run this application native on Ubuntu?
Somewhat. You have to modify avs2bdnxml.c, about line 227:
p_param->i_width = 1920;
p_param->i_height = 1080;
p_param->i_fps_den = 30000;
p_param->i_fps_num = 1001;

Adjust the video parameters hardcoded there. The input file has to be raw RGBA data without any headers. Using a FIFO should work. (NB: The Linux makefile isn't really intended for general use, just for hunting memory bugs with valgrind.)

[twazerty]

If it runs via Wine it is okay too. I see I didn't use the latest version. Now it almost runs okay. The output only contains the first word of each line (srt file).

Installed Avisynth under wine. Loaded the VSFilter in the avs. Any ideas?

[hamletiii]

General update: I'm finally getting around to working on this again. The SUP writer is much improved, but some new bugs were introduced, which I have to fix before I can release it. Well, at least there's progress after long last.


That's great to hear you finally start working on it. I'll be glad to test it once you are done.

[ps auxw]

If it runs via Wine it is okay too. I see I didn't use the latest version. Now it almost runs okay. The output only contains the first word of each line (srt file).

Installed Avisynth under wine. Loaded the VSFilter in the avs. Any ideas?
I'm afraid not. I could never get VSFilter to produce usable results under wine.

That's great to hear you finally start working on it. I'll be glad to test it once you are done.
I have just released avs2bdnxml v2.04 (http://ps-auxw.de/avs2bdnxml/avs2bdnxml-2.04.tar.bz2), which hopefully fixes some/most problems with SUP output. My apologies for the repeated delays.

[twazerty]

I'm afraid not. I could never get VSFilter to produce usable results under wine.

Indeed I see VSFilter problems with Wine on the internet.

It would be very very nice if this application works on linux. Just some ideas:
Use ffmpeg input instead of avs. Maybe you can play a purple screen in ffmpeg with subtitles on top of it and then filter this purple color out of the image leaving pure subtitle images.

The reason I am looking for this is because I am updating AVCHDCoder so it works on linux. And your application is one of the features I haven't found an alternative for. Since VSFilter doesn't work correctly under wine I was hoping for another solution.

[sneaker_ger]

Try AssRender (http://forum.doom9.org/showthread.php?t=148926) instead of vsfilter.

[twazerty]

Try AssRender (http://forum.doom9.org/showthread.php?t=148926) instead of vsfilter.

I get an error that it isn't an Avisynth 2.5 plugin???

[sneaker_ger]

Try loadCplugin("assrender.dll").

[twazerty]

Try loadCplugin("assrender.dll").

Thanx. Lines show up when I test it in MPCHC on Windows but avs2bdnxml doesn't detect any lines on linux via wine.

Edit: After putting some fonts in the fonts dir in Wine it runs fine and it produces correct images. Huge Thanx!!!

[ps auxw]

Good to hear you got it working for you, although, since you're not using VSFilter anymore, you might want to give PunkGraphicStream (http://forum.doom9.org/showthread.php?t=148030) a try. It's based on the same subtitle rendering library as assrender.dll.

Apart from that, I don't suppose anybody has given the reworked SUP output a try yet?

[river1]

avs2bdnxml crashed when using -o twice
like this
avs2bdnxml -t Undefined -l und -v 1080p -f 23.976 -a1 -b1 -p1 -m3 -o output.sup -o output.xml input.avs
CPU: Using SSE2 optimized functions.
avis [info]: 1920x1080 @ 23.98 fps (4300 frames)
Progress: 112/4300 - Lines: 1Assertion failed: pal != NULL, file avs2bdnxml.c, line 1145

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.

[ps auxw]

avs2bdnxml crashed when using -o twice
like this
avs2bdnxml -t Undefined -l und -v 1080p -f 23.976 -a1 -b1 -p1 -m3 -o output.sup -o output.xml input.avs
CPU: Using SSE2 optimized functions.
avis [info]: 1920x1080 @ 23.98 fps (4300 frames)
Progress: 112/4300 - Lines: 1Assertion failed: pal != NULL, file avs2bdnxml.c, line 1145

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
Oops. Thanks for reporting this. avs2bdnxml v2.05 (http://ps-auxw.de/avs2bdnxml/avs2bdnxml-2.05.tar.bz2) should fix the problem.

[FILLIFY]

easyavs2bdnxml
GUI supports last avs2bdnxml v2.05.
Original Chinese version's here http://bbs.siluhd.com/thread-385883-1-1.html
English/Russian version (with authored interface) is here http://www.adrive.com/public/n7BQPV/...nxml-0.3.4.zip

[OptimusX]

I was last around for this program during the early revisions, but now I'm back with renewed interest and happy to see SUP export and some solutions to the buffer overflow problems! I have some simple colliding subs that appear on the top and bottom that I could never get to stop flickering, but now I'm going to give this another go!

[OptimusX]

easyavs2bdnxml
GUI supports last avs2bdnxml v2.05.
Original Chinese version's here http://bbs.siluhd.com/thread-385883-1-1.html
English/Russian version (with authored interface) is here http://www.adrive.com/public/n7BQPV/...nxml-0.3.4.zip

Hmmm it seems the english version is no longer available. I think your posted link is corrupted...I get "...nxml-0.3.4.zip" as part of the link.

[qknet]

Try AssRender (http://forum.doom9.org/showthread.php?t=148926) instead of vsfilter.

Can you give an example of avs file? I have failed to produce one to work with avs2bdnxml.

Currently I stole the avs created by EasySub, which works, but it uses vsFilter. I would like to see if AssRender can make it faster

loadplugin("C:\Program Files\AviSynth 2.5\plugins\VSFilter.dll")
nw=1920
nh=1080
b=MaskSub("D:\Documents\Dropbox\subtitles\test.ass" ,nw, nh, 23.976, 161478)



d=b.killaudio

return d