OGM Development

[Nic]

This thread is just to replace the old OGM thread that got rather large. Please post your OGM suggestions/wishlist and bugs here. And ill keep this up to date over the weeks with development.

Here is OGMuxer powered version of OggMux by Koepi:
http://nic.dnsalias.com/OggMux_Nic.zip
Just place those two files in a directory and that can be used as an alternative for muxing OggMux.

AAC can now be muxed from there and work perfectly well with the current Ogg DShow filters (available at http://tobias.everwicked.com ) using CoreAAC filters for playback.

(The current CVS version of OggMux is based on DirectShow and also has Matroska support, my version will merge into that soon)

The current progress on the unreleased OggDS Filters is as follow:

Now uses CoreVorbis filter for Vorbis decoding rather than internal filter
Uses VSFilter instead of SubTitDS, for Subtitles. Subtitles are now activated from VSFilter rather than from the OGM filter itself. Switching languages, stability is better because of it.
Memory leaks have been fixed.
Eshetic Improvements


A beta to demonstrate the features will be up soon.

Next proposed new features are:
1)
Use of a Stream Switcher instead of outputing audio through the same pin. This will then work in audio switching the same way as matroska does now, and will allow for different audio samplerates/channels to be switched easily and without speeding up or down of the video.

2)
Improve the Unicode support, and general bug fixing.

Please post comments/bugs below. Try to keep them technical and to the point as possible.

Cheers,
-Nic

[Nic]

Almost forgot, Source code for my current branch of OggMux and my ammendments to OGMuxer by Cyrius are at:
http://nic.dnsalias.com/OggMux_Src.zip
http://nic.dnsalias.com/OGMuxer_src.zip

At present the code is a bit rough, it will be neatened up over the next few days.

[BlackSun]

Nic, you know what I want already ;) . Btw you talk about a stream switcher, I have one for you: http://www.corecodec.org/projects/mmswitch

[ssjkakaroto]

hey Nic since you'll be using VSFilter for subtitles do you think it'll be possible to mux other subtitle formats in OGM besides srt?

tia

[Gaia]

Ogm + aac seems to work great but Ogm playback with aac+ audio is very very slow. It might be problem in latest "temporary experimental" CoreAAC build or just me. Anyone else having problems with ogm + aac+?

Because atleast i can't mux aac+ with current tools to that other container what can handle aac audio i can't test if it is CoreAAC problem or somenthing else.

I am testing with music video so my audio track is just 2 channels and no subtitles ofcouse.

I am using Win98SE...

[h9903209]

Thx for ongoing improvement for ogm.

I really want to know whether the problem of the speed of jumping to non-keyframe would be solved or not. I often want to jump e.g. 5,10 seconds "accurately" with speed as fast as avi/mkv. But in ogm if I choose jump to keyframe then the jump cannot be accurate while if I choose jump to non-keyframe the speed is terribly slow... T__T

[Hiro2k]

I have one little problem with the chapters in your new hybrid oggmux :p

When I mux in the chapters I always get this error
"Chapter 44] is not of a known format". So I deleted chapter 44 and then I get "Chapter 43] is not of a known format", and so on. I stopped trying at 35.

I also had a question about the chapters, does it split the chapters when it splits the video? That's what I'm really looking for in an OGM muxing tool, I know Suircy's can do it, but it is in commandline. I was hoping your tool could do it since it uses Suircy's to mux.

[Koepi]

From the oggmux sourceforge forum:

- add the possibility to force subtitles (like when they speak elvish in LOTR). (hehehe, some people sure can't devide oggmux from oggds ;) fortunately we now have the possibility to work on both).

Thank you for your nice work Nic, you're project admin now and can add your sources to the project! (NB: melgish now takes a look at OggDS as well and already tests some improvements...)

Best regards
Koepi

EDIT: @Hiro2k:
could you email me the chapter-file you try to import (find my email address on my webpage). I'll have a look of it.
My first idea is: hit enter after the last line, i think the chapter import function i wrote is always trying to read whole lines to the CR/LF.

[Nic]

Thanks for all the feedback.
1) Yup, Ill try and get different types of subs into OGM. I know nothing about subs though, so its a bit of a steep learning curve.
2) Ill look into the Nero AAC+ thing, it shouldn't make any difference between AAC and AAC+ so its a bit weird. (what could be happening is AAC+ looks like it has a lower samplingrate due to SBR, I think, so it may be a problem with CoreAAC, ill check)
3) Ill work on the seeking, I looked at the code and couldn't spot the problem, but ill try again.

@BlackSun: The morgan stream switcher...nice. Ill look at that asap.

@Hiro2k: Ill fix that, damn I should have noticed that during testing, if you could email me that chapter file too, that would be great (nic at nic.dnsalias.com)

-Nic

[DAvenger]

Error message when I clicked 'Mux File!'

'E:\My' is not recognized as an internal or external command,
operable program or batch file.


Unpacked in E:\My Downloads\OggMux_Nic :rolleyes:

P.S Morgan stream switcher suxx badly.

[Nic]

@DA: Doh! I Put " around the parameters but not around the Exe. Anyway, Ive re-uploaded a fixed version. Thanks for letting me know. :)

Whats a good alternative to the Morgan Stream Switcher then?

[mgerald21]

Hi NIC,

Have you already commited the new/changed OggDS code to the CVS??
I've checked out the code and didn't find the changes you mentioned.

Greets

[Nic]

No as yet, actually I have no access to the Xiph CVS. Xiph haven't been very forthcoming recently. But ill mail them soon. Ill post the source code soon (still testing if its stable) and/or setup a sourceforge project for it.

-Nic

[DAvenger]

Originally posted by Nic
Whats a good alternative to the Morgan Stream Switcher then?

I do not think that Splitter should handle stream switching. Players should :)

[Koepi]

I added OggDS to the OggMux project. You can upload everything there: OggDS, OggDS sources - and you can make a branch in the CVS.

It' s the easiest way to do the development(IMO of course).

DAvenger, if you want CVS access there too I need your sourceforge handle :)=

Regards
Koepi

[DAvenger]

if you want CVS access there too I need your sourceforge handle

I hate CVS and besides, I do not code ;) But thanks.

[ssjkakaroto]

Originally posted by DAvenger
I do not think that Splitter should handle stream switching. Players should :) you limit a lot the users with that DAvenger, i think everyone should be able to switch streams independently of the player they're using plus stream switch with the current oggds works great (at least for me), so a stream switcher filter based on that code should work great too
edit: do you think if nic and koepi work on morgan ss it would still suck? ;)

[Nic]

@DA: ssjkakaroto is right, it does limit the player used. Ill have to look into this more. Also the corecodec guys, I assume, will be working on the morgan stream switcher...what do they use at present to switch streams in matroska?

-Nic

[DAvenger]

External switcher filter is 'ok' :rolleyes: but to put streamswitching code into the actuall Ogm splitter ... hmm /me does not like this idea :( But maybe I got you all wrong :confused: I am not really sure how does it currently work in OggDS. :o

[tiki4]

Well,

currently, switching between streams is done in OggDS. So I think drawing this out of the filter make sense. AFAIK switching in Matroska is done by the player (TCMP or MPC) or via Morgan Stream Switcher. Maybe the OggDS and Matroska developers can ask Gabest to make his MPC internal stream switcher a stand alone DirectShow filter. Well, I guess Gabest would only need some days for that (if he likes to do so).

Cheers,

tiki4

[bond]

Originally posted by Nic
Now uses CoreVorbis filter for Vorbis decoding rather than internal filternic, i hope this doesnt mean that you put corevorbis into oggds.dll!?
i wouldnt totally ban tobias' vorbis decoder but i would give it a low merit value so that it is easy for people to use corevorbis by just giving corevorbis a higher merit...

or a second possibility is (what i would prefer the most) to split the oggds.dll into four or five independent filter, so that everyone really can only take the parts he really wants...
- ogg splitter
- ogg muxer
- vorbis decoder (tobias' one)
- vorbis encoder
- and perhaps the stream switcher (which leads to the next point -> )

Originally posted by Nic
Use of a Stream Switcher instead of outputing audio through the same pin. This will then work in audio switching the same way as matroska does now, and will allow for different audio samplerates/channels to be switched easily and without speeding up or down of the video.
Originally posted by DAvenger
P.S Morgan stream switcher suxx badly.Originally posted by DAvenger
I do not think that Splitter should handle stream switching. Players shouldi really like this idea and i also think (like ssjkakaroto) that there should be external switching possible by a filter
i think using and continue working on morgan switcher is a great idea, cause this filter can also be used for .avi for example, so people just need to install one filter to switch in all containers available...

Originally posted by Nic
Ill post the source code soon (still testing if its stable) and/or setup a sourceforge project for it.Originally posted by Koepi
I added OggDS to the OggMux project. You can upload everything there: OggDS, OggDS sources - and you can make a branch in the CVS.

It' s the easiest way to do the development(IMO of course).i have the same opinion as koepi, a sourceforge project would really help (especially if more people want to work on the source (koepi, melgish) and nic, also think if you loose interest and somebody wants to continue your work than he can just join the project and the sources wouldnt get split up into thousand pieces, sourceforge helps to keep the sources together imho)

but i think it would be better to give the oggds filter an own project page and not put it together with the oggmux project imho as the filter deserve their own project as they are the major part of the ogg/ogm system imho and oggmux will not be only oggds related soon anymore (matroska support etc...)


and last but not least a sourceforge project would also allow posting different components (similar to gabest' guliverkli) for people to choose (for example you could post the worked on morgan switcher and tobias' decoder seperatly from the other components and also an "all-inclusive" oggds.dll which includes all components...)
in that case people would have the possbility to take for example only the splitter and somebody else also has the possibility to choose the all-inclusive version (and both would be happy :) )

[Nic]

Very valid points Bond:
1) Ok, it seems only me that wants to keep them as one DLL file. So ill split them into seperate projects. Although when i distribute them (in an installer) all will be contained in there as seperate files (i.e. I dont want people to be able to download them seperately, but they can install only the ones they want from the installer). Sound fair?

2) Ill talk to the corecodec guys about that switcher BlackSun pointed out. I dont know enough about the stream switching yet to comment

3) Nope, CoreVorbis isn't in OggDS. But at the same the old OggDS Vorbis filter can't be used even with a lower merit. They work too differently. Maybe I can change the one in OggDS to work like CoreVorbis, but in someways Id like to force the use of CoreVorbis, as at present that is the most stable and functioning. (Toff's filters are very nicely written)

4) Yup, a new sourceforge for OggDS would be good. But im hoping Xiph will show interest and we could keep everything there instead. Ill contact them by the end of the week (im waiting on something else first)

Sound cool?

[bond]

Originally posted by Nic
when i distribute them (in an installer) all will be contained in there as seperate files (i.e. I dont want people to be able to download them seperately, but they can install only the ones they want from the installer). Sound fair?yep, that would be definitly the best way to satisfy everybody :)

Ill talk to the corecodec guys about that switcher BlackSun pointed out. I dont know enough about the stream switching yet to commentgabest also wrote a stream switcher (audio) which is only in cvs, no downloadable build on guliverkli till now, so i dont know how and if it works already...
find it here (http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/guliverkli/guliverkli/src/filters/switcher/audioswitcher/)!

imo it would be very usefull to choose a switcher which can be also used with .avi and .mkv so that there is no need to install more than one filter (one for each container for example)...

Nope, CoreVorbis isn't in OggDSperhaps you can include corevorbis in the installer too, i dont think that toff has something against it (and also the "external" switcher could be included) hm, i really like the installer idea :)

But at the same the old OggDS Vorbis filter can't be used even with a lower merit. They work too differently. Maybe I can change the one in OggDS to work like CoreVorbis, but in someways Id like to force the use of CoreVorbis, as at present that is the most stable and functioning. (Toff's filters are very nicely written)i see

Yup, a new sourceforge for OggDS would be good. But im hoping Xiph will show interest and we could keep everything there instead. Ill contact them by the end of the week (im waiting on something else first)this would be also a possiblity, let's hope that they want to support oggds...

[Nic]

Im off seeing a band tonight (funeral for a friend) but tomorrow ill get to #matroska and ask about the stream switching. I agree, using the same stream switcher would be great for all formats. I won't include CoreVorbis, but say "if you need to decode ogg, click on this link." (unless CoreVorbis becomes finalized, then ill include it)

-Nic

[robUx4]

Then ask Gabest or Toff about it.
It's almost all done by the player (disabling pins I think).

[Nic]

Oh, ok. Thanks robUx4. ill ask them about it, or look into it myself soon :)

[kxy]

This has being mentioned before, I am not sure if it is on your list. With old oggDS, you can not seek to the last chapter by using the chapter feature.

[bond]

nic, i just found an "old" oggds filter feedback thread (http://forum.doom9.org/showthread.php?s=&threadid=33854)!
the bugs in there seem to be still present as 0995 was already available (page 2)!

Originally posted by Nic
I won't include CoreVorbis, but say "if you need to decode ogg, click on this link."good idea, you could also point to coreaac and valex' ac3filter and the working switcher too...

[tiki4]

Well,

I really shouldn't interfere here, but splitting OggDS in different components and distributing them independently maybe a good idea. Although if I take a closer look at the guliverkli at the moment I must say, that it becomes really difficult for newbies to handle all those filters and things. Look at VSFilter for example, one has to install VobSub 2.23 package at the moment, manually unregister dvobsub.ax and register VSFilter.dll then. After that everything works as before (pre-configuring the filter for example). Nevertheless I'd like Gabest to bundle those newer tools again into a VobSub package with an installer, such that more inexperienced users can handle the installing as well.

This should be considered IMHO in a new 'OggDS' package as well, maybe with a NSIS installer so that everyone can choose the components he wants without registering and downloading everything manually.

Just my thoughts...

tiki4

[Edit] NIS -> NSIS (too much UNIX) [\Edit]

[DAvenger]

NSIS is nice :D You can even use its on-demand plugin and download/install only selected components ;) Not sure if it is worth it tho :rolleyes:

[Nic]

@tiki4: Everyone should interfere :) Thats how I perceived it too. Sourceforge is nice for developers to work from, but for users to download from, its a little confusing (Downloading the latest ffdshow is a bit of a similar problem)....
A NSIS installer will always include all the filters (I dont want them available seperately) but you can then select the ones you want from it...
(I think thats for the best)

@Bond: Thanks for the link (reading it now) and for the input :)

edit:
@Hiro2k: Thanks for the chapter file! That will be fixed tonight

(Just noticed I didn't zip the OGMuxer source correctly! That will be fixed too)

-Nic

[pixolex]

Originally posted by tiki4
Well,

I really shouldn't interfere here, but splitting OggDS in different components and distributing them independently maybe a good idea. Although if I take a closer look at the guliverkli at the moment I must say, that it becomes really difficult for newbies to handle all those filters and things. Look at VSFilter for example, one has to install VobSub 2.23 package at the moment, manually unregister dvobsub.ax and register VSFilter.dll then. After that everything works as before (pre-configuring the filter for example). Nevertheless I'd like Gabest to bundle those newer tools again into a VobSub package with an installer, such that more inexperienced users can handle the installing as well.

tiki4

OR make a little DESCRPTION in SF about EVERY single file...is very dificult see what's for each file...for a newbi like me :(

i hope gabest read this :P ;)

[Nic]

My version of OggMux has been updated so that the chapter support works (OGMuxer expected only a file name in the koepi style .omx not the whole chapter list).

The links for the download are in the first post of this thread. Thanks goto Hiro2k for the chapter file :)

-Nic

[Hiro2k]

Are you sure you updated the zip? I keep getting the same error as before.

[Koepi]

Hiro2k:

You may get the old file from the browser's cache or proxy.

Try to d/l without proxy, use "shift"+left mouse button click to enforce the reload from the net.

Hope this helps,

Regards
Koepi

[Hiro2k]

I had the correct version. I tried what you said but I got the same results. I forgot to back up the older version, but the new one says 0.9.5.2 Nic Branch Beta. Were you guys actually able to mux in the file I gave you? Is anyone else able to mux in chapters?

[tiki4]

@Nic:

Please have a look at the end of the CoreAAC thread. There is a problem with muxing HE-AAC into another container or rather with de-muxing it from the MP4 container. I didn't investigate too much yet, but Nero 6 exports LC-AAC files, even if the MP4-file is HE-AAC. What you get using OggMux is a AAC stripped off its SBR content. The solution I use at the moment and which works well is to connect 3ivx' MP4 splitter with the Tobias' Ogg Muxer filter directly. Maybe we need direct muxing from MP4 without creating AAC first...

Just a few thoughts,

tiki4

Additional remark: Even MP4IP tools are not HE-AAC aware at the moment. So mp4creator60 and mp4UI can't help either.

[Nic]

@hiro2k: It does work fine with the chapter file you sent me? I do think koepi is right and it got cached somewhere. Try re-downloading, ive just uploaded it again but now the title bar should read:
OggMux 0.9.5.2 - Nic Branch - Beta Build 4

That version seems to definitely work with your chapter file. Please let me know if you can't get that file or if it doesn't work next time you try :) Thanks for all your help :)

@tiki: Ill look into that now :)

-Nic

[kastro68]

If anyone has the time and talent, I was wondering if they are willing to make some icons for ogm.

@Nic, Koepi
With the changes you are making, is their a risk that you could break backwards compatibility?

[Koepi]

kastro: nope.

Icon? Hm. Who's designing a nice icon for us? :) (I wonder if we should abuse that Xiph-fish as it's used in OggDrop [and in the about-page of OgGDS as well])

Regards
Koepi

[DAvenger]

Originally posted by kastro68
If anyone has the time and talent, I was wondering if they are willing to make some icons for ogm.

http://www.radlight.com/davenger/ogm.png

:p Like them? These shall be included in the next alpha of RadLight (we have MKV too :D ) If anyone wants I may post them here :)

edit: sorry, no fish :rolleyes:

[Koepi]

Who needs fish, it smells bad anyways ;)

Those icons would sure be ok as well i think!

Regards
Koepi

[DAvenger]

Ok. Here's the link : http://www.radlight.com/davenger/OGM.zip

If you want more you know where to find them ;) (this weekend)

[kastro68]

@DAvenger

Thanks. They look great :)

[Liisachan]

Tested:
MPEG-2 AAC by PsyTEL Aacenc;
MPEG-4 AAC by QuickTime, Nero;
HE AAC by Nero.

Basically, all works fine. Just sometimes (NOT always) Media Player Classic crashes to try to play OGM with MPEG-2 AAC, but I don't think this is related to the muxer. To begin with, I'm not sure whether or not the difference between MPEG-2 AAC and MPEG-4 AAC is important for OGM.

Originally posted by Gaia
Ogm + aac seems to work great but Ogm playback with aac+ audio is very very slow. It might be problem in latest "temporary experimental" CoreAAC build or just me. Anyone else having problems with ogm + aac+? I have this problem too, but I guess this is related to CoreAAC, because if you use 3ivx to play AAC, the same file (made by Nic's tool) works fine.

Originally posted by tiki4
There is a problem with muxing HE-AAC into another container or rather with de-muxing it from the MP4 container. I didn't investigate too much yet, but Nero 6 exports LC-AAC files, even if the MP4-file is HE-AAC. What you get using OggMux is a AAC stripped off its SBR content. The solution I use at the moment and which works well is to connect 3ivx' MP4 splitter with the Tobias' Ogg Muxer filter directly. This trick works for me too. If you mux OGM using GraphEdit with 3ivx's splitter, CoreAAC can play he-aac ogm fine. I don't know well about AAC, but according to the ADTS spec (http://faac.sourceforge.net/wiki/index.php?page=ADTS) and my binary editor, Nero' s HE-AAC does identify iteself as LC in its ADTS header (profile = 01)... is this wrong? i.e. Nero's HE-AAC = LC + SBR ? or main + SBR? or ...:confused:

[Nic]

Well MPEG-4 LC and MPEG-2 LC AAC is only one bit...infact I wrote a program to change that bit and released it and the source on Hyrdoaudio (but I doubt there is any difference)

As the AAC thread describes, the samplerate needs to be made higher...this should be easy to sort out using any manner of tricks, etc. But ill see how the corecodec team handle it first.

-Nic

ps
Im having such a pain with the audio stream switching. But once I get this bug out of the way its that part done.

[Palikrovol]

Any date for the first improved "looking forward" release? :)

[Phanton_13]

I like a litle but good feture for me in the OggDS filter, its for preselecting audio and subtitles using a simples rules, it cam be incorporate ar level of player but I like it more at level of filter.
exanple, on top rule is the dominat and if the language or sub dont exist skip the rule but this is evident:
----------------------------------------------
if language = spanish play ( play spanish without subtile )
if language = english play ( play englsih without subtile )
if subtile = spanish play sub ( play with subtile spanish and anything language )
if subtile = english play sub ( play with subtile english and anything language )
if language = german & subtitle = german play ( play german with subtitle in german )
----------------------------------------------
For solve a randomly audio with the case in they arent defined the filter play the first track but if exist a default track marked in one tag the filter play it.

[Liisachan]

Originally posted by Nic
Well MPEG-4 LC and MPEG-2 LC AAC is only one bit...infact I wrote a program to change that bit and released it and the source on Hyrdoaudio Such a tool should be handy for Psytel AACenc users :)
because QT doesn't like mp4 movies with MPEG-2 AAC in my experience.

(but I doubt there is any difference)
I verified it...you are right, I think.

I made an MPEG-2 AAC and converted it into an MPEG-4 AAC (by mp4creator60)--the difference is only 1 bit in each ADTS header like you said--, and then made 2 OGMs having the same video--one with mpeg2 aac, the other with mpeg4 aac--using your tool. The result is almost the same in binary, only a few bytes seem to be random but they are probably padding bytes with no special meaning; I didn't find ADTS header(s) anymore. I assume OGM doesn't "save" the ADTS header, and mpeg2 aac and mpeg4 aac will be exactly same in OGM (maybe...).

Anyway thanks again for a nice tool. This is way more handy than graphEdit, when you want to make OGM with AAC.

[Nic]

Just an update:
1) Got the audio stream switching working well. The Morgan Stream Switcher is a bit c**p, but at least it has the same problems with OGM as it does with matroska (i.e. doesn't like swapping between float and non-float samples). Media Player Classic works fine.
(seeking and possible uncontrolled memory growth, need to be fixed but im getting there)

2) Gabest's next release will have HTML parsing in it (it's already in CVS). No need for SubTitDs anymore :) (Tobias hasn't written back to me :( )

3) Spoke to Toff, adding proper SBR support is very easy to add to OGMuxer, but you may have to specify where the AAC your adding is definitely SBR or not in the GUI (its hard to detect it from the .aac file)

Getting there,
-Nic

ps
Well, got stream switching working nice :) Now in MPC, etc you can have any type of audio in there without it speeding up, static, etc.
If I can add SSA and proper AAC+ (to OGMuxer)...ill be well on my way :)

[bond]

Nic,
on the xiph monthly meeting in IRC (on irc.xiph.org server, #xiphmeet channel) the xiph guys were talking about the ds filter!

it seems that someone called "oddsock (http://wiki.xiph.org/Oddsock)" is also working a little bit on the sources too

i think its time to join the xiph people in irc as it seems that they are really interested in oggds now!

[Nic]

@Bond: Thanks for the tip, ill try and hang out there. I've noticed a buzz about it. I can't help but feel it's in reaction to matroska coming together...They're a bit late now ;)

@all: After Sgt_Strider's thread; Just to illustrate again, I am not trying to compete with matroska, im just bug fixing and updating OGM. So please don't talk about the two comparatively in this forum (and especially not in this thread), as it was done to death on the last OGM thread. If you'd like to talk about it privately or on IRC then feel free to PM me or whatever.

Cheers,
-Nic

ps
DAvenger sent me the log link:
http://xiph.org/~jack/200308.txt
(ill mail jack today...haven't spoken to him in ages)

[DaddyC]

i got a problem with OggMux 0.9.5.2 - Nic Branch - Beta 4.

i set everything (inputfile(s), outputfile aso), as soon as i click on Mux File! OggMux launches another instance of OggMux (the gui itself). and sometimes i get the following error message:

Der Befehl ""OggMux.exe"" ist entweder falsch geschrieben oder
konnte nicht gefunden werden.

(it is the errormessage you get at the console when program is not in path or the exe name is missspelled)

edit: on english os it would be '"OggMux.exe"' is not recognized as an internal or external command, operable program or batch file. (i think)

my os is WinXPSP1, im running OggMux from C:\1\ (i thought it has something todo with the path to OggMux.exe, the former one contained spaces), the avi i try to mux is div3 with mp3 audio.

edit: muxing with the old OggMux is working as well as with OGMuxerGUI by zulu.

[Nic]

Please download again from the same link (on the first post of this thread) to get beta build 5 of OggMux.

I don't know why that doesn't work...but if it can't find it normally now it trys to find it by another two methods. So this should work. Make sure when you download it, it says Build 5 at the top :)

-Nic

[DaddyC]

great nic, thanks.

now it is working.

[mgerald21]

Hi Nic,

It seems, that only you are playing around with the oggds filters doing bug fixing and improvments.
Is there a TODO or WISHLIST ??

Maybe I can give you a helping hand :)

[Nic]

Cool, this thread gives links to things people want as well as the original OGM thread that was eventually locked. Ill post my source code to as far as Ive got with OggDS later today.

Seeking needs to be improved, thats one thing I haven't got round to looking at yet, and is in a seperate .cpp file in the splitter so you could start there if you wanted :)

Also, I really want to add support for SSA into OGMuxer, but can't get the motivation ;) If you look through mkvtoolnix on Mosu's site you can see how Mosu's does it with mkvmerge. So you can try that as well if you wanted, ive only made a start on it.

Cheers,
-Nic

[Kurosu]

Originally posted by Nic
2) Gabest's next release will have HTML parsing in it (it's already in CVS). No need for SubTitDs anymore :) (Tobias hasn't written back to me :( )
All the other news are good news, but as we previously discussed, I jump on this occasion to thank him for that.

Now, too bad he extensively use VC++ 7 (.NET) that make me unable to compile all of his work on my VC++ 6. :(

[BlackSun]

Nic, the new filter give me some problems, I think you need to clean a bit the IAMStreamSelect interface because every audio/sub are still available in it. Do you plan to change the CLSID ?

[Nic]

I probably gave you a version where they are turned on (Its useful for testing some things). Normally the subs and audio are turned off on that interface. No I don't plan to change the CLSID because then it overrides the Splitter nicely and other things that search for OggSplitter still can...

...Do you think it would be a good idea to change the CLSID?

-Nic

[BlackSun]

Originally posted by Nic

...Do you think it would be a good idea to change the CLSID?

-Nic

No, in this case keeping the same CLSID is a good idea, this will force people to use the new filter and I won't have to support both filters ;)

Ok for the IAMStreamSelect interface, when you have some time think about me so I can work on it ;)

Keep up the good work mate.

[bond]

Nic,
any news? when will we see the first working version :D

did you contact any of the xiph guys, are they interested?

[Nic]

gotta be quick, not my computer;
1) My computer died :( Gonna delay any news till after wednesday im afraid :(
2) Xiph haven't contacted me
3) Tobias has contacted me, should get the sources to an improved SubTitDS soo :)

-Nic

[ookzDVD]

I don't know what is wrong,
but muxing the chap info with Nic's Ogmuxer the result is the chap info can't be displayed, but use Cyrius's is working well.

[Nic]

Could you be more specific? Is it the latest download build you've used? "chap info can't be displayed" Does that mean you can't see the chapter info in the systray icon menu?

Thanks,
-Nic

[bond]

hey nic,
did you still have no contact with xiph? :(

btw. i just found out that theora already stores video in ogg (yeah i know old news) did you have a look at these files?

[Nic]

Nope, I haven't heard anything. Sorry for the lack of posting...so busy at work. Hopefully it will clear up by next week.

-Nic

[RexManning]

Sorry to interrupt you, but is anyone working on removing the 2 GB file size limit for OGM? 'cause I've a couple of 4 GB OGMs and would love to seek in them :)

TIA

[Sagittaire]

possible to have compatibility between WMW9 and Ogg Vorbis in ogm container ... ?

[ukb007]

Apologise for pushing in like a bull in a china shop. Is there an information page somewhere where one could learn MORE about OGM ? I'd like to use this container for my XCDs because of its inherent error correction.

Regards.

[Animaniac]

Originally posted by ukb007
Apologise for pushing in like a bull in a china shop. Is there an information page somewhere where one could learn MORE about OGM ? I'd like to use this container for my XCDs because of its inherent error correction.

Regards.

Error correction? It's just error resilient. Errors in the file do not prevent playback of the rest of the file, unlike AVI.

[ChristianHJW]

LIke animaniac was stating, OGM has error detection ( EDC ) based on CRC32 checksums, but no error correction. XCD was supposed to use another layer ( mode 2 form 1 ) to store the OGM headers, including another ECC ( Error Correction ) layer which is coming with form 1 ....

[ukb007]

Thanks, Animaniac. Thanks, ChristianHJW.

I wonder if there's more information regarding OGM somewhere outside the forum discussions.

I'm excited about matroska, too, but where can I learn more in PLAIN English ? I have been through the Matroska website, and what I could get made me excited.

I've repeatedly tried to visit the Matroska Support Pages from your link at http://cvs.corecodec.org/cgi-bin/cvsweb.cgi/~checkout~/matroska/doc/website/contact/contact.html but repeatedly got an HTTP 404 error. I'll try again later.

I understand that it, too has error correction, and would be suitable for XCDs. Is there a How-To regarding this aspect, or is it too early yet ?

Regards.

[Phanton_13]

The error corection inplement in a A/V file format is a bad question, they produce a big overhead for a good eror correction an this is unaceptable, but the real error corection in the world is based on soft or hard implemetation, a shoft example is if the file has a error detection like ogg if you download a file you can be detect on the fly this bad parts and request it to the server, in hardware the base of it is to try to read the bad data up to obtain a good data, the cd implemetation is a litle complexe but the base is this, they have 2 error algoritmes for the data one is the basic and is imprecise and the other produce de data used in the XCD for the files but is more precise to detect the bad part and it can directly correct the data in the most comon situations, in reality is a hybrid system. we can compare it with a raid 5 system with 8 HD we only have the capacity of 7 and the other disk is used for estore the error corection data, For this teh posibility to have a A/V file format with ECC capabilitis is only a imposible wish for now, in the future with a many times better format compresion than the actual formats and a bigger storage systems they can be posible ( I don't expect this for less than 10 years, not for the technical capabilities only for the market absorbtion of the new tecnologies, optical blue laser disks are now on the marcket but if you see the DVD has ned about 4-5 years for a good marcket implemetation, the new disck need a similar time for the same penetration, you can see my point of view about teh expected time).