[PATCH][RFC] TechnoTrend PVA Demuxer
Hi, In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one. http://wiki.multimedia.cx/index.php?title=PVA http://samples.mplayerhq.hu/PVA/ All samples play correctly. I did not however implement seeking yet. That's the RFC. How should I proceed? There's no index. PVA files are encapsulated MPEG-ES (for video) and MPEG-PES (for audio) packets. Should I just rely on the caller (MPlayer, ...) feeding it a new byte-position and resync on the AV-syncword at the start of read_packet() ? And if I do so, can I use url_fseek or does that ruin the ability to pipe the input to ffmpeg? --Ivo
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
Hi,
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
http://wiki.multimedia.cx/index.php?title=PVA http://samples.mplayerhq.hu/PVA/
All samples play correctly.
I did not however implement seeking yet. That's the RFC. How should I proceed? There's no index. PVA files are encapsulated MPEG-ES (for video) and MPEG-PES (for audio) packets. Should I just rely on the caller (MPlayer, ...) feeding it a new byte-position and resync on the AV-syncword at the start of read_packet() ? And if I do so, can I use url_fseek or does that ruin the ability to pipe the input to ffmpeg?
use av_add_index_entry() every time a packet is demuxed and implement read_timestamp() and feel free to add seeking as a seperate patch if you like [...]
+static int pva_read_close(AVFormatContext *s) { + return 0; +}
unneeded ... [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB I wish the Xiph folks would stop pretending they've got something they do not. Somehow I fear this will remain a wish. -- M?ns Rullg?rd
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
BTW, nice initiative :) Diego
On Monday 31 December 2007 18:10, Diego Biurrun wrote:
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
BTW, nice initiative :)
Thanks :) I made a list of what needs to be done: LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty) VIVO Demuxer VQF Demuxer (TwinVQ) XMMS Demuxer? Nemesi? TwinVQ might get problematic. Well, I could write a lavf demuxer for that, but there's no vqf decoder in lavc. MPlayer uses a binary blob to play those files. Perhaps that could be reverse engineered. The Linux codec is a .a file that has quite some information in it. --Ivo
Il Wednesday 02 January 2008 11:58:15 Ivo ha scritto:
On Monday 31 December 2007 18:10, Diego Biurrun wrote:
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
BTW, nice initiative :)
Thanks :) I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty) VIVO Demuxer VQF Demuxer (TwinVQ)
XMMS Demuxer? Nemesi?
you are welcome to port my SL demuxer from demux_ts to libavformat if you like, provided it's still gpl and not lgpl :)
On Wednesday 02 January 2008 12:03, Nico Sabbi wrote:
Il Wednesday 02 January 2008 11:58:15 Ivo ha scritto:
On Monday 31 December 2007 18:10, Diego Biurrun wrote:
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer
wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
BTW, nice initiative :)
Thanks :) I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty) VIVO Demuxer VQF Demuxer (TwinVQ)
XMMS Demuxer? Nemesi?
you are welcome to port my SL demuxer from demux_ts to libavformat if you like, provided it's still gpl and not lgpl :)
Are there any sample files on mphq? I couldn't find them by just looking at file- and directorynames. --Ivo
Il Friday 04 January 2008 11:19:06 Ivo ha scritto:
On Wednesday 02 January 2008 12:03, Nico Sabbi wrote:
Il Wednesday 02 January 2008 11:58:15 Ivo ha scritto:
On Monday 31 December 2007 18:10, Diego Biurrun wrote:
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer
wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
BTW, nice initiative :)
Thanks :) I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty) VIVO Demuxer VQF Demuxer (TwinVQ)
XMMS Demuxer? Nemesi?
you are welcome to port my SL demuxer from demux_ts to libavformat if you like, provided it's still gpl and not lgpl :)
Are there any sample files on mphq? I couldn't find them by just looking at file- and directorynames.
--Ivo
all samples that are listed as dmb/ in this file: ftp://ftp1.mplayerhq.hu/MPlayer/samples/allsamples.txt but apparentlty someone removed them ... BTW, if you decide to go on you will need to export some functions from the mov demuxer and remove the mp4_* functions that I added in demux_ts.c. Thanks, Nico
In "dca.c" the call to "lfe_interpolation_fir" (there's only one call) must be changed to this: lfe_interpolation_fir(s->lfe, 2 * s->lfe, s->lfe_data + lfe_samples + 2 * s->lfe * subsubframe, &s->samples[256 * i_channels], 256.0 /*8388608.0*/, 0 /*s->bias */); This will fix the missing LFE channel when decoding DTS. I'm sorry, I don't know how to create patch files. But the change is simple enough. Hopefully one of the SVN maintainers can take this patch over? Thanks! Gr??e *?* Regards... *Mathias Rauen* Ernst-M?ller-Allee 10 Germany *?* 37520 Osterode dear at madshi.net <mailto:dear at madshi.net> *?* www.madshi.net <http://www.madshi.net>
On Fri, Jan 04, 2008 at 11:41:13AM +0100, madshi wrote:
[...]
Thread hijacking is a good way to get ignored. Also, please learn how to create a patch. Diego
Did I hijack a thread? If so, it was not done intentionally. I meant to start a new thread... Gr??e *?* Regards... *Mathias Rauen* Ernst-M?ller-Allee 10 Germany *?* 37520 Osterode dear at madshi.net <mailto:dear at madshi.net> *?* www.madshi.net <http://www.madshi.net> Diego Biurrun schrieb:
On Fri, Jan 04, 2008 at 11:41:13AM +0100, madshi wrote:
[...]
Thread hijacking is a good way to get ignored. Also, please learn how to create a patch.
Diego _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel at mplayerhq.hu http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-devel
Hi madshi wrote:
Did I hijack a thread? If so, it was not done intentionally. I meant to start a new thread...
See how it looks in most mail readers: http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/60951/focus=61009 You should use "Compose", not "Reply" to create a new thread... People in this this list are pretty strict about the list rules, so also do not top-post. [...]
Diego Biurrun schrieb:
On Fri, Jan 04, 2008 at 11:41:13AM +0100, madshi wrote:
[...]
Thread hijacking is a good way to get ignored. Also, please learn how to create a patch.
If you don't know how to create a patch (the agreed upon way to propose changes) try using the command "svn diff" in the path of the modified source code. -Vitor
Thanks Vitor for your friendly reply. Did use "reply", but entered a new subject line. I thought I'd create a new thread this way. Obviously I was wrong. Heh, as you can see I'm a noob with mailing lists... This may be a stupid question: But if I "Compose", don't I create a top-post that way? What do I need to do to make sure that I don't create a top-post accidently? Thanks! P.S: Will try to use svn diff next time. Gr??e *?* Regards... *Mathias Rauen* Ernst-M?ller-Allee 10 Germany *?* 37520 Osterode dear at madshi.net <mailto:dear at madshi.net> *?* www.madshi.net <http://www.madshi.net> Vitor Sessak schrieb:
Hi
madshi wrote:
Did I hijack a thread? If so, it was not done intentionally. I meant to start a new thread...
See how it looks in most mail readers: http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/60951/focus=61009
You should use "Compose", not "Reply" to create a new thread...
People in this this list are pretty strict about the list rules, so also do not top-post.
[...]
Diego Biurrun schrieb:
On Fri, Jan 04, 2008 at 11:41:13AM +0100, madshi wrote:
[...]
Thread hijacking is a good way to get ignored. Also, please learn how to create a patch.
If you don't know how to create a patch (the agreed upon way to propose changes) try using the command "svn diff" in the path of the modified source code.
-Vitor _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel at mplayerhq.hu http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-devel
Hi madshi wrote:
Thanks Vitor for your friendly reply.
[...]
This may be a stupid question: But if I "Compose", don't I create a top-post that way? What do I need to do to make sure that I don't create a top-post accidently?
Actually, you top-posted again. See http://en.wikipedia.org/wiki/Top-posting#Top-posting . Top-post only have a meaning when you reply to a post, not when creating a new thread.
P.S: Will try to use svn diff next time.
You will have to resend your bugfix as a patch if you want it reviewed for inclusion... -Vitor PS: It's also a good idea to point to the thread where you described the problem: http://permalink.gmane.org/gmane.comp.video.ffmpeg.devel/60896
Vitor Sessak schrieb:
Actually, you top-posted again. See http://en.wikipedia.org/wiki/Top-posting#Top-posting . Top-post only have a meaning when you reply to a post, not when creating a new thread.
P.S: Will try to use svn diff next time.
You will have to resend your bugfix as a patch if you want it reviewed for inclusion...
-Vitor
PS: It's also a good idea to point to the thread where you described the problem: http://permalink.gmane.org/gmane.comp.video.ffmpeg.devel/60896
Thanks Vitor, will start a new thread with a patch and not top-post, anymore... :-)
On Fri, Jan 04, 2008 at 12:48:50PM +0100, madshi wrote:
[...]
Stop wasting our time please, Google can answer your questions just fine (first hit will do). Also, our mailing list page should already have explained all of this, but you chose to ignore it. Come back when you have a clue about email etiquette, thanks a bundle. Diego
Diego Biurrun schrieb:
Stop wasting our time please, Google can answer your questions just fine (first hit will do). Also, our mailing list page should already have explained all of this, but you chose to ignore it. Come back when you have a clue about email etiquette, thanks a bundle.
Diego
Thank you for the kind words. You really have a gift for motivating people.
On Fri, Jan 04, 2008 at 02:04:27PM +0100, madshi wrote:
Diego Biurrun schrieb:
Stop wasting our time please, Google can answer your questions just fine (first hit will do). Also, our mailing list page should already have explained all of this, but you chose to ignore it. Come back when you have a clue about email etiquette, thanks a bundle.
Thank you for the kind words. You really have a gift for motivating people.
I can understand your frustration, now please understand mine. I spelled the rules out quite clearly on the mailing lists page. What do you suggest I do so that they do not get ignored? Diego
Hi Diego Biurrun wrote:
On Fri, Jan 04, 2008 at 02:04:27PM +0100, madshi wrote:
Diego Biurrun schrieb:
Stop wasting our time please, Google can answer your questions just fine (first hit will do). Also, our mailing list page should already have explained all of this, but you chose to ignore it. Come back when you have a clue about email etiquette, thanks a bundle. Thank you for the kind words. You really have a gift for motivating people.
I can understand your frustration, now please understand mine. I
For him to understand your frustration he has to search the list archives and see that the 95% of the top-posters have the attitude of "I want you to spend hours debugging to fix my problem but I'm too lazy to spend 30 seconds to learn not to top-post". Sadly, just the fact that madshi didn't ignore what was said to him about top-posting makes him a rare exception... -Vitor
Vitor Sessak schrieb:
Diego Biurrun wrote:
I can understand your frustration, now please understand mine. I
For him to understand your frustration he has to search the list archives and see that the 95% of the top-posters have the attitude of "I want you to spend hours debugging to fix my problem but I'm too lazy to spend 30 seconds to learn not to top-post".
Sadly, just the fact that madshi didn't ignore what was said to him about top-posting makes him a rare exception...
-Vitor
Thanks, you've been kind and helpful... :-)
Diego Biurrun schrieb:
On Fri, Jan 04, 2008 at 02:04:27PM +0100, madshi wrote:
Diego Biurrun schrieb:
Stop wasting our time please, Google can answer your questions just fine (first hit will do). Also, our mailing list page should already have explained all of this, but you chose to ignore it. Come back when you have a clue about email etiquette, thanks a bundle.
Thank you for the kind words. You really have a gift for motivating people.
I can understand your frustration, now please understand mine. I spelled the rules out quite clearly on the mailing lists page. What do you suggest I do so that they do not get ignored?
Diego, I can understand that you're frustrated with people ignoring the mailing list's rules. I really do understand that. There are often people posting in my own forums ignoring all the rules I set. So I really do feel your pain. But what I don't understand is how you're weighing my actions. Yes, I did violate some mailing list rules because I didn't read the full instructions. That was wrong and I have already apologized for that (and changed my posting style). But on the positive side I have also come with a fully tested and working bugfix for "your" code. I think you're a developer yourself, aren't you? So you surely know that debugging other people's code is no fun. Furthermore I'm a Delphi developer by heart and doing C++ stuff is really painful for me. Now instead of posting a "thanks" for me spending hours on debugging bugs in "your" code you were welcoming me with a "go away" attitude just because I dared to post a bugfix before studying the mailing list's rules from A to Z. Reading your posts it seems to me that the mailing list formatting rules are more important to you than getting bugs fixed in ffmpeg. I don't think that's the right approach. Anyway, no harm done (on my side, at least). P.S: Of course I was and still am not expecting any "thank you"s. After all you're all doing work for the public without getting much reward back yourself.
Hello, On 04/01/2008, madshi <dear at madshi.net> wrote:
Diego, I can understand that you're frustrated with people ignoring the mailing list's rules. I really do understand that. There are often people posting in my own forums ignoring all the rules I set. So I really do feel your pain.
But what I don't understand is how you're weighing my actions. Yes, I did violate some mailing list rules because I didn't read the full instructions. That was wrong and I have already apologized for that (and changed my posting style). But on the positive side I have also come with a fully tested and working bugfix for "your" code. I think you're a developer yourself, aren't you? So you surely know that debugging other people's code is no fun. Furthermore I'm a Delphi developer by heart and doing C++ stuff is really painful for me. Now instead of posting a "thanks" for me spending hours on debugging bugs in "your" code you were welcoming me with a "go away" attitude just because I dared to post a bugfix before studying the mailing list's rules from A to Z. Reading your posts it seems to me that the mailing list formatting rules are more important to you than getting bugs fixed in ffmpeg. I don't think that's the right approach.
Anyway, no harm done (on my side, at least).
P.S: Of course I was and still am not expecting any "thank you"s. After all you're all doing work for the public without getting much reward back yourself.
1) Thank you for your effort. 2) We use C not C++. :) Regards, Rob
On Fri, Jan 04, 2008 at 06:38:09PM +0100, madshi wrote:
Diego Biurrun schrieb:
On Fri, Jan 04, 2008 at 02:04:27PM +0100, madshi wrote:
Diego Biurrun schrieb:
Stop wasting our time please, Google can answer your questions just fine (first hit will do). Also, our mailing list page should already have explained all of this, but you chose to ignore it. Come back when you have a clue about email etiquette, thanks a bundle.
Thank you for the kind words. You really have a gift for motivating people.
I can understand your frustration, now please understand mine. I spelled the rules out quite clearly on the mailing lists page. What do you suggest I do so that they do not get ignored?
Diego, I can understand that you're frustrated with people ignoring the mailing list's rules. I really do understand that. There are often people posting in my own forums ignoring all the rules I set. So I really do feel your pain.
But what I don't understand is how you're weighing my actions. Yes, I did violate some mailing list rules because I didn't read the full instructions. That was wrong and I have already apologized for that (and changed my posting style). But on the positive side I have also come with a fully tested and working bugfix for "your" code. I think you're a developer yourself, aren't you? So you surely know that debugging other people's code is no fun. Furthermore I'm a Delphi developer by heart and doing C++ stuff is really painful for me.
Do not confuse C with C++, they are by no means the same thing.
Now instead of posting a "thanks" for me spending hours on debugging bugs in "your" code you were welcoming me with a "go away" attitude just because I dared to post a bugfix before studying the mailing list's rules from A to Z. Reading your posts it seems to me that the mailing list formatting rules are more important to you than getting bugs fixed in ffmpeg. I don't think that's the right approach.
You were debugging a problem you yourself were having, I was not having that problem. It is welcome that you share your findings with us, but that does not excuse you from following the mailing list rules. I never told you to go away. I told you to come back when you learned how to follow the rules. That's a big difference. My first message also just told you that thread hijacking is a good way to get ignored. This is true. Most thread hijackers do get ignored completely. I chose to point this detail out to you so that you had a chance to avoid this fate. I maintain that it is rude to ask questions that Google answers perfectly on the first hit. We have explained what top-posting and thread hijacking is countless times around here. Spamming the list with more messages about the subject just degrades the signal to noise ratio. Naturally if somebody chooses to explain stuff to you in private (as I have done countless times) that is welcome. best regards Diego
Diego Biurrun schrieb:
You were debugging a problem you yourself were having, I was not having that problem.
Then obviously you've never used a more recent SVN of the libav DTS decoder. (No offense to the DTS decoder guys. It's working great apart from that little but important bug.)
It is welcome that you share your findings with us, but that does not excuse you from following the mailing list rules.
And I already agreed with that and apologized.
My first message also just told you that thread hijacking is a good way to get ignored. This is true. Most thread hijackers do get ignored completely. I chose to point this detail out to you so that you had a chance to avoid this fate.
So you posted with good will, I appreciate that. However, your post also made me feel somewhat like a criminal at the same time. You made it sound as if I was *intentionally* hijacking and violating rules while in fact it was an accident. Btw, I got a reply from one of the DTS decoder maintainers directly after my first (thread hijacking) post. Thankfully fixing bugs in his code seems to be more important to him than some mailing list behaviour rules. He saw that my intention was good and kindly ignored my accidental mistake. Kudos to him.
I maintain that it is rude to ask questions that Google answers perfectly on the first hit.
You know, I thought I knew what "top-posting" is. I thought it would mean adding a post to the beginning of a thread instead of at the end. I just couldn't see how you could top-post (in the way I thought) in a mailing list. That's why I didn't even think about asking google. Anyway, asking a stupid question may be stupid but not rude in my book. Let's stop it here. I'm neither angry nor offended. I hope the same is true with you. So let's just call it a day. Ok?
On Fri, Jan 04, 2008 at 08:11:00PM +0100, madshi wrote:
Let's stop it here. I'm neither angry nor offended. I hope the same is true with you. So let's just call it a day. Ok?
Neither am I, let's call it a day. BTW, I'm looking into getting the mailing list to send out welcome messages that explain the rules better, possibly with a reminder in regular intervals. Hopefully this will reduce incidents like this one. Diego
Diego Biurrun schrieb:
BTW, I'm looking into getting the mailing list to send out welcome messages that explain the rules better
Sounds like a very good idea to me! Just a thought: Maybe you could limit the list of rules in that welcome message to the top 2 or 3 things that go wrong most often according to your experience. That would increase the chance that at least those 2-3 rules would be read... :-) The welcome message could contain a link to a website with a more complete list of rules. Regards, madshi.
Hi Diego Biurrun wrote: [...]
I maintain that it is rude to ask questions that Google answers perfectly on the first hit. We have explained what top-posting and thread hijacking is countless times around here. Spamming the list with more messages about the subject just degrades the signal to noise ratio.
Naturally if somebody chooses to explain stuff to you in private (as I have done countless times) that is welcome.
You do have a point about explaining these things in pvt. -Vitor
Hi madshi, madshi wrote:
This will fix the missing LFE channel when decoding DTS. I'm sorry, I don't know how to create patch files. But the change is simple enough. Hopefully one of the SVN maintainers can take this patch over? Thanks!
have a look at http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2008-January/039886.html if you want to learn how to create patch files. -Kenneth
Kenneth Vermeirsch schrieb:
have a look at http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2008-January/039886.html if you want to learn how to create patch files. Thank you, Kenneth.
Ivo a ?crit :
On Monday 31 December 2007 18:10, Diego Biurrun wrote:
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one. BTW, nice initiative :)
Thanks :) I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty) VIVO Demuxer VQF Demuxer (TwinVQ)
Is that all that is left in MPlayer and not in FFMPEG ? Maybe an exhaustive list deserves to be commited in MPlayer's TODO/Wishlist file ? Also it _might_ be good starting to removing unnecessary demuxers from MPlayer to really know what still needs to be done.
XMMS Demuxer? Nemesi?
Hum, doesn't nemesi "only" do RTP/RTSP ? FFMPEG can already do that iirc. Ben
Hi all, Benjamin Zores wrote: [...]
XMMS Demuxer? Nemesi?
Hum, doesn't nemesi "only" do RTP/RTSP ? FFMPEG can already do that iirc.
The RTSP demuxer in libavformat currently has some problems/bugs (volunteers to fix them are welcome :), while I read that libnemesi should work fairly well... Maybe we can provide libnemesi support until our RTSP demuxer is fixed? Luca B. is our RTSP maintainer, and is involved in libnemesi development too, so maybe he can comment on the opportunity of providing libnemesi support in libavformat. Luca
On Wednesday 02 January 2008 12:19, Benjamin Zores wrote:
Ivo a ?crit :
On Monday 31 December 2007 18:10, Diego Biurrun wrote:
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
BTW, nice initiative :)
Thanks :) I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty) VIVO Demuxer VQF Demuxer (TwinVQ)
Is that all that is left in MPlayer and not in FFMPEG ?
I looked over all the demux_* files in libmpdemux, but I might have missed something, like SL support for mpegts which Nico pointed out.
Maybe an exhaustive list deserves to be commited in MPlayer's TODO/Wishlist file ?
I could do that if others agree.
Also it _might_ be good starting to removing unnecessary demuxers from MPlayer to really know what still needs to be done.
Yes, or at least prefer lavf demuxers over native ones (if available) and then remove them after say six months if nobody complained about them not working as good as before. --Ivo
Hello, On Wed, Jan 02, 2008 at 11:58:15AM +0100, Ivo wrote:
On Monday 31 December 2007 18:10, Diego Biurrun wrote:
On Mon, Dec 31, 2007 at 05:38:49PM +0100, Michael Niedermayer wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one.
BTW, nice initiative :)
Thanks :) I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty)
If someone will work on TiVo feel free to contact me, I have a bit of experience with it (though I lack some samples, there are playback problems with long files > 2 hours - I have a patch for that but not enough samples to be sure if it is right). Greetings, Reimar D?ffinger
On Sunday 06 January 2008 19:40, Reimar D?ffinger wrote:
On Wed, Jan 02, 2008 at 11:58:15AM +0100, Ivo wrote:
I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty)
If someone will work on TiVo feel free to contact me, I have a bit of experience with it (though I lack some samples, there are playback problems with long files > 2 hours - I have a patch for that but not enough samples to be sure if it is right).
I'm currently investigating the different TiVo formats and documenting it on the wiki. Are you referring to ty streams or the encrypted TiVoToGo streams? And to be sure there's no misunderstanding, you have a mostly working demuxer for ffmpeg already? As for sample files, our repo at samples.mphq is a bit lacking. Perhaps somebody reading this list could upload a few other samples? Also, there are no TiVoToGo samples at all. Might be tricky because of the private Media Access Key needed to decrypt them, but perhaps something can be arranged off-list. Or somebody can donate a real device with recorded content on it :-) --Ivo
Hello, On Sun, Jan 06, 2008 at 09:37:00PM +0100, Ivo wrote:
On Sunday 06 January 2008 19:40, Reimar D?ffinger wrote:
On Wed, Jan 02, 2008 at 11:58:15AM +0100, Ivo wrote:
I made a list of what needs to be done:
LMLM4 Demuxer PVA Demuxer TiVo Demuxer (demux_ty)
If someone will work on TiVo feel free to contact me, I have a bit of experience with it (though I lack some samples, there are playback problems with long files > 2 hours - I have a patch for that but not enough samples to be sure if it is right).
I'm currently investigating the different TiVo formats and documenting it on the wiki. Are you referring to ty streams or the encrypted TiVoToGo streams? And to be sure there's no misunderstanding, you have a mostly working demuxer for ffmpeg already?
Sorry, no I meant I specifically have a bit of an idea of MPlayer's demux_ty code and can answer question both about it and about the format as MPlayer "understands" it.
As for sample files, our repo at samples.mphq is a bit lacking. Perhaps somebody reading this list could upload a few other samples?
That would be nice, it took me a lot of asking just to get those few samples we have... Greetings, Reimar D?ffinger
On Monday 31 December 2007 17:38, Michael Niedermayer wrote:
On Mon, Dec 31, 2007 at 01:12:31AM +0100, Ivo wrote:
I did not however implement seeking yet. That's the RFC. How should I proceed? There's no index. PVA files are encapsulated MPEG-ES (for video) and MPEG-PES (for audio) packets. Should I just rely on the caller (MPlayer, ...) feeding it a new byte-position and resync on the AV-syncword at the start of read_packet() ? And if I do so, can I use url_fseek or does that ruin the ability to pipe the input to ffmpeg?
use av_add_index_entry() every time a packet is demuxed and implement read_timestamp()
Ok, I will look into it. It seems that I could use .flags= AVFMT_GENERIC_INDEX, for that and don't need any special code inside read_packet. And then implement read_timestamp() and perhaps read_seek(). But I'll need to RTFS a bit more.
and feel free to add seeking as a seperate patch if you like
Yes, I will shortly (hopefully :) ).
+static int pva_read_close(AVFormatContext *s) { + return 0; +}
unneeded ...
Removed. Patch comitted. --Ivo
participants (11)
-
ben＠geexbox.org -
dear＠madshi.net -
diego＠biurrun.de -
ivop＠euronet.nl -
kenneth.vermeirsch＠elis.ugent.be -
lucabe72＠email.it -
michaelni＠gmx.at -
Nicola.Sabbi＠poste.it -
Reimar.Doeffinger＠stud.uni-karlsruhe.de -
robert.swain＠gmail.com -
vitor1001＠gmail.com