1. Sam Bull
  2. pygame
  3. Branches

Branches

tylerthemovie

View source
tylerthemovie
  • Contributors
    1. Loading...
Author Commit Message Date Builds
zeroth
Merged with trunk rev 2613, and moved interactive movie test and sample to __movie_test.py.
zeroth
Pre-check in before merging with trunk. Fixed video display issues thanks to an always false if conditional.
zeroth
Added new doc file as per Vicent Martin's suggestion in the doc folder. As well, added basic GPL and copyright notices to each of my files, and began documenting the functions and their uses within the code.
zeroth
Fixed surface issue. Surfaces now don't get resized. It is recommended to make sure that the surface you use is the same aspect ratio as the original video.
zeroth
Added video display destruction on a stop. However, replaying causes a weird issue with corrupted data. Must examine why.
zeroth
Fixed memory leak issue with Pascal's test video. Apparently, the video has a negative duration value, so the moving to a keyframe cycle would continue forever.
zeroth
Fixed sound issues(hopefully). I did not hear static on my end, and neither did any friends. However, I did take the time to make the sound process more efficient, do less copying of buffers(before I was copying buffers 3-4 times...XD), and this should also hopefully deal better with memory leak issues.
zeroth
Successful compilation and installation state for module, just uncomment _movie line in Setup.in
zeroth
Began work on modifications for Windows compatibility.
zeroth
Merged with latest trunk rev.
zeroth
Made change to surface display code(no more skipped frames)
zeroth
Fixed seek issues, memory load issue, fixed bug in the movie_test. Added proper sound seeking so that audio and video don't desync after seeking. Issue still not resolved, where the audio and video desync with a spot of stuttering sound, until seek is used after the bad portion, on the Doctor Horrible Ep. 1 video. No other video I've used has this problem.
zeroth
Conditional fix of sound system causes much better syncing. I /think/ it is slightly off in some places, but gets synced back very quickly. Must test.
zeroth
Modifications to try to solve desync problem.
zeroth
Made improvements to sound/video syncing... but they eventually desync unfortunately.
zeroth
Added filepath checks. Also began working on desync issue... corrected audio clock issue which doesn't reset if the video is seeking. Fixed now with a beautiful little calculation and check, based on the pts of audio_packets. Now I realize that the video clock doesn't update wholly properly.
zeroth
Fixed a few more small bugs, created a macro which makes command allocation much more forgetfulness-proof. (Never be idiot proof... this IS C after all...)
zeroth
Tested expansion of command queue to all interactions, fixed several bugs. Also cleared flickering bug... which was a bug in the python test, not my module. Apparently if you don't change the movie to surface mode, and then call pygame.display.update() you get flickering! Silly me. But I did find a bug with the surface application, and determined that things look 100% smoother with RGB24 vs RGBA. So surfaces work, just need to figure out the desync issue...
zeroth
Finished expansion of CommandQueue. Results in a much more stable interface. Still a "hiccup" when starting a video up, which can cause desync... As well, using surfaces right now causes flickering. Must investigate.
zeroth
Finished addition of CommandQueue, will add ytop, xleft, surface stuff to command queue now. Also fixed an issue when stopping and playing a video. Will be trying to have ffmpeg write directly to the surface array.
zeroth
Did some basic if conditionals on library version numbers to work around older libraries of ffmpeg. Started construction of a command queue, to properly have the movie module respond to a command fully in time, before processing another.
zeroth
Removed audioconvert.h as it was not needed. Macroed out the subtitle code, as that is the biggest difference between versions of ffmpeg. Have investigated and found that all uses of ffmpeg references in the module are supported all the way to Ubuntu 8.04 and Debian Lenny. Found a small issue where calling pause after a seek causes spotty sound issues due to a conflict of threads. Current solution is to wait at least half a second before calling pause. Will consider how to make them safe to use so close together.
zeroth
Added seeking, but discovered desync issue. Working on solving it...
zeroth
Began working on seeking.
zeroth
Fixed slowdown issue that happened on odd-numbered playthroughs
zeroth
Finished trunk rev merge 2544
Lenard Lindstrom
addition of ffmpeg to Windows deps build chain; incomplete
Lenard Lindstrom
unpdate Windows deps build chain in preporation to adding ffmpeg
zeroth
Finished small stats package for checking render times. Surface rendering is much more consistent than Overlay rendering. Why?
zeroth
Updated backends to (mostly) mirror the functionality of the finished member in the ffmpeg backend.
zeroth
Added "finished" member to the movie player. Will be true when the video has finished playing.
zeroth
Rolled back threaded code. Moved only audio to its own thread, to not be stopped by the frame issues of large surfaces. Wrote a slightly more optimized routine for transferring data from an AVPicture to an SDL_Surface.
zeroth
Rolled back to revision 2506 before threading changes...
zeroth
Working on moving video and audio decode to threaded. Video is not displaying. Investigate, use profilers...
zeroth
Fixed reference issue, deleting works properly, except fifth time in a row.
zeroth
Finished subtitle support, sort of.
zeroth
Began subtitle support work. Need to alter decoder, and add a subtitle_render function.
zeroth
Fixed surface crash issue.
zeroth
Been cleaning up code, writing comments for more obtuse areas. Unfortunately, it segfaults right now with surfaces... must fix.
zeroth
Expanded unit-tests for ffmpeg backend. Will work on unittests for other two backends.
zeroth
Worked on fixing surface corruption issue, waiting on an answer from libsdl mailing list. As well, worked on bringing other backends up to interface compat with ffmpeg backend. Unfortunately vlc backend is... bad. It doesn't seem to work...at all.
zeroth
Merged with trunk rev 2482
zeroth
Fixed overlay release and surface resizing bug. Surface however remains corrupted. Need to test changes with full test...
zeroth
Fixed occassional speedup after unpausing. Now we just save and then restore the audio clock after unpausing.
zeroth
Fixed crashes on loops.
zeroth
Error is either an unpredicted code-path, or somehow a length of zero... mix->alen==0 causes the error.
zeroth
Changes made to try to locate the source of the looping issue. Looks like calling a thread swap causes a deadlock...
zeroth
Trying to find why looping causes crashes... yet again.
zeroth
Fixed unpausing issue. No more drift!
zeroth
Merged with trunk rev 2455.
zeroth
Fixed video and sound issues...however video is very fast after unpausing. This is likely related to last_showtime..
zeroth
Got the audio sync and stutter issue mostly solved.. Video is weird now, slowing down and speeding up. At least its synced up! Hmm...
zeroth
Fixed unpause issue, fixed sound restart issue, and started working on sound/audio drift issue. When synced to audio, the sound is very staticky and not right.
zeroth
Finally got video looping to work. Audio still doesn't work... but everything's just a bit less hackish now! Yay functions.
zeroth
Last code state, looping still does not work. Waiting on info from ffmpeg people.
zeroth
Further refactoring, to help solve loop issue. Apparently. the AVFormatContext struct needs to be reinitialized, as the streams have reached their EOF. I need a way to... initialize them properly.
zeroth
Fixed up some kind of hacky code I was worried about, fixed a small sound issue. Now it doesn't loop... yet again...
zeroth
Added fixes for my desktop, which was causing some double free errors in cb_mixer in gsound.c
zeroth
Testing, and attempts at understanding the cause of the bugs. Current major bugs: -unpause doesn't work -on a loop or replay, sound doesn't work -audio and video drift out of sync I'm also concerned that I need to rewrite and refactor a bunch of code to handle better the stuff solely needed for initialization, then routines for the stuff needed for playback, and the same for shutting down. It will be like this: open, start, end, start, end, start, end clo…
zeroth
Made some small alterations, trying to fix audio and unpausing.
zeroth
Tested on desktop computer, fixed issue I was having.
zeroth
Fixed an issue where pausing would not pause sound. Unpausing still doesn't work quite right.
zeroth
Fixed output of tonnes of error messages from ffmpeg. However, unpause doesn't quite work right still...
zeroth
Fixed up some play/stop/pause issues, used better thread primitives in several functions, fixed a few deadlocks. Discovered for updating the display, I need to let the GIL loose with a call to time.sleep before updating, else we get a deadlock.
zeroth
Fixed various bugs in stop and pause. Managed to eliminate most of the music errors that came up before, stuck only with the video errors spouted off by ffmpeg. Must investigate further. Hrm.
zeroth
Changed stream_open to not return a struct, which for some reason was not working right on my desktop at home... but not on my laptop.
zeroth
Merged with trubk rev 2393, and made some fixes in sound and movie interaction.
zeroth
Added fixes for stopping and restarting playback. Forgot to set abort_request back to zero... whoops. Also fixed some issues in sound, where I didn't call the right functions at the right times, as well as checks within the sound module itself. Had to add a new channel int to the audio portion of PyMovie. In addition, fixed a few small bugs in my test script.
zeroth
Added a sample test file, and made a small modification to gmovie.c
zeroth
Merged with trunk rev 2383
zeroth
Fixed up the unit test slightly, as an example of how to use the movie module.
zeroth
Added vlc backend module, which minimally works, and a dummy-backend module which has the right interface and members, and right status indicators, but does nothing. Last-ditch fallback module. Added single module which deals with importing which movie module, does the system checks for a file...
zeroth
Fixed problems in unit test file.
zeroth
Added beginnings of ffplay wrapper code.
zeroth
Moved unit test to proper directory. Began modifications of ffplay to listen on stdin to make it controllable, added a _backend member to the ffmpeg wrapper code.
zeroth
Found endian flags, now it is little vs big endian portable.
zeroth
Fixed color issue. Its an endian-ness issue, so need to find a flag for ifdefing.
zeroth
Fixed threading issues and finally have a snappy gil-response! As well, fixed a small issue with movie->resize being expanded to indicate only either height or width has been changed, ie, resize_w or resize_h.
zeroth
Added basic documentation
zeroth
Merge update, added resizing to movie module. Cleaned up several really long functions to be smaller and simpler.
zeroth
Renamed the test file.
zeroth
Renaming work, added support for misc. surfaces. Unfortunately, due to threading issues, its... uh... not very effective. No image gets displayed for some reason...
zeroth
Working on adding surface support.
zeroth
Cleared out old, non-working code.
zeroth
Worked on various issues in the files.
zeroth
Fixed some closing issues, and turned off all the extra writes to the console.
zeroth
Finally made sound work!
zeroth
Intermittent work on sound
zeroth
Fixed pointer/address issues of buffers, hopefully.
zeroth
Added timestamp representation. Trying to call repr seems to hang after playback... due to the other threads grabbing and releasing the GIL all the time.
zeroth
Put module back into workable state, turned off audio connections. Must read up on buffers/pointers, etc.
zeroth
Added sound library, began attempts at making sound work properly.
zeroth
Added new sound stuff, bunch of deadlocks at the moment. Must plan out better.
zeroth
Made loops argument work like the loops argument for Sound
zeroth
Fixed memory leak issues. Now only gains 0.1 MB per play.
zeroth
Slight fixes, still a memory leak, ~2.625 MB/play
zeroth
Worked on proper looping construct(a wrapper around decoder), to minimize memory leaks. Memory leak is still continuing.
zeroth
Fixed accidental increment when I should have decremented, added decrements for some hanging returns.
zeroth
Fixed minor pointer issues.
zeroth
Moved code-base into threaded mode, fixed basic threading issues. Fixed stop, pause, and play issues.
zeroth
Fixed potential memory leak, but larger memory leak issue still present.
zeroth
Fixed lots of warnings that gcc was giving.
zeroth
Added height and width members, with only getters for now. Moved main brunt of initialization to stream_open, from decode_thread Moved main brunt of cleanup to stream_close, from decode_thread Still a memory leak issue
zeroth
Cleaned up code, fixed loop issues. Currently a small memory leak of ~1.124 MB per loop is observed.
zeroth
Looping works properly on positive number integers. If you put in a 3, you will get 3 plays. Infinite play(negative integers) causes a seemingly random lockup/infinite loop, as well as a memory leak.
zeroth
Looping works once through, check further looping... -Tyler
zeroth
Video is properly synchronized in the absence of audio info. :D -Tyler
zeroth
Further progress, made it so queue_picture displays a picture when there is one meant to be shown. For some reason, this shows all frames, just really fast. -Tyler
zeroth
Added video-queue, but some images are still being overwritten. Must investigate. -Tyler
zeroth
Added proper loading of the pts to the frame, according to page 5 of the ffmpeg tutorial
zeroth
Experiments with synchronizing, and audio. Did not work, video still not syncing properly. Audio just gives a bunch of errors and a big squawk. -Tyler
zeroth
Fixed green, offcenter bug, fixed dying at loop 3606(shutdown did not happen), fixed timing, but timing still needs audio to work correctly, I think. Probably should look at how others do syncing, especially on video files with no audio. -Tyler
zeroth
Finally got an image working. It is however, green-shifted, and for some reason wraps around the block. The timing as well needs to be fixed.
zeroth
Changed over to thread-free design, _movie_init blocks tillplay is finished. Bug being thrown on line 573:_gmovie.c
zeroth
Completed rebuild, no more streams object. (For now.) Got it compiled, installed, and importing. Testing will come next. -Tyler
zeroth
Modifications and fixes. Will do rewrite of code. Too many bad design decisions. -tyler
zeroth
Newline at end of header
zeroth
Moved video and audio stuff to separate header files, created a generic header file for video, audio, and subtitles.
zeroth
Progress
zeroth
Working on adding all necessary python objects and methods for pyvideostream, pyaudio, and pysubtitle. Also added compiler flag -ggdb
zeroth
Merged to trunk revision 2269
zeroth
Merged with trunk revision 2262
zeroth
Merged to trunk revision 2256
zeroth
Fixed non-init of streams object .gdbinit added for eclipse debugging
zeroth
Merged with trunk revision 2245, lib/midi.py and test/midi_test.py are not merged. -Tyler
zeroth
Merged to trunk revision 2231
zeroth
zeroth
Play still doesn't work.
zeroth
Merged to branch revision 2222
zeroth
Almost operational
zeroth
Updates no merges yet
zeroth
FInally compiled and linked, and installed. The module however does not successfully create a movie object yet.
zeroth
Merged to trunk revision 2214
zeroth
Updates and near-operation.
zeroth
Updates.
zeroth
Merged to revision 2209 in trunk.
zeroth
Compilation works, just need to get linking to work.
zeroth
Compilation fixes
zeroth
Updates, and almost compilation state. -Tyler
zeroth
Managed to start compilation process for ff_movie.c. Still lots of errors to fix before compilation will be fixed.
zeroth
Nearly finished python interfaces. Still need to construct interfaces for the individual streams. -Tyler
zeroth
Massive updates to ff_movie.c, lots of memory holes fixed. Just need to do python interface functions now. Also merged with trunk revision 2185
zeroth
Merged trunk changes up to revision 2180
zeroth
Merged changes up to revision 2173
zeroth
First workings of the movie module in the branch. Unittests already developed, about halfway between adapting ffplay code to my design. -Tyler
zeroth
Started movie branch. -Tyler