Files - bd.m2ts, bd.meta, bd.iso, and bbb-1920x1080-cfg06.mkv test file - are at: https://cinelerra-gg.org/download/testing/tsmuxer_problem_files.tar My guess of whether or not they will execute the AppImage (link below) is below 1% probability: https://cinelerra-gg.org/download/testing/cin-x86_64.AppImage
On Wed, Apr 27, 2022 at 1:04 PM Andrew Randrianasulu < randrianas...@gmail.com> wrote: > > > On Wednesday, April 27, 2022, Phyllis Smith <phylsmith2...@gmail.com> > wrote: > >> The 30 second and 1 minute 30 second files worked. So I think the last 2 >> patches are good. I am just going to see if I can put the 99 and 999a >> patch in 1 file and test that. If that goes well, I think I can check it >> into GIT. >> > > yes, good idea - sorry I forgot about this detail (5.1 vs 5.1_back) while > preparing patches. Hopefully wider testing will show if I get channel order > right or there is some need for moving audio tracks into specific > configuration... {not sure if this covered in manual} > > > >> I will prepare an appimage and upload the files as you requested but it >> will not be done until later this evening. >> > > > thanks a lot! > >> >> On Wed, Apr 27, 2022 at 12:16 PM Andrew Randrianasulu < >> randrianas...@gmail.com> wrote: >> >>> >>> >>> On Wednesday, April 27, 2022, Phyllis Smith <phylsmith2...@gmail.com> >>> wrote: >>> >>>> The file is Big Buck Bunny bbb-1920x1080-cfg06.mkv and probably came >>>> from: >>>> https://test-videos.co.uk/bigbuckbunny/mkv >>>> >>>> Meanwhile I will try a different file -- I use bbb because it is bigger >>>> than most of my test cases and then is still only 10 minutes long. >>>> >>> >>> >>> I think most tsmuxer users use it from win/mac.... so, our linux-only >>> program driving Linux tsmuxer with hand-hacked ffmpeg might hit some >>> less-known codepath.. so if you can upload resultant m2ts file somewhere >>> I'll link it into issue, because this is what usually asked tehre - >>> problematic ts file to look at.. .. >>> >>> If you can prepare appimage with all those patches applied may be >>> someone will test it on Linux too, but I do not hold my breath for this... >>> >>> >>> >>>> On Wed, Apr 27, 2022 at 11:19 AM Andrew Randrianasulu < >>>> randrianas...@gmail.com> wrote: >>>> >>>>> >>>>> >>>>> On Wednesday, April 27, 2022, Phyllis Smith <phylsmith2...@gmail.com> >>>>> wrote: >>>>> >>>>>> Much better - now tsmuxer and profile value seem to be reset >>>>>> appropriately. >>>>>> But now tsmuxer error: >>>>>> >>>>>> tsMuxeR version 2.6.16-dev. github.com/justdan96/tsMuxer >>>>>> Decoding H264 stream (track 1): Profile: High@4.1 Resolution: >>>>>> 1920:1080i Frame rate: 29.97 >>>>>> H.264 muxing fps is not set. Get fps from stream. Value: 29.97 >>>>>> 0.0% complete >>>>>> Decoding LPCM stream (track 2): Bitrate: 4608Kbps Sample Rate: 48KHz >>>>>> Channels: 5.1 Bits per sample: 16bit >>>>>> 18.2% complete >>>>>> Finalize ISO disk >>>>>> Reading buffer overflow. Possible container streams are not >>>>>> syncronized. Please, verify stream fps. File name: >>>>>> /tmp/bd_20220427-090605/bd.m2ts >>>>>> >>>>> >>>>> >>>>> can you provide example stream for tsmuxer's developers? >>>>> https://github.com/justdan96/tsMuxer/issues >>>>> >>>>> I can enter issue for you, just need link to file in question... >>>>> >>>>> >>>>>> On Tue, Apr 26, 2022 at 7:13 PM Andrew Randrianasulu < >>>>>> randrianas...@gmail.com> wrote: >>>>>> >>>>>>> >>>>>>> >>>>>>> On Wednesday, April 27, 2022, Phyllis Smith <phylsmith2...@gmail.com> >>>>>>> wrote: >>>>>>> >>>>>>>> Adding use_tsmuxer = 0 after use_farmed = 0 did not work (I did >>>>>>>> full rebuild just to make sure). There must be a different spot >>>>>>>> It is possible that the 5:1 audio limitation is Fedora? >>>>>>>> >>>>>>> >>>>>>> no, just diff between av_ch_layout_5point1 and >>>>>>> av_ch_layout_5point1_back .... >>>>>>> >>>>>>> >>>>>>> try attached patches? one is for thirdparty/src another for cin >>>>>>> herself.. >>>>>>> >>>>>>> >>>>>>> >>>>>>>> On Tue, Apr 26, 2022 at 6:26 PM Andrew Randrianasulu < >>>>>>>> randrianas...@gmail.com> wrote: >>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> On Wednesday, April 27, 2022, Phyllis Smith < >>>>>>>>> phylsmith2...@gmail.com> wrote: >>>>>>>>> >>>>>>>>>> Andrew, 10 million "mega" tests later, I finally figured it out. >>>>>>>>>> Neither the chosen "Profile" or "tsmuxer" setting is getting reset to >>>>>>>>>> "Bluray" for profile and "no tsmuxer" for tsmuxer. >>>>>>>>>> >>>>>>>>>> So what was happening is I loaded a file with 5:1 audio, chose >>>>>>>>>> LPCM and TSMUXER and the error message of: >>>>>>>>>> * [pcm_bluray @ 0x7f002b5ff9c0] Specified channel layout >>>>>>>>>> '5.1' is not supported* >>>>>>>>>> which is fine as long as I document that. But then I would load >>>>>>>>>> Big Buck Bunny (also 5:1 audio) and since I saw that tsmuxer was not >>>>>>>>>> checked and lpcm was not highlighted, I thought is should work and >>>>>>>>>> it did >>>>>>>>>> not because both LPCM and TSMUXER will still "on". >>>>>>>>>> >>>>>>>>>> Since I am not sure if I can get it right, could you provide the >>>>>>>>>> lines of code to change so that the profile and tsmuxer are reset to >>>>>>>>>> default "not on" values every time BD Render is chosen from the File >>>>>>>>>> pulldown? Then I can do a final test and check into GIT. Thanks, >>>>>>>>>> Phyllis >>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> try to add 'use_tsmuxer = 0; >>>>>>>>> >>>>>>>>> at this block? >>>>>>>>> >>>>>>>>> int CreateBD_Thread:: >>>>>>>>> option_presets() >>>>>>>>> { >>>>>>>>> // reset only probed options >>>>>>>>> use_deinterlace = 0; >>>>>>>>> use_scale = Rescale::none; >>>>>>>>> use_resize_tracks = 0; >>>>>>>>> use_wide_audio = 0; >>>>>>>>> use_labeled = 0; >>>>>>>>> use_farmed = 0; >>>>>>>>> >>>>>>>>> in cinelerra/bdcreate.C? >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>> On Tue, Apr 26, 2022 at 7:22 AM Andrew Randrianasulu < >>>>>>>>>> randrianas...@gmail.com> wrote: >>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Tuesday, April 26, 2022, Phyllis Smith < >>>>>>>>>>> phylsmith2...@gmail.com> wrote: >>>>>>>>>>> >>>>>>>>>>>> Andrew, if instead of choosing lpcm as the profile with the 2 >>>>>>>>>>>> patches in, I choose the standard bluray, it works. >>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> note, even if file named like c file it actually diff.. >>>>>>>>>>>>> >>>>>>>>>>>>> but both patches as send works for me as after full rebuild.. >>>>>>>>>>>>> / >>>>>>>>>>>>> >>>>>>>>>>>>> if you (re)move those last two patches from tirdparty/src >>>>>>>>>>>>> error (can't find codec..) disappears? >>>>>>>>>>>>> >>>>>>>>>>>> Yes. >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> do you have some other libavcodec headers installed? >>>>>>>>>>> >>>>>>>>>>> what ldd bin/cin (assuming you run from src root for single >>>>>>>>>>> user) says? >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>>> but strange. you rebuild cin executable too? (because we >>>>>>>>>>>>>>> link statically I typed "touch cinelerra/new.C" and this >>>>>>>>>>>>>>> forced rebuild >>>>>>>>>>>>>>> of this file and re-linking of executable...) >>>>>>>>>>>>>>> >>>>>>>>>>>>>> I do full rebuilds by running "make clean" which deletes all >>>>>>>>>>>>>> of the thirdparty libraries and then do a full rebuild with >>>>>>>>>>>>>> "./bld.sh". >>>>>>>>>>>>>> Goes pretty fast on this laptop and it is not worth the >>>>>>>>>>>>>> aggravation to only >>>>>>>>>>>>>> rebuild changed itemsl >>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> does ffmpeg from thirdparty lists pcm_bluray as supported >>>>>>>>>>>>>>> encoder? >>>>>>>>>>>>>>> >>>>>>>>>>>>>> It must because what is checked into GIT, lets me choose the >>>>>>>>>>>>>> lpcm profile and on the bluray player, the info button clearly >>>>>>>>>>>>>> displays it >>>>>>>>>>>>>> as lpcm. >>>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> try thirdparty/ffmpeg-4.4/ffmpeg -codecs | grep bluray >>>>>>>>>>>>> >>>>>>>>>>>>> [root@keystone cinelerra-5.1]# thirdparty/ffmpeg-4.4/ffmpeg >>>>>>>>>>>> -codecs | grep bluray >>>>>>>>>>>> DEAI.S pcm_bluray PCM signed 16|20|24-bit big-endian >>>>>>>>>>>> for Blu-ray media >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> On Mon, Apr 25, 2022 at 2:00 PM Andrew Randrianasulu < >>>>>>>>>>>>>>>> randrianas...@gmail.com> wrote: >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> On Monday, April 25, 2022, Phyllis Smith < >>>>>>>>>>>>>>>>> phylsmith2...@gmail.com> wrote: >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> Andrew, applied ffmpeg-4.4.patch_12 and >>>>>>>>>>>>>>>>>> ffmpeg-4.4.patch99 and it compiles but fails with error >>>>>>>>>>>>>>>>>> message "open >>>>>>>>>>>>>>>>>> failed pcm_bluray:/tmp/bd_date_tme/bd.m2ts" as seen in the >>>>>>>>>>>>>>>>>> attached png >>>>>>>>>>>>>>>>>> file. >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> strange. >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> can you double-check patches applied correctly (by >>>>>>>>>>>>>>>>> comparing pcm-bluenc.c I attached earlier and actual file in >>>>>>>>>>>>>>>>> thirdparty/ffmpeg/libavcodec? ) >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> On Sun, Apr 24, 2022 at 9:18 AM Andrew Randrianasulu < >>>>>>>>>>>>>>>>>> randrianas...@gmail.com> wrote: >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> Paul fixed bug preventing tsmuxer from recognizing our >>>>>>>>>>>>>>>>>>> lpcm audio stream >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> I backshoed fix into our ffmpeg 4.4.. >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> I think having it all worked in monthly release quite >>>>>>>>>>>>>>>>>>> important, even if names of patches kinda high... >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> also attached raw pcm-bluenc file as I hacked it, and >>>>>>>>>>>>>>>>>>> fix/patch as downloaded from gitweb.. >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> https://github.com/justdan96/tsMuxer/issues/580 >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> https://git.ffmpeg.org/gitweb/ffmpeg.git/blobdiff/037fa0437d2aead40c1f786c6ad4d343afe0546d..058a1ff9b4cdbd2e1a93acc8bf0365e7d01cc9f4:/libavcodec/pcm-blurayenc.c >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> https://git.ffmpeg.org/gitweb/ffmpeg.git/commit/93dfb6afddb278218da7f8c696bf35ede2ba05aa >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>
-- Cin mailing list Cin@lists.cinelerra-gg.org https://lists.cinelerra-gg.org/mailman/listinfo/cin