[FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
Hello, I am trying to get a framemd5 output from an atomised Avid 1:1 10bit.mxf. I'm using Media Composer 8.1.0. Ignore the mention of v210 in the filename, I was just hoping to run some tests (I was wondering how much lossy degradation occurs when capturing using Avid 1:1 10 bit and then exporting as v210.mov ). I don't know how to compile the latest source so I took a build from http://ffmpegmac.net . I would normally use zeranoe's nightly builds when on windows, but I'm currently on a mac station. I apologise if I've broken the rules. Let me know if you want a sample or mediainfo output. Any help greatly appreciated. Full console output: i:~ admin$ /Users/admin/Downloads/SnowLeopard_Lion_Mountain_Lion_Mavericks_Yosemite_24/ffmpeg -i /Volumes/Stardom/glitch\ tests/v210atomV01.559F9800.770AD0.mxf -f framemd5 /Volumes/Stardom/glitch\ tests/v210atomV01.559F9800.770AD0.framemd5 ffmpeg version 2.7.1 Copyright (c) 2000-2015 the FFmpeg developers built with llvm-gcc 4.2.1 (LLVM build 2336.11.00) configuration: --prefix=/Volumes/Ramdisk/sw --enable-gpl --enable-pthreads --enable-version3 --enable-libspeex --enable-libvpx --disable-decoder=libvpx --enable-libmp3lame --enable-libtheora --enable-libvorbis --enable-libx264 --enable-avfilter --enable-libopencore_amrwb --enable-libopencore_amrnb --enable-filters --enable-libgsm --enable-libvidstab --enable-libx265 --disable-doc --arch=x86_64 --enable-runtime-cpudetect libavutil 54. 27.100 / 54. 27.100 libavcodec 56. 41.100 / 56. 41.100 libavformat56. 36.100 / 56. 36.100 libavdevice56. 4.100 / 56. 4.100 libavfilter 5. 16.101 / 5. 16.101 libswscale 3. 1.101 / 3. 1.101 libswresample 1. 2.100 / 1. 2.100 libpostproc53. 3.100 / 53. 3.100 [rawvideo @ 0x7fdb2b801000] Invalid pixel format. [mxf @ 0x7fdb2b004a00] Failed to open codec in av_find_stream_info [rawvideo @ 0x7fdb2b801000] Invalid pixel format. Input #0, mxf, from '/Volumes/Stardom/glitch tests/v210atomV01.559F9800.770AD0.mxf': Metadata: project_name: mxf v210 test uid : 7d8145ea-c6da-f74b-99c0-846d937db67f generation_uid : b6bcfcab-70ff-7331-e18b-b3bb69de11d2 application_platform: AAFSDK (MacOS X) modification_date: 2015-07-10 10:01:36 product_uid : acfbf03a-4f42-a231-d0b7-c06ecd3d4ad7 product_version : Unknown version product_name: Avid Media Composer 8.1.0.34535 company_name: Avid Technology, Inc. material_package_umid: 0x060A2B340101010101010F001300559F980070AA0E77060E2B347F7F2A80 material_package_name: mxf v210 test Bin.01 Duration: 00:00:08.72, start: 0.00, bitrate: 215006 kb/s Stream #0:0: Video: rawvideo, none, 720x592, SAR 148:135 DAR 4:3, 25 fps, 25 tbr, 25 tbn, 25 tbc Metadata: file_package_umid: 0x060A2B340101010101010F001300559F980070AD0E77060E2B347F7F2A80 file_package_name: Video file reel_umid : 0x060A2B340101010101010F001300559F96D0CC540E72060E2B347F7F2A80 reel_name : v210atomised test timecode: 01:03:23:09 File '/Volumes/Stardom/glitch tests/v210atomV01.559F9800.770AD0.framemd5' already exists. Overwrite ? [y/N] y [buffer @ 0x7fdb2ad01100] Unable to parse option value -1 as pixel format Last message repeated 1 times [buffer @ 0x7fdb2ad01100] Error setting option pix_fmt to value -1. [graph 0 input from stream 0:0 @ 0x7fdb2ad011e0] Error applying options to the filter. Error opening filters! -- This email has been scanned for spam and malware by The Email Laundry. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
On 7/10/15, Kieran O'Leary kieran.ole...@irishfilm.ie wrote: Please also provide the mov file. Carl Eugen ___ This is three frames of colour bars, ingested as avid 1:1 10 bit.mxf, exported same as source.mov. https://dl.dropboxusercontent.com/u/3740512/samesource01.mov -- This email has been scanned for spam and malware by The Email Laundry. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user It looks to me both files are 8bit only. The first one with extra stuff added at begining. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
Kieran O'Leary kieran.oleary at irishfilm.ie writes: I exported the sequence as a same as source quicktime. Please also provide the mov file. Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
On 7/10/15, Kieran O'Leary kieran.ole...@irishfilm.ie wrote: framemd5 won't help you determine 'how much' visual difference there is between an exported file and a reference file. Depending on how you run it, framemd5 would only confirm if the two files decode to the same raw frame or different ones, it's a yes/no answer and not a 'how much'. I suggest sending both streams as input to the psnr or ssim filters which are better able to quantify visual difference. Also ffmpeg is not detecting the pixel format of your mxf file, you may have to declare the pixel format before the input. Dave Rice -- Hi Dave, Thanks for that. I will set the pixel format later and see if that gets it working. Re the framemd5, I should have been more clear. I wanted to use the framemd5s to prove that the video stream was altered when transcoding from 1:1 10bit to v210. I realise that it would just give a yes or no answer to this question. I honestly didn't have a clue how to compare them after that and see how much alteration occurs. Thanks yet again for your invaluable help. AVup is just packed uyvy422. But something is not correct as there is greenish thing on top, also the dimensions doesnt seems to be correct. -K -- This email has been scanned for spam and malware by The Email Laundry. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
framemd5 won't help you determine 'how much' visual difference there is between an exported file and a reference file. Depending on how you run it, framemd5 would only confirm if the two files decode to the same raw frame or different ones, it's a yes/no answer and not a 'how much'. I suggest sending both streams as input to the psnr or ssim filters which are better able to quantify visual difference. Also ffmpeg is not detecting the pixel format of your mxf file, you may have to declare the pixel format before the input. Dave Rice -- Hi Dave, Thanks for that. I will set the pixel format later and see if that gets it working. Re the framemd5, I should have been more clear. I wanted to use the framemd5s to prove that the video stream was altered when transcoding from 1:1 10bit to v210. I realise that it would just give a yes or no answer to this question. I honestly didn't have a clue how to compare them after that and see how much alteration occurs. Thanks yet again for your invaluable help. -K -- This email has been scanned for spam and malware by The Email Laundry. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
Kieran O'Leary kieran.oleary at irishfilm.ie writes: That scares me! The bitrate looks close to v210 (220Mbps vs 213Mbps), but it's actually 8-bit video? You can use the following inlined patch to check for yourself. PLEASE fix your mailer! Paul, will you implement a patch for the original sample? Carl Eugen diff --git a/libavformat/mxfdec.c b/libavformat/mxfdec.c index b3c25b7..e623f50 100644 --- a/libavformat/mxfdec.c +++ b/libavformat/mxfdec.c @@ -1964,6 +1964,7 @@ st-codec-codec_id = (enum AVCodecID)codec_ul-id; } +st-codec-codec_tag = MKTAG('A', 'V', 'u', 'p'); av_log(mxf-fc, AV_LOG_VERBOSE, %s: Universal Label: , avcodec_get_name(st-codec-codec_id)); for (k = 0; k 16; k++) { ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
It looks to me both files are 8bit only. The first one with extra stuff added at begining. ___ That scares me! The bitrate looks close to v210 (220Mbps vs 213Mbps), but it's actually 8-bit video? ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
On 7/10/15, Carl Eugen Hoyos ceho...@ag.or.at wrote: Kieran O'Leary kieran.oleary at irishfilm.ie writes: That scares me! The bitrate looks close to v210 (220Mbps vs 213Mbps), but it's actually 8-bit video? You can use the following inlined patch to check for yourself. PLEASE fix your mailer! Paul, will you implement a patch for the original sample? Nope, its too much broken. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user
Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf
-Original Message- From: ffmpeg-user [mailto:ffmpeg-user-boun...@ffmpeg.org] On Behalf Of Paul B Mahol Sent: 10 July 2015 15:23 To: FFmpeg user questions Subject: Re: [FFmpeg-user] Framemd5 filter error for atomised Avid 1:10 bit.mxf On 7/10/15, Kieran O'Leary kieran.ole...@irishfilm.ie wrote: framemd5 won't help you determine 'how much' visual difference there is between an exported file and a reference file. Depending on how you run it, framemd5 would only confirm if the two files decode to the same raw frame or different ones, it's a yes/no answer and not a 'how much'. I suggest sending both streams as input to the psnr or ssim filters which are better able to quantify visual difference. Also ffmpeg is not detecting the pixel format of your mxf file, you may have to declare the pixel format before the input. Dave Rice -- Hi Dave, Thanks for that. I will set the pixel format later and see if that gets it working. Re the framemd5, I should have been more clear. I wanted to use the framemd5s to prove that the video stream was altered when transcoding from 1:1 10bit to v210. I realise that it would just give a yes or no answer to this question. I honestly didn't have a clue how to compare them after that and see how much alteration occurs. Thanks yet again for your invaluable help. AVup is just packed uyvy422. But something is not correct as there is greenish thing on top, also the dimensions doesnt seems to be correct. I'm getting out of my depth here, but isn't uyvy422 8-bit? Would avup 10 bit be uyvy42210? I'm having a little difficulty with the term packed. Are you referring to the packed/planar distinction? And yes, the 720*592 dimension seems odd. Seems to pop up here as well: https://trac.ffmpeg.org/ticket/162 ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-user