Re: [FFmpeg-user] Yadif_cuda and Pix_fmt
> Am 25.08.2020 um 06:38 schrieb Andy Sheimo : > > I'm trying to use yadif_cuda during a transcode along with converting from > 8bit to 10bit. Command line and complete, uncut console output missing. Carl Eugen ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-user] Yadif_cuda and Pix_fmt
I'm trying to use yadif_cuda during a transcode along with converting from 8bit to 10bit. The process errors out: Impossible to convert between the formats supported by the filter 'Parsed_yadif_cuda_0' and the filter 'auto_scaler_0' Error reinitializing filters! Failed to inject frame into filter network: Function not implemented Error while processing the decoded data for stream #0:0 If I don't set the pix_fmt then the transcode works without issue. I'm trying to understand why I am getting that failure. It would seem to be that the yadif_cuda filter would take the input stream, perform the deinterlacing then the hevc_nvenc encoder should accept pix_fmt p010le. What am I missing here. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] RTMP issue
I had avformat_network_init() in the code before and have found no difference with or without it. I added back and reran the test again with the verbose logging and received the same results and same log. pi@pi2cam:/userland/build/bin $ ./rcam -D dmic_sv -d 180 -q 30 -w 1920 -h 1080 rtmp://a.rtmp.youtube.com/live2/ [tcp @ 0x199bb70] Starting connection attempt to 2607:f8b0:401a:26::8 port 1935 [tcp @ 0x199bb70] Successfully connected to 2607:f8b0:401a:26::8 port 1935 [aac @ 0x19a9620] Qavg: 116.588 [AVIOContext @ 0x199dfd0] Statistics: 0 seeks, 12400 writeouts Thanks please keep the ideas coming, William From: ffmpeg-user on behalf of Chris Miceli Sent: Monday, August 24, 2020 6:05 PM To: FFmpeg user questions Subject: Re: [FFmpeg-user] RTMP issue Hi William, I'm by no means an expert on the matter but I tried having a look for you. At least one issue would be that you are not calling avformat_network_init() which is a requirement for any protocol which needs network [1] I would have tested and tried this for you, however I do not have the hardware laying around. I wish you the best of luck and let me know how it goes. [1] - libavformat/libavformat.h lines 40-41 *Chris Miceli* On Mon, Aug 24, 2020 at 10:17 AM william keeling wrote: > I added the AV_LOG_VERBOSE and got the following output: > > pi@pi2cam:/userland/build/bin $ ./rcam -D dmic_sv -d 5 -q 30 -w 1920 -h > 1080 rtmp://a.rtmp.youtube.com/live2/ > [tcp @ 0x18acb70] Starting connection attempt to 2607:f8b0:401a:7::8 port > 1935 > [tcp @ 0x18acb70] Successfully connected to 2607:f8b0:401a:7::8 port 1935 > > [aac @ 0x18ba620] Qavg: 118.451 > [AVIOContext @ 0x18aefd0] Statistics: 0 seeks, 341 writeouts > > > If you can point me to an example of logging callback I will try and add > it if you think it will yield more useful information. > > Thanks in advance for your ideas and help, > William > > > From: ffmpeg-user on behalf of Moritz > Barsnick > Sent: Monday, August 24, 2020 9:25 AM > To: FFmpeg user discussions > Subject: Re: [FFmpeg-user] RTMP issue > > Hi William > > On Mon, Aug 24, 2020 at 00:40:04 +, william keeling wrote: > > > My question is an issue with a program written to live stream to > > YouTube working is all modes other than rtmp live stream. > > For programming with the ffmpeg libav* libraries, you should usually > turn to the libav-user mailing list, But perhaps we can help. > > > finishes the stream is never processed by YouTube. When the stream is > > piped (pipe:1 URL) thru FFMPEG (demuxed and muxed only -- ffmpeg -i - > > -vcodec copy -acodec copy -f flv rtmp:) the live stream works > > without issues. > > At least that means that your content encoding, bit rate, number of > audio channels, and so on, are correct for YouTube. It's just the > muxing, or perhaps something like a flag or metadata. > > > So, I need to understand what FFMPEG command line is doing with rtmp: > URL that I am not doing with the FFMPEG API. > > My source code is at github.com/wkeeling63/rcam. Any ideas you have I > > would love to hear as I am stumped. > > Too bad ths only compiles on Raspberry Pi, so there's no way to > reproduce your muxing issue. > > Since you didn't provide a log callback using av_log_set_callback(), > shouldn't your application show all the usual ffmpeg/libav* log output > on stdout? Can you provide us with that ouptut? Raise the log level > with e.g. av_log_set_level(AV_LOG_VERBOSE), if required. > > Thanks, > Moritz > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] RTMP issue
Hi William, I'm by no means an expert on the matter but I tried having a look for you. At least one issue would be that you are not calling avformat_network_init() which is a requirement for any protocol which needs network [1] I would have tested and tried this for you, however I do not have the hardware laying around. I wish you the best of luck and let me know how it goes. [1] - libavformat/libavformat.h lines 40-41 *Chris Miceli* On Mon, Aug 24, 2020 at 10:17 AM william keeling wrote: > I added the AV_LOG_VERBOSE and got the following output: > > pi@pi2cam:/userland/build/bin $ ./rcam -D dmic_sv -d 5 -q 30 -w 1920 -h > 1080 rtmp://a.rtmp.youtube.com/live2/ > [tcp @ 0x18acb70] Starting connection attempt to 2607:f8b0:401a:7::8 port > 1935 > [tcp @ 0x18acb70] Successfully connected to 2607:f8b0:401a:7::8 port 1935 > > [aac @ 0x18ba620] Qavg: 118.451 > [AVIOContext @ 0x18aefd0] Statistics: 0 seeks, 341 writeouts > > > If you can point me to an example of logging callback I will try and add > it if you think it will yield more useful information. > > Thanks in advance for your ideas and help, > William > > > From: ffmpeg-user on behalf of Moritz > Barsnick > Sent: Monday, August 24, 2020 9:25 AM > To: FFmpeg user discussions > Subject: Re: [FFmpeg-user] RTMP issue > > Hi William > > On Mon, Aug 24, 2020 at 00:40:04 +, william keeling wrote: > > > My question is an issue with a program written to live stream to > > YouTube working is all modes other than rtmp live stream. > > For programming with the ffmpeg libav* libraries, you should usually > turn to the libav-user mailing list, But perhaps we can help. > > > finishes the stream is never processed by YouTube. When the stream is > > piped (pipe:1 URL) thru FFMPEG (demuxed and muxed only -- ffmpeg -i - > > -vcodec copy -acodec copy -f flv rtmp:) the live stream works > > without issues. > > At least that means that your content encoding, bit rate, number of > audio channels, and so on, are correct for YouTube. It's just the > muxing, or perhaps something like a flag or metadata. > > > So, I need to understand what FFMPEG command line is doing with rtmp: > URL that I am not doing with the FFMPEG API. > > My source code is at github.com/wkeeling63/rcam. Any ideas you have I > > would love to hear as I am stumped. > > Too bad ths only compiles on Raspberry Pi, so there's no way to > reproduce your muxing issue. > > Since you didn't provide a log callback using av_log_set_callback(), > shouldn't your application show all the usual ffmpeg/libav* log output > on stdout? Can you provide us with that ouptut? Raise the log level > with e.g. av_log_set_level(AV_LOG_VERBOSE), if required. > > Thanks, > Moritz > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". > ___ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] Some more thoughts on documentation
On 2020-08-24 02:11, Phil Rhodes via ffmpeg-user wrote: On Sunday, 23 August 2020, 21:27:17 BST, Carl Zwanzig wrote: I think we'll have to disagree on that, most of the basic logic should be fairly clear to someone who knows the 'c' language. …There should always be a narrative description, as short as possible but no shorter, of how a chunk of the code is intended to be used and what the general control flow is. The definition of "chunk" in this context is a matter of opinion, but a general overview of the intent of the code is a massive time-saver. Examples help, but a narrative description is hugely helpful, especially to people who are new to the situation. No, "read the source" is not an answer to this, and neither is doxygen. Doxygen tells you what individual functions do. It doesn't generally tell you what the whole thing does. Microsoft do this sort of thing incredibly well: https://docs.microsoft.com/en-us/windows/win32/directshow/how-to-play-a-file I agree about the value of narrative descriptions. For instance, in reading the DirectShow "How to Play a File" article, I think of how very helpful it would be to have a narrative description of "How FFmpeg calls a video filter", which describes the entry points to a filter and when FFmpeg calls them and what the filter is supposed to do in response. Also, each filter would benefit from a narrative description of what the author intends the filter to do, and how to use it within an FFmepg invocation. FFmpeg is a large and complex system. It is worth thinking of it as a standard library, or an OS, when planning what documentation is useful. Don't think of it as a single function or as a small standalone program. ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] RTMP issue
I added the AV_LOG_VERBOSE and got the following output: pi@pi2cam:/userland/build/bin $ ./rcam -D dmic_sv -d 5 -q 30 -w 1920 -h 1080 rtmp://a.rtmp.youtube.com/live2/ [tcp @ 0x18acb70] Starting connection attempt to 2607:f8b0:401a:7::8 port 1935 [tcp @ 0x18acb70] Successfully connected to 2607:f8b0:401a:7::8 port 1935 [aac @ 0x18ba620] Qavg: 118.451 [AVIOContext @ 0x18aefd0] Statistics: 0 seeks, 341 writeouts If you can point me to an example of logging callback I will try and add it if you think it will yield more useful information. Thanks in advance for your ideas and help, William From: ffmpeg-user on behalf of Moritz Barsnick Sent: Monday, August 24, 2020 9:25 AM To: FFmpeg user discussions Subject: Re: [FFmpeg-user] RTMP issue Hi William On Mon, Aug 24, 2020 at 00:40:04 +, william keeling wrote: > My question is an issue with a program written to live stream to > YouTube working is all modes other than rtmp live stream. For programming with the ffmpeg libav* libraries, you should usually turn to the libav-user mailing list, But perhaps we can help. > finishes the stream is never processed by YouTube. When the stream is > piped (pipe:1 URL) thru FFMPEG (demuxed and muxed only -- ffmpeg -i - > -vcodec copy -acodec copy -f flv rtmp:) the live stream works > without issues. At least that means that your content encoding, bit rate, number of audio channels, and so on, are correct for YouTube. It's just the muxing, or perhaps something like a flag or metadata. > So, I need to understand what FFMPEG command line is doing with rtmp: URL > that I am not doing with the FFMPEG API. > My source code is at github.com/wkeeling63/rcam. Any ideas you have I > would love to hear as I am stumped. Too bad ths only compiles on Raspberry Pi, so there's no way to reproduce your muxing issue. Since you didn't provide a log callback using av_log_set_callback(), shouldn't your application show all the usual ffmpeg/libav* log output on stdout? Can you provide us with that ouptut? Raise the log level with e.g. av_log_set_level(AV_LOG_VERBOSE), if required. Thanks, Moritz ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] RTMP issue
Hi William On Mon, Aug 24, 2020 at 00:40:04 +, william keeling wrote: > My question is an issue with a program written to live stream to > YouTube working is all modes other than rtmp live stream. For programming with the ffmpeg libav* libraries, you should usually turn to the libav-user mailing list, But perhaps we can help. > finishes the stream is never processed by YouTube. When the stream is > piped (pipe:1 URL) thru FFMPEG (demuxed and muxed only -- ffmpeg -i - > -vcodec copy -acodec copy -f flv rtmp:) the live stream works > without issues. At least that means that your content encoding, bit rate, number of audio channels, and so on, are correct for YouTube. It's just the muxing, or perhaps something like a flag or metadata. > So, I need to understand what FFMPEG command line is doing with rtmp: URL > that I am not doing with the FFMPEG API. > My source code is at github.com/wkeeling63/rcam. Any ideas you have I > would love to hear as I am stumped. Too bad ths only compiles on Raspberry Pi, so there's no way to reproduce your muxing issue. Since you didn't provide a log callback using av_log_set_callback(), shouldn't your application show all the usual ffmpeg/libav* log output on stdout? Can you provide us with that ouptut? Raise the log level with e.g. av_log_set_level(AV_LOG_VERBOSE), if required. Thanks, Moritz ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] Some more thoughts on documentation
On Sunday, 23 August 2020, 21:27:17 BST, Carl Zwanzig wrote: > I think we'll have to disagree on that, most of the basic logic should be > > fairly clear to someone who knows the 'c' language. That tends to be true for, say, one function, or a few functions. And sure, OK, if you want to spend a lot of time scanning around in potentially several source files, you can work things out, though it's longwinded. There should always be a narrative description, as short as possible but no shorter, of how a chunk of the code is intended to be used and what the general control flow is. The definition of "chunk" in this context is a matter of opinion, but a general overview of the intent of the code is a massive time-saver. Examples help, but a narrative description is hugely helpful, especially to people who are new to the situation. No, "read the source" is not an answer to this, and neither is doxygen. Doxygen tells you what individual functions do. It doesn't generally tell you what the whole thing does. Microsoft do this sort of thing incredibly well: https://docs.microsoft.com/en-us/windows/win32/directshow/how-to-play-a-file P ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-user] Some more thoughts on documentation
On 2020-08-23 11:07, Carl Zwanzig wrote: …I'm fairly sure that someone will object to a point below. Go ahead, but also suggest something better…. …Original content: … Process to create output docs: … Output organization & formats: … "Rules of the game"? … … This sounds like fun. I'll play! I think it would help to state explicitly some fairly fundamental premises. We should state that documentation is important, and adds value to the project. Something along the lines of a vision and mission and values statement for the documentation effort. In most projects this would go without saying, but in the FFmpeg project I believe these premises aren't accepted by everyone. Sooner or later in some patch debate, we will have to appeal to these premises to justify a documentation change. I think it would help to write down personas[1], in the sense of interaction design, to explain who we think reads the documentation. This will help us make decisions about what level of detail is appropriate, and what kind of material to include. One persona is the person who isn't interested in the details of digital video, they just want to convert from format A to format B, and heard ffmpeg could do that. Another persona is the digital video specialist who is looking for a versatile processing tool. Another persona is the system integrator connecting a camera type A to streaming channel B with FFmpeg providing the glue. Another persona is the devops expert at BigComputerCompany who is deploying FFmpeg to a cloud-based image processing pipeline. Another persona is the digital media software developer who is working on adding support for format X to Ffmpeg. And so on. By defining these personas, we defend their needs against those who might dismiss them as "lazy beginners". I think it would help to articulate quality standards for the documentation. For instance, documentation must be complete. Every parameter to a filter or format must be included in the documentation, and no parameter may be in the documentation which is not supported by the software. Every data type or keyword accepted by a parameter must be defined in the documentation or via at most one click on a hyperlink. Etc. I think it would help to come up with a different global structure for the documentation. Separate tutorial, and cookbook, and reference material. Build a chapter and section structure which suits the content we actually have (the present structure looks like the outcome of a lot of accretion over time without restructuring). Come up with a URL scheme that permits effective cross-linking to shared definitions, with URLs that are stable over time. I think there is no escaping that people are going to be using a lot of different versions of FFmpeg. It is desirable to snapshot the documentation at major version changes, so that someone with an old version of FFmpeg has a chance of finding an old version of the docs that matches their software. Those are a few thoughts to add to the conversation. [1] https://www.interaction-design.org/literature/article/personas-why-and-how-you-should-use-them ___ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
