unai-d / unai.vitc Goto Github PK

Hi.
Thanks for the nice tool.

I noticed that bit faster output at the 30fps.

The default duration is 1m40s while the output is 1m23s.
The result of readvitc by ffmpeg is also bit faster.

An example of command is below.
Unai.VITC.exe | ffmpeg.exe -f rawvideo -video_size 90x1 -pixel_format gray -framerate 30 -i - -f mp4 -c:v h264 -pix_fmt yuv420p -vf scale=360:4:flags=neighbor vitc90x4-30fps.mp4

At 25fps, works will fine.

Are there any possible causes?

Thanks.

Congratulations on the first release!

Just some little things I thought I would mention making things easier and or more useful to users.

Accessibility

ffmepg/.net runtimes should be bundled in the release for self-contained operation (avoids the whole chicken and egg archive issue)

Detailed Example Commands & Test Data

Ideally, the readme would be expanded on its okay but it needs more demo examples for people.

• SMPTE/EBU Colour Bar Tests

• Commands to take 720x576 PAL & 720x480 NTSC inputs and pad to broadcast standard 720x608 & 720x512 for example.

• Considerations for commands to generate data suited for ld-chroma-encoder as that allows software side generation of full spec 4fsc composite or s-video data in .TBC and it can be played back to analogue via a 10USD Fl2000 VGA adapter on Windows/Linux FL2K

would this project be able to generate the tbc from ether video or hacktv output and pipe it to the hacktv as input
for use with analouge transmition?
https://github.com/fsphil/hacktv

could this be possible?

thank you.

Sincerly,
Michael

Firstly, love the concept of the Unai.VITC tool. Thank-you for putting this utility together.

I have successfully compiled the tool for macOS x86 using dotnet & .NET3.x using...
$ dotnet publish -r osx-x64 -p:PublishSingleFile=true --self-contained true

However, I am struggling to use the example commands on the README.md to create a result that produce a result that FFmpeg readvitc will recognize from the first frame. Would you be kind enough to share the commands that you used to create the APNG examples?

Here is the command I ran (based on the example in README.md). I have tried various framerates, padded inputs, overlays, unscaled, various scale-factors etc. I have also tried using an intermediary file, rather than a pipe.

# Long command is broken up into separate lines with \ for readability on Github...
#
$ ./Unai.VITC -fps 25 -t "00:01:00:00" -tc "00:00:00:00" \
  | ffmpeg -hide_banner -report \
  -f "rawvideo" -video_size "90x1" -pixel_format "gray" -framerate 25 -i - \
  -filter:v "scale=width=360:height=4:flags=neighbor,readvitc=scan_max=-1,metadata=print:file=readvitc.txt" \
  -t "00:01:00.000" -f "null" - -y
$ cat readvitc.txt

readvitc.txt - unexpected result is the output of readvitc filter showing that no VITC is identified until 4 seconds/100 frames into the file. It then stops detecting at 5 seconds, then starts working at 10 seconds.

lavfi.readvitc.found=0
frame:99   pts:99      pts_time:3.96
lavfi.readvitc.found=0
frame:100  pts:100     pts_time:4
lavfi.readvitc.found=1
lavfi.readvitc.tc_str=00:00:04:00
frame:101  pts:101     pts_time:4.04
lavfi.readvitc.found=1
lavfi.readvitc.tc_str=00:00:04:01
frame:102  pts:102     pts_time:4.08

However, if I run FFmpeg against your animated PNG hosted samples on github, the VITC is detected and decoded from the first frame, as expected. Here is the working.txt result.

$ ffmpeg -hide_banner -report \
-i "https://raw.githubusercontent.com/unai-d/Unai.VITC/main/img/readvitc-pal.png" \
-filter:v "readvitc=scan_max=-1,metadata=print:file=working.txt" \
-f "null" - -y
$ cat working.txt

frame:0    pts:0       pts_time:0
lavfi.readvitc.found=1
lavfi.readvitc.tc_str=00:00:00:00
frame:1    pts:4000    pts_time:0.04
lavfi.readvitc.found=1
lavfi.readvitc.tc_str=00:00:00:01
frame:2    pts:8000    pts_time:0.08
lavfi.readvitc.found=1
lavfi.readvitc.tc_str=00:00:00:02
frame:3    pts:12000   pts_time:0.12

I can imagine how you have used FFmpeg's readvitc, metadata, drawtext filters to create your animated PNGs. I'm familiar with using metadata and drawtext filters.

Would you be kind enough to share your full command for generating the APNG's and try my command on your build, in case it is specific to my build. I'm running a recent version of FFmpeg.

$ ffmpeg -version
ffmpeg version 4.4 Copyright (c) 2000-2021 the FFmpeg developers
built with Apple clang version 12.0.0 (clang-1200.0.32.29)

Thanks for the tool. I've been looking for a tool to generate VITC and embed 608 captions as Line 21 for a few years. Your project is the first open-source vitsee rendering tool. Appreciated.

Dear developers,
I am a noob when it comes to programming and I have no clue how to use these c# files.
But I REALLY need this tool for a large scale event where I will do video mapping. (I am a VJ)
If anyone can compile this to an executable for either Mac or Windows that I can use, I would be forever grateful!