matroska | A Rust library for reading Matroska files

I have ben struggling with this for quite some time; Essentially what I'm trying to do is broadcast my webcam from within the Browser to Ffmpeg on the server to convert into a .m3u8 HLS live stream. But I am having a few problems with video playback I initially thought it was CORS and cross origin headers however now I am thinking its to do with the way I'm trying to encode the media.

I did a ffprobe on the input this was the response;

I am recording screen capture of my Electron app to a file, as suggested here: Save captured video to file in Electron

It works great, but the file is a “transport stream”. I can play it in Chrome browser, but can’t adjust the time slider.

The suggestion was to use ffmpeg to post-process the file. The simplest, straight-forward command I found is:

I am trying to convert some mkv videos with hevc codec to avi format. For this I run the following command:

I'm trying to make a discord music bot, but for some reason, the audio just crashes.

Here's the code:

In my case I have 3 .webm files the first one is audio only, the second one is video only, the third one is audio and video.

I want to concatenate them into a single file which shows black screen for audio only parts, video for video only parts, and plays both for the parts that have audio and video.

The video codec is VP8, the audio codec is Opus.

concat.txt contains the entries for the three files

I am using the following command to concatenate them.

I'm trying to scale down a webm video with transparency using this line of commands (based off of an answer in this stackoverflow question @ How to keep transparency when scale webm file with ffmpeg):

ffmpeg -c:v libvpx -i in.webm -c:v libvpx -vf scale=400:416 -auto-alt-ref 0 out.webm

but I keep getting errors (for example):

I have the values 88 40 B0 00. They are is hex. I don't understand how, but they represent a number of nanoseconds. I need help with understand this document which outlines Matroska and WebM metadata encoding. I am relatively new to this, but I am using JavaScript to alter a file duration. What I am currently using works, but the goal is to know how to set a custom duration of the WebM video file. The document shows the following for the duration parameters:

• Element Name
  • Duration
• The level within an EBML tree
  • 2
• Element ID displayed as octets
  • 0x4489
• Mandatory
  • False
• May occur multiple times
  • False
• Contained in
  • 1
  • 2
  • 3
  • 4
• Available to use in WebM
  • True
• Description
  • Duration of the Segment in nanoseconds based on TimestampScale

I am using the default time stamp scale (1000000). My question is how to I get those hex values, and turn them into hours:minutes:seconds:milliseconds. I simply am stumped :( 0x8840B000 doesn't help me.

I know that you are reading this, and I would appreciate a comment if do not have an answer because I feel that I am alone in trying to figure this out.

I have thoroughly read this document. It mentions an element called TimestampScale. I have downloaded probably thirty WebM example files and seen that the values following the HEX { 0x2AD7B1 or 2A D7 B1 or ['2A','D7','B1'] } or NON-HEX integers { 42 215 177 or [42, 215, 177] } are always the same: HEX { 0x830F4240 or 83 0F 42 40 or ['83', '0F', '42', '40'] } or NON-HEX integers { 131 15 66 64 or [131,15,66,64] }. The values all SHOULD be the default of 1000000 as stated in the Matroska Element Specification. So the question is... How does 0x830F4240 end up as 1000000?

Incase Above Was Too Sloppy For You or You Want More Explanation:

I'm trying to write a C++ class handling MP4 movies via ffmpeg.

First I created a couple of functions to use with std::unique<>() so that way things get released even on exceptions.

However, I get a double free when I try to free the BitStreamFilter object, yet the documentation clearly says that each av_bsf_alloc() must be paired with an av_bsf_free() call.

@param ctx a pointer into which the pointer to the newly-allocated context will be written. It must be freed with av_bsf_free() after the filtering is done.

Note: emphasis mine.

However, at the time I call the avformat_close_input() I get a double free error, even if I haven't used the two contexts for anything!? I'm thinking that there may be a packet that both allocate and both try to free. But since these two contexts are not directly connected, I really don't understand how they end up freeing something twice.

Below is code which reproduce the error (at least on an amd64 platform). Once compiled, you can just execute it. Make sure to specify a filename as in: