general uniforms support #21
Conversation
|
need more test |
|
This looks awesome @turengege! Will have some time to view more in-depth this weekend 💯 Thanks!! |
|
Int and float vectors are supported now: ffmpeg -i media/0.mp4 -i media/1.mp4 -filter_complex "gltransition=duration=2:offset=1.5:source=fadecolor.glsl:uniforms='color=vec3(0.0,0.5,0.0)',format=yuv420p" -y out.mp4I don't have much C Coding experience. Hope for code reviewing and suggestions. @transitive-bullshit |
| if(sa.len >0){ | ||
| for(int i = 0;i<sa.len;i+=2){ | ||
| GLint location = glGetUniformLocation(c->program, sa.strings[i]); | ||
| if (location >= 0) { |
There was a problem hiding this comment.
let's throw an error if this is not the case; better to be strict with potential errors here rather than ignoring it silently.
There was a problem hiding this comment.
There is an err message can't get location of uniform .. if location is not valid, but I agree that throwing an error is the right choice.
| @@ -98,6 +99,7 @@ typedef struct { | |||
| double duration; | |||
| double offset; | |||
| char *source; | |||
| char *uniforms; | |||
There was a problem hiding this comment.
these uniforms should be freed in uninit
There was a problem hiding this comment.
I think FFmpeg will do this cleaning up. I am not sure, will check it out later.
|
Added a few quick comments -- overall, this looks really great! |
| int cursor_start = 0; | ||
| int cursor_end= 0; | ||
| int cnt_seg = 1; | ||
| auto len = (int) (strlen(str_query)); |
There was a problem hiding this comment.
let's opt for strnlen_s here, as it's safer, and str_querymay be NULL.
There was a problem hiding this comment.
It will require users to use C11 if we choose strnlen_s, because FFmpeg seems to be built on C99 by default.
|
awesome! why not merge it ? |
|
Hello, which ffmpeg version were you using? |
|
+1 |
I think I use the latest version at that time. |
|
@transitive-bullshit @xinghuolei What's missing here to merge? I am willing to step in and help finish this one |
Hi
This feature is unfinished but I think it should help.
Usage:
Limits:
only
intandfloatare supported