FFmpeg-音频链接dshow输入的仅延迟视频流

Question

I'm having a slight issue trying to sync my audio and video up with an acceptable margin of error. Here is my command:

ffmpeg -y -thread_queue_size 9999 -indexmem 9999 -guess_layout_max 0 -f dshow -video_size 3440x1440 -rtbufsize 2147.48M ^
-framerate 100 -pixel_format nv12 -i video="Video (00 Pro Capture HDMI 4K+)":audio="SPDIF/ADAT (1+2) (RME Fireface UC)" ^
-map 0:0,0:1 -map 0:1 -flags +cgop -force_key_frames expr:gte(t,n_forced*2) -c:v h264_nvenc -preset: llhp -pix_fmt nv12 ^
-b:v 250M -minrate 250M -maxrate 250M -bufsize 250M -c:a aac -ar 44100 -b:a 384k -ac 2 -r 100 -af "aresample=async=250" ^
-vsync 1 -max_muxing_queue_size 9999 -f segment -segment_time 600 -segment_wrap 9 -reset_timestamps 1 ^
C:\Users\djcim\Videos\PC\PC\PC%02d.ts

My problem is the video comes in slightly ahead of the audio, I can use -itsoffset but then I have to call the video and audio as separate inputs as -itsoffset offsets both audio and video. While this may seem the obvious solution it causes inconsistent audio synchronization if the audio isn't called with the video. Basically if both audio and video aren't called at the same time the video can now be ahead or behind with a 2-3 frame margin. When I call them at the same time the video consistently comes in 2 frames ahead of the audio, every time. I just need a way to delay only the video stream without delaying the audio while keeping both audio and video linked from the beginning. I've tried this with no luck:

ffmpeg -y -thread_queue_size 9999 -indexmem 9999 -guess_layout_max 0 -f dshow -video_size 3440x1440 -rtbufsize 2147.48M ^
-framerate 200 -pixel_format nv12 -i video="Video (00 Pro Capture HDMI 4K+)":audio="SPDIF/ADAT (1+2) (RME Fireface UC)" ^
-flags +cgop -force_key_frames expr:gte(t,n_forced*2) -c:v h264_nvenc -preset: llhp -pix_fmt nv12 -b:v 250M ^
-minrate 250M -maxrate 250M -bufsize 250M -c:a aac -ar 44100 -b:a 384k -ac 2 -r 100 ^
-filter_complex "[0:v] setpts=PTS-STARTPTS+.032/TB [v]; [0:a] asetpts=PTS-STARTPTS, aresample=async=250 [a]" -map [v] ^
-map [a] -vsync 1 -max_muxing_queue_size 9999 -f segment -segment_time 600 -segment_wrap 9 -reset_timestamps 1 ^
C:\Users\djcim\Videos\PC\PC\PC%02d.ts

Just like -itsoffset both video and audio are being delayed. You can delay solely audio with adelay, but there doesn't seem to be a video delaying equivalent.

Any help or advice would be greatly appreciated.

Answer 1

As stated by Gyan in the comments, atrim worked. While it isn't delaying the video it is still lining everything up by ditching part of the audio stream.

ffmpeg - y -thread_queue_size 9999 -indexmem 9999 -guess_layout_max 0 -f dshow -video_size 3440x1440 -rtbufsize 2147.48M ^
-framerate 200 -pixel_format nv12 -i video="Video (00 Pro Capture HDMI 4K+)":audio="SPDIF/ADAT (1+2) (RME Fireface UC)" ^
-map 0:0,0:1 -map 0:1 -flags +cgop -force_key_frames expr:gte(t,n_forced*2) -c:v h264_nvenc -preset: llhp -pix_fmt nv12 ^
-b:v 250M -minrate 250M -maxrate 250M -bufsize 250M -c:a aac -ar 44100 -b:a 384k -ac 2 -r 100 ^
-af "atrim=0.038, asetpts=PTS-STARTPTS, aresample=async=250" -vsync 1 -ss 00:00:01.096 -max_muxing_queue_size 9999 ^
-f segment -segment_time 600 -segment_wrap 9 -reset_timestamps 1 C:\Users\djcim\Videos\PC\PC\PC%02d.ts