I’m trying to add and audio track to my screencast (some programming video, mainly text editor, 1920x1080 avc1 constant framerate) video with shotcut.
However, despite export quality settings, the resulting exported file has blurry text, if it is itself colored (highlighted keywords in text editor) on colored background (looks like compression-artifact). Black text on white background stays crisp.
I wonder if it is a problem of shotcut’s rendering engine or if any workaround exits.
p.s. i have pictures from src video and from dst video of the same fragment
p.p.s. video mode is set to 1920x1080. also tried exporting as bmp-stills, and artifacts are still there
Yes, that is what i’ve done to avoid filters messing with frames: open src and export it without timeline
btw, the artifacts can be seen in preview
but i definitely do not see artifacts when playing src-vid with media player - image looks crisp and readable, like it is a screenshot
also tried fallback to software (vs opengl) - no change
changing “Color Range” from “Limited MPEG” to “Full JPEG” also changes slightly palette in preview (everything becomes slightly lighter), but artifacts are still there
I see the format of your source is yuv444. That means the chroma format is 4:4:4 - one Y, Cr and Cb sample for each pixel. Your export is probably 4:2:0 - one Y for each pixel, but Cr and Cb are shared for 4 pixels. So your chroma resolution will be reduced.
The Shotcut preview window also operates in 4:2:0. so you will not be able to see the full chroma resolution in the preview pane.
For export, you can open the advanced panel, click on the “other” tab, and add this line:
pix_fmt=yuv422p
or
pix_fmt=yuv444p
Other additional parameters may be needed to make those work properly. For example, you may need to choose the H.264 High Profile preset to start with.
Sub-sampled Chroma is the standard way that video is distributed and delivered. You can output non-subsampled chroma from Shotcut to avoid this, but some players and devices may have a problem. Certainly, if you upload it to a web video service like YouTube or Vimeo, it will be down-sampled for their deliveries, which is always 4:2:0.