Why do I have to specify a deinterlace method in the export window when interlaced is selected. How does the setting affect? Is it relevant for interlaced?
Highwayman
Typically it is ignored. However, there may be a reason that the underlying framework needs to deinterlace the source even if the output is interlaced. Some of the filters only work on progressive frames. The text filter comes to mind as an example.
Oups, that would mean that the clip would first be deinterlaced and then transformed back to interlaced again. But that would result in a loss of quality.
The result clip is definitely interlaced.
I’m currently having problems with the quality of h.264 High Profile exports. The video sometimes stutters when the scene is moving.
Which do you mean:
- The resulting clip is “signaled” as interlaced
OR - When you view the clip you can see interlacing combing artifacts in the images
?
What Video Mode have you selected for your project?
I would suggest to narrow the problem by removing variables. Try this:
- Close all projects
- Set the video mode to your desired interlaced output format
- Open one of your source clips in Shotcut
- Do not put anything on the timeline, do not add any filters or transitions
- Export the source clip using your selected export settings
In this test scenario, I would expect the interlaced fields to pass through Shotcut without any modification.
Does this test show a problem?
@brian, would it be possible to add more Deinterlacing options for the Convert To Edit Friendly menu? There just seems to be a one size fits all approach which doesn’t always give a clean result depending on the video source.
It is possible, but I am not sure what more you would want. In the Convert to Edit Friendly dialog, the Deinterlace option applies the FFmpeg bwdif filter with frame doubling - which is the highest qualify deinterlace filter that is available in the open source community.
Also, OP is asking about the Deinterlacer option in the advanced export panel - which has 4 options. So your comment/question might be off topic.
Answer to the second question:
Video mode is custom:
Width=720
height=576
sample_aspect_num=16
sample_aspect_den=15
display_aspect_num=4
display_aspect_den=3
progressive=0
colorspace=601
frame_rate_num=25000000
frame_rate_den=1000000
Answer to the first question:
Both.
I am curious why you made it custom. Those parameters perfectly match the existing profile for “SD PAL”.
But the output for that profile should be signaled as interlaced if the codec you choose to export supports interlaced…
That is interesting.
What is the Convert To Edit Friendly menu, where can I find it?
Is that meant:
Which option applies the FFmpeg bwdif filter ?
Click on a clip, click on Prperties, at the bottom is the Convert tab. See below for an explanation:
I tested that. Same problems with interlaced output. Video is jerky or shaky in moving scenes. No problems with progressive output.
If I convert the clip with the deinterlace option (Convert to Edit-friendly), I have the same glitches in the resulting progressive mp4 file. What’s wrong with interlaced?
What exactly means “If the source is progressive, each interlaced field will be converted to a progressive frame resulting in double frame rate”?
Do progressive sources have interlaced fields?
I converted the clip with deinterlace and double frame rate (50 fps). Something is wrong with the conversion. Flicker when moving. But if I only select every second frame (selectevery(2, 0) it’s OK.
I do not reproduce this. I am not sure we have determined that interlaced sources are the problem. Maybe there is a different atribute of your clips that cause the glitches. Can you provide screenshots of the glitches?
It is a form of field extension deinterlacing:
No. And that is why the conversion is necessary when going from interlaced to progressive.
What does this mean? How are you selecting frames?
I would understand:
If the source is interlaced, each interlaced field will be converted to a progressive frame resulting in double frame rate.
Or:
If the output format is progressive, each interlaced field will be converted to a progressive frame resulting in double frame rate.
Why should I convert progressive to progressive?
With Avisynth+.
You can’t see it in individual frames, only in a running video. The individual frames are OK, just not the juxtaposition. So it must have something to do with the fields and their compilation / connection / linking. Either wrong order or horizontal line offset (one line).
You discovered a typo in the tool tip. It should say “If the source is interlaced, each interlaced field will be converted to a progressive frame resulting in double frame rate”
I will fix that for the next release.
I do not really know why your clips are not encoding cleanly. I tried my own test using a popular interlaced test clip.
When I export it from Shotcut, the motion is smooth.
Also, I did this experiment with success:
- Open the clip in Shotcut
- Use convert to edit friendly with the Deinterlace option
- In Shotcut, choose the video mode 720p 60 (progressive)
- Open the converted file
- Observe that the clip is 60fps and each frame is unique (no duplicated frames)
- Observe there are no interlaced combing artifacts
Can you try that experiment with the test clip I linked and then again with your own test clip (except I think your clip is 25fps, so the resulting file would be 50fps)?
I tested that.
With the fireball clip there are probably no problems. But there is no camera movement or zoom involved here. And it’s NTSC.
I still have the same problems with my videos (see video link).
And I also have this kind of effects when I simply export interlaced.
What is wrong?