I have some 2704x2028 (2.7k @ 4:3) video from a GoPro and I would like to extract the center portion pixels to make a 1920x1080 (1080p 16:9) video clip. Based on what I read I should be using the crop source filter to get the best quality, and as far as I can tell, just selecting the Center option on the filter seems to accomplish this, could you please confirm that this is the correct way of doing this? Why if I tried to set the Top, Bottom, Left, and Right parameters manually using 474, 474, 392, and 392 respectively it doesn’t seem to crop the picture correctly?
Using Shotcut 21.08.29
Video Mode: 1080p 60fps
Image size: 2704x2028
Screenshot shows the Crop:Source filter deactivated.
You are wanting to crop the top and bottom of your source video/image to fit the 16:9 aspect ratio of 1080p.
135 Top & 135 Bottom.
With Center applied the same thing is accomplished.
Thanks for the prompt response but I am a bit confused with the 135 pixels number in your response. If the resolution of 1080p 16:9 video 1920x1080 pixels, shouldn’t we be removing 392 from the top and bottom, and 474 from the right and left if my original video is 2704x2028?
Video Mode never changes. You are cropping the clip to the bounds of the Video Mode’s resolution.
When you insert media larger, or smaller to the Video Mode, Shotcut resizes the media to either width or height of the Video Mode, not the source.
Your source of 2704x2028 now essentially becomes 1440x1080 in a 1920x1080 space.
For the width you still have 1920 (blank space on both sides 135px each).
Ok thanks, I see now where the 135 number comes from and what I am doing when I select the crop source with the center option selected, but I am not trying to remove the black bars on the side to get the largest possible frame. I am trying to zoom into my source and extract the inner 1920x1080 pixels from the frame. I would think that I should be able to do it by manually removing 392 pixels from the sides (2704-1920 / 2) and 474 from top and bottom (2028-1080 / 2), but this does not seem to produce the right image when using the crop source filter. I believe that I can also achieve what I want by using an SPR filter by making the size 2704x1920 and setting the position to -392 and -474, but I though it would be better to use the crop source filter. Can you please clarify why the crop source option does not seem to work? Is the SPR filter the only way and best ay to achieve what I want?
Is this what you are trying to accomplish?
This is what the inside 1920x1080 looks like with the aid of GIMP
Setting the SPR filter size to 2560x1920
Crop: Source used to operate in terms of source resolution but no longer since proxy was added because proxy clips are reduced resolution, and the engine does not know anything about proxy. You can see the numeric range of the sliders are based on the video mode resolution. So you simply need to scale your numbers:
Top/Bottom = 474/2028*1080 = 252.426 ~= 252
Left/Right = 392/2704*1920 = 278.343 ~= 278
Now, since you cannot enter decimals enter the rounded down values, but there will be a single column of black pixels on the right edge. To compensate for that, you can add one to the top or bottom to remove the black column. However, that adds a row of black pixels to the bottom; add one to the left or right to get rid of that. Essentially, these +1s are accounting for the fractional part * 2.
This worked for me:
Ok, so I tried this with my video and it worked, thank you. Back to my original question, is using the crop source filter a better option than using an SPL filter for this application, specially given this rounding problem and extra black line?
BTW, this whole thing was no very intuitive and I am not sure if it is documented anywhere. I would like to suggest adding a new filter or modifying the existing crop source filter to facilitate doing this and dealing with the rounding issue. Perhaps another option for Center where instead of removing black lines to match the video mode, it crops as required to match it. I personally shoot video at higher resolutions and 4:3 format to give me flexibility in selecting the best 1920x1080 area within the frame that contains my subject. Ideally I just want to extract the original pixels in a 1920x1080 within the original frame and preserve as much as possible the integrity of those pixels.
Thank you again for your support on this.