Hm, yet another definition; "Picture Aspect Ratio" used.
пн, 8 янв. 2024 г., 01:52 Terje J. Hanssen <terjejhanssen@gmail.com>:
Den 07.01.2024 23:35, skrev Andrew Randrianasulu:
пн, 8 янв. 2024 г., 01:31 Andrea paz <gamberucci.andrea@gmail.com>:
I don't quite understand how ffmpeg considers the various aspect
ratios. However from the site lurkertech you pointed to (and from
which I had taken some definitions put in the manual) PAR and SAR are
not the same thing: PAR is about pixel shape and size; SAR is about
frame shape and size.
To change the manual I would like to be sure that ffmpeg really
considers PAR=SAR.
lets wait until ffmpeg-users reply to just send email :)
The wikipedia article section "Confusion with display aspect ratio" points out:
https://en.wikipedia.org/wiki/Pixel_aspect_ratio#Confusion_with_display_aspect_ratio
Pixel aspect ratio is often confused with different types of image aspect ratios; the ratio of the image width and height. Due to non-squareness of pixels in Standard-definition TV, there are two types of such aspect ratios: storage aspect ratio (SAR) and display aspect ratio (abbreviated DAR, also known as image aspect ratio and picture aspect ratio). Also, pixel aspect ratio (PAR) is also known as sample aspect ratio (abbreviated SAR) in some industrial standards (such as H.264[2]) and output of programs (such as ffmpeg[3]). Note the reuse of the abbreviations PAR and SAR. This article uses only the terms pixel aspect ratio, display aspect ratio and storage aspect ratio to avoid ambiguity.
And in Note 3
"ffprobe shows PAR as SAR". ffmpeg.org. Retrieved 2022-06-10.
https://trac.ffmpeg.org/ticket/1776
well, both cingg and ffmpeg can interface with digitizing boards, but after another pass at reading
I hopefully hope that only DEVICE DRIVERS for analog input cards must deal with picture aspect ratio, as they deal with sampling of analog video.
One can also hope that modern smartphone cameras also follow square pixels convention, because they typically do not send video over SDI but hand it into system-on-chip shared memory where encoder/display blocks deal with it directly.
In Settings → Format there is the additional possibility to vary the shape of the pixels from 1:1 (square) to handle anamorphic formats. In such cases we use:
PAR = DAR/SAR
Is it correct that Canvas size "W Ratio" and "H Ratio" define PAR, while "Aspect Ratio" defines DAR in the Settings > Format widget?