joe_scotto wrote:Coasterghost wrote:CaptainHook wrote:What's New in Blackmagic Camera 1.1
* Support for recording at 23.98, 29.97 and 59.94 fps.
I did some 10 second clip testing this evening and only the 24FPS is still CFR, everything else is VFR.
Some observations of my very short test clips (all are Apple Log at ProRes 422 LT recorded to Internal memeory):
23.98 reads in MediaInfo as 22.999 FPS with a MIN: 22.222 FPS and a MAX: 23.077 FPS
29.97 reads in MediaInfo as 28.993 FPS with a MIN: 28.571 FPS and a MAX: 30.000 FPS
59.94 reads in MediaInfo as 58.995 FPS with a MIN: 54.545 FPS and a MAX: 60.000 FPS
Additional 422 tests for 23.98 and 59.94 are the same exact numbers as well.
24.000 reads in Media Info as Constant
25.000 reads in MediaInfo as 25.000 FPS with a MIN: 24.000 FPS and a MAX: 26.087 FPS
30.000 reads in MediaInfo as 30.000 FPS with a MIN: 28.571 FPS and a MAX: 31.579 FPS
48.000 reads in MediaInfo as 48.000 FPS with a MIN: 46.154 FPS and a MAX: 50.000 FPS
50.000 reads in MediaInfo as 50.000 FPS with a MIN: 46.154 FPS and a MAX: 54.545 FPS
60.000 reads in MediaInfo as 60.000 FPS with a MIN: 54.545 FPS and a MAX: 66.667 FPS
I do also believe that there is a bug on the H265 with Rec2020 - HDR where its showing under automatic settings as peaking while Rec709 is not. but i haven't really tested that option and would require others to test aswell.
Phone used was the 256GB iPhone 15 Pro Max by the way.
Use ffmpeg to check the footage, I can assure you that the 24 footage is actually VFR. Mediainfo only checks the first few seconds of the video and if it's constant, will mark the entire clip as constant when it's not. This is a limitation of the iOS camera API and nothing can be done by BMD unfortunately.
Using in command prompt:
- Code: Select all
ffmpeg -i input -vf vfrdet -an -f null -
23.98: [Parsed_vfrdet_0 @ 000001b3ec16aa40] VFR:0.176471 (54/252) min: 26 max: 27 avg: 26
24.00: [Parsed_vfrdet_0 @ 000002a6ee88c800] VFR:0.000000 (0/255) [/code]
25.00: [Parsed_vfrdet_0 @ 0000023fb6b55b40] VFR:0.017544 (4/224) min: 23 max: 25 avg: 24
29.97: [Parsed_vfrdet_0 @ 0000026f6e872e40] VFR:0.613333 (184/116) min: 20 max: 21 avg: 20
30.00: [Parsed_vfrdet_0 @ 000002917262c800] VFR:0.013812 (5/357) min: 19 max: 21 avg: 20
48.00: [Parsed_vfrdet_0 @ 00000203bfffe500] VFR:0.994071 (503/3) min: 12 max: 13 avg: 12
50.00: [Parsed_vfrdet_0 @ 00000188b9217e00] VFR:0.012963 (7/533) min: 11 max: 13 avg: 11
59.94: [Parsed_vfrdet_0 @ 00000172c80bb140] VFR:0.341308 (214/413) min: 10 max: 11 avg: 10
60.00: [Parsed_vfrdet_0 @ 000001c130784000] VFR:0.008711 (5/569) min: 9 max: 11 avg: 10
Beyond that, my longest test clip has been 3:25 at 24fps with 422LT on V1.01 and did have a VFR of [Parsed_vfrdet_0 @ 00000278ecd121c0] VFR:0.007692 (38/4902) min: 24 max: 26 avg: 24
But the second longest for 24fps at the same 422LT recorded at 44 seconds is CFR [Parsed_vfrdet_0 @ 000001f613663e80] VFR:0.000000 (0/1059)
I hope to have the time to do some more tests that are longer over the upcoming weekend.
Even going back on past recordings I’ve made back to the iPhone 8 using the app ProMovie, while it’s only 4K Rec 709 at 120Mbps H265, there was VFR, but it was always constant. A 2hr and 11 recording would show in media info, 60fps with a min of 59.850 and a max of 60.150, so you can limit how VFR is done. That came back with:
[Parsed_vfrdet_0 @ 00000202698567c0] VFR:0.111731 (52813/419865) min: 399 max: 401 avg: 399