Flowframes Early Access Download: 1.37.1
Added 2022-05-04 15:16:56 +0000 UTC
So initially I wanted to fix deduplication and scene detection for fractional factors but it took a bit longer than expected, but that's not a big deal because I managed to add a bunch of other useful stuff in the meantime.
Custom interpolation are now enabled by default (instead of being a hidden config option), you can now set a target framerate with RIFE-NCNN, Flowframes should prevent Windows from going to sleep, video encoding (esp. AV1) is faster by default, plus some other small fixes.
Full changelog on Discord and in Flowframes.
Under the gear icon settings, under Export Options, just leave "Maximum Output FPS Encoding Mode" and "Make Output Match Output Duration" at default values. So keep Maximum Output FPS Encoding Mode to "Only Create Output with Limited FPS" and keep Make Output Match Input Duration UNchecked. That's what fixed it for me. I haven't gotten the Auto-Encode encountered an error message ever again. I don't have enuff karma to post on Reddit
GamingCanBeFun (YouTuber)
2022-05-14 00:17:52 +0000 UTC
can you please also share it to reddit? I try to avoid discord as I fucking hate it
Kenneth Flournoy
2022-05-13 22:50:53 +0000 UTC
Use RIFE-NCNN
LTSM
2022-05-12 21:58:27 +0000 UTC
I'm running into errors saying the Python runtime can't be found. It suggests I redownload flowframes with the embedded python runtime or install python myself. Python is installed already, though :)
BonEte
2022-05-12 21:56:03 +0000 UTC
I fixed it. Thanx anyway. I put the screenshot of what I did to fix the error in the same Discord channel for future reference to help with any development ;)
GamingCanBeFun (YouTuber)
2022-05-11 01:33:19 +0000 UTC
I DM'ed "flowframes-support" in the Discord channel n uploaded both my ffmpeg log and session log. I'm under the name "GamingCanBeFun"
GamingCanBeFun (YouTuber)
2022-05-11 00:39:07 +0000 UTC
DM me on Discord with the error and logs
LTSM
2022-05-10 23:39:40 +0000 UTC
Ty, but no matter what settings I use with RIFE, it keeps coming back to me saying "Auto-Encode encountered an error" then interpolation cancels automatically... I'm trying to turn a 4K 60FPS video to a 4K 240fps video. It worked in the free version on RIFE 3.1 just fine with HEVC, H.265 and H.264... So I know my hardware supports that interpolation frame rate. I'm gonna have to go back to the free version until this is fixed. I haven't tried it in ur older paid ones yet tho
GamingCanBeFun (YouTuber)
2022-05-10 23:32:22 +0000 UTC
Cool, thanks for the info, I hate relying on assumptions as I never make decisions with them in mind!
Kezo
2022-05-07 17:49:40 +0000 UTC
Yeah pretty much, it's the sensitivity of the detection
LTSM
2022-05-07 15:48:57 +0000 UTC
Thanks for the update, I love the prog, I just need to know about the dd settings, does the 2.0 setting represent the amount by which the prog finds dupes?
As in, the higher the number, the higher the percentage of dupes found, vs lower numbers that represent lower percentages of dupes?
Kezo
2022-05-06 13:12:34 +0000 UTC
It does not work yet, so 24>60 should only be used for content without scene cuts or duplicates, for now
LTSM
2022-05-05 11:26:30 +0000 UTC
Hi, thank you for this amazing program and work! I had trouble understanding your sentence: "so initially I wanted to fix deduplication and scene detection for fractional factors" <-- does that mean you did actually fix it, or, that's still something presently being worked on and probably fixed in the next version? I'm one of those weirdos doing 24>60fps, lol. TIA
Pewburrito
2022-05-05 10:37:25 +0000 UTC
Not at all, CPU encoding is always better, it just takes longer
LTSM
2022-05-04 20:38:02 +0000 UTC
Last question, is H265 NVENC the optimal encoding method speed/quality wise, assuming you have a high end computer.
Jake
2022-05-04 20:30:04 +0000 UTC
Deduplication should work fine fine for non-fractional factors but I recently encountered a bug where it detected dupes on a video that did not have any. But if it does have dupes it should work normally.
LTSM
2022-05-04 20:04:51 +0000 UTC
I was using dedpulication recently, so it's fine for normal use right? What was being fixed specifically if you don't mind me asking?
Jake
2022-05-04 20:03:51 +0000 UTC
4.0 is my go-to, it's fast and quality seems to be good if not the best.
LTSM
2022-05-04 15:19:54 +0000 UTC
Is there a comparison between the different models to find out which one is best for which use case? Do you have a go-to model you mostly get the best results from?
Love your work btw! <3
mynamejeremy
2022-05-04 15:19:00 +0000 UTC