I would find useful a way to make easier to take previews a single frames of the video, save them and compare to others closely inside the tool, instant swap one exact frame to the same frame, and being able to do it with zoom, and being able to identify the ai model and parameters used on each one.
This is very useful when someone is trying to discover which ai model better apply for the video.
But I found more useful any way to export to video lossless or pseudo-lossless, so it could be optimized for future modifications in other programs without adding noise on each lossy encode.
Hello, can anyone help me? Iām running this version with these settings:
TheiaFidelity
Sharpen 60
Detail 30
Noise 42
200% Scale
MP4 + attempt to sync audio on
and all 4 of my video export with 8 seconds of black at the beginning. The audio plays just fine under the black and each, after 8 seconds, displays the video again.
I switched off keep audio and merged after, and itās still out of sync. Video starts fine and synced, but begins to lag about 60% through a 30 minute video. This is with Atremis-HQ. Iāll try another model, but waiting almost 5 hours to experiment is a big waste, especially at 1 AM, like it is now. Donāt know if this is the same problem as @danielpando?
Iām noticing some grid looking thing with the Theia-Detail model (the setting Iām using is Sharpen 60, Details 100, Noise Reduction 5, the settings are not relevant as it happens on other values too).
Take a look at this picture and pay attention to the green background at the center, there are grid looking thing thatās apparent where color gradients are:
It was constant, but after applying decomb filter, it changed to variable. I hope thatās it, because the latest Atremis-HQ model is phenomenal for my old Gilliganās Island episodes. So good that I decided to redo season 1 and 2 completely over again.
UPDATE: Still looses sync about 60% through.
Original video:
Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L3
Format settings : CABAC / 3 Ref Frames
Format settings, CABAC : Yes
Format settings, Reference frames : 3 frames
Format settings, GOP : M=3, N=24
Codec ID : V_MPEG4/ISO/AVC
Duration : 25 min 35 s
Bit rate : 23.0 Mb/s
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 4:3 Frame rate mode : Constant Frame rate : 23.976 (24000/1001) FPS
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 2.773
Stream size : 4.11 GiB (97%)
Default : Yes
Forced : No
Color range : Limited
Color primaries : BT.601 NTSC
Transfer characteristics : BT.709
Matrix coefficients : BT.601
Has anyone noticed the drag and drop is broken in this build?
Minor annoyance but I tend to use it more than loading a batch as I have had the horizontal scroll bar not appear if you initially load more than 10-12 files all at once for batching.
Drag and Drop works as usual for me on this version. However I did notice it not working if you select ālaunch video enhance AIā right after the installation, it requires closing the application and relaunch to get drag and drop working normally.
BTWā¦ I have tested to import a video with lagarith-codec and I think VEAI donāt like it.
Sometimes the video jumps between the images ā¦ I think the codec is in general a bit buggy ā¦
Iām working with Huffyuv and FFV1 with no problems