I used mp4 twice and I guess I got lucky. Once I saw it had lossless image output options, I’ve been using png instead. Png gives you some nice extra options- changing framerate is simple and I can do editing of bad frames with Photoshop.
With the new Nvidia drivers 457.51, VEAI is constantly crashing at launch now. It doesn’t just crash and close VEAI, it causes my screen to blank while the driver resets itself.
Might be your set-up. I use the latest studio driver and it is very stable, no crashes.
@virtutis in your previous post you said that you started working on NTSC sources and I’d like to ask you a question:
are you trying adding the PAL audio to the NTSC video?
Because I tried but many episodes have got more frames in the NTSC version than in the PAL, resulting into the impossibility to align PAL audio with NTSC video, did you manage to solve that problem?
Thanks!
Framerate option is missing when choosing an image sequance input and MP4 output.
1 Like
You have to make what is called a slowdown from PAL to NTSC
you can use an avisynth script to do this
LWLibavAudioSource(“youraudiofile”)
ConvertAudioToFloat(last)
SSRC(Round((AudioRate()*1001.0)/960.0)).AssumeSampleRate(AudioRate())
return last
And if you want to correct the pitch, replace SSRC with
TimeStretch(tempo=(96.0/1.001))
But timestretch is more relevant on NTSC to PAL than PAL to NTSC
or with ffmpeg I think this is
ffmpeg -i videofile -filter:a “atempo=0.959040959040959” -vn outputvideofile
@jphilip-649517 thank you for your answer! I start with a variabile framerate NTSC source: I convert It to 23.976fps CFR and then upscale exporting as image sequence. Then I convert those images to video at 25fps and finally I try to merge PAL audio with the video from NTSC upscaled frames, so I am matching a PAL audio with an NTSC video with increased Speed to 25 fps.
The issue Is that NTSC source has got a couple of frames( 200ca) more than the PAL source, so the audio goes out of Sync for those missing frames in the NTSC.
I already tried keeping 23.976 fps and slow down audio but It remains out of sync, with an audio that goes forward and backward during the video.
I’d like to remove those 200ca frames from the NTSC source but it’s pretty impossible to fine them with precision.
Any idea on that point?
Thank you!
That’s an interesting method I’ve not seen before. How fast is it for pitch correction?
How is the pitch correction quality compared to doing it in Adobe Audition with the high quality settings? Audition lets your enter an exact audio length which I find more intuitive and more useful when fractions fail to create the exact length needed (Audition lets you do either). And they usually fail, and you have to redo the audio with a correction, like + or - 100ms. Pitch correction is a 3 to 4 hour process for a 90 minute movie on a quad core CPU.
Right… I’ve had that before. It means your audio and video lengths are not exactly the same. And can be very difficult to fix. That’s the issue I have for that 90s TV series I’m working on. And the reason I’m creating a Plural Eyes that handles the issue better. ETA is at least 6 months before I have an alpha version working. If I had money, I’d start a software company with better developers as it’s a definite money maker (IMO). Finding investors is not something I’m good at unfortunately.
I was hoping on a quicker solution but I imagine that’s not that easy to fix that kind of issue. I’m trying to work on PAL sources only for those episode but well, those sources are really bad compared to NTSC ones. When your PluralEyes will be ready I’ll keep it in consideration if I can’t find a solution in the meantime.
@jphilip-649517 have got any idea on that?
Thanks to both of you!
If you convert vfr to cfr that’s what usually happens duplicate or worse dropped frames occurs. What I would do to try to sync the audio is use an NLE (Premiere Pro for example) ,
- create a sequence/timeline at 25 fps
- then add each video on seperate tracks
- reduce the size at around 25% of one the tracks and move it somewhere on a corner (PIP)
- Then start to move the timeline slider or just playback the video and when you start to see mismatch frames stop and and go back to find exactly where the mismatch happens on the NTSC track then delete frames until it rematch the PAL video
- and so goes on and yes this will be a very tedious and long work if there are 200 frames to remove
- once finish re-export your new NTSC video that matches your PAL audio.
@videogeek I don’t have Audition so I can’t tell. But it’s very fast on avisynth. And in general slow methods aren’t always the best, I tried once with Audacity high quality pitch correction that takes forever compared to the standard mode and the result was far more worst than the fast standard.
@jphilip-649517 Interesting method, which seems similar to me to the one of videogeek, I tried to do that for almost two hours on an episode but I couldn’t cut frames precisely and the end result was a messy video, with audio more out of sync and jumps between one to another frame were I cut. I’m sure your method can work but I’m not good at it and doing it on every episode would require too much time and work, I think I’ll keep working on PAL sources only for those episode if there isn’t a simplier solution.
If you’ve got any other idea that you want to share I really appreciate!
Thank you for your help!
Still looking for a comparison of performance between RX 6800 XT and RTX 3080 on Gaia CG so I know which card to buy. Looks like they are pretty comparable on the gaming side so this application will likely be the deciding factor.
I’m getting a wierd half-crash using Artemis V8 HQ. From DVD to 1080p using 1080Ti
It will process for hours then just stops, no movement on screen, no HDD activity and GPU fans reduced to normal. You can move mouse and some other apps will open, some won’t. Only way to get back is to reset PC.
It works using older v7 model but quality not as good as v8 - i did two before which worked but now just half-crashes. No idea why, I’m on Ryzen 3700X PC which is 6 months old.
Any ideas?
Exactly how is black on white or white on black “lack of contrast”? It sounds like you’re blaming something on Windows that is the fault of your own tweaks.
You aren’t going to get a definitive answer on something like that.
I don’t believe the RX 6800 XT is shipping yet. It’s moot though. Support for AMD wasn’t added until 1.7.0. The Gaia-CG in 1.6.1 is better quality than in 1.7.1. So much better that I’m staying on 1.6.1 for now. I’m also testing 1.2.0 right now, as it has the Artemis-MQ model, which according to the posts on this forum, sometimes does higher quality (less noise) with film DVD sources than Gaia-CG 1.6.1.
I have a film from DVD that still has too much noise despite pre and post processing with QTGMC interlaced and progressive modes, so last night I figured out how to get 1.2.0 and 1.6.1 to coexist with their own registry settings (no details as the process is alpha and I don’t know yet if it’s stable).
1.7.1 also has blocking artifact issues with Artemis v8. Read through this thread and you’ll find examples.
Regarding your crashes, do you have at least an 850watt power supply? Crashes with high power GPUs and CPUs are sometimes caused by insufficient power.
I’m not sure what you are refering to. I didn’t make any post regarding “lack of contrast”.
What is this studio driver I hear people talk about? Can you get this for 1080Ti?
@paulgates I’m not the best at explaining those things but here’s my answer.
Nvidia GPUs have got two kinds of driver:
Game ready driver: they are optimized for videogames and for a traditional use of the GPU, they’re the most commonly used and they focus on the gpu’s output quality sent to the monitor, considering parameters like fps and other ones.
Studio driver: they are optimized for use in video editing software that required more computational speed instead of a perfect video output sent to the monitor.
About that I’m not sure but as far as I know Studio Drivers are better for use in software like this one because they use all the GPU power for computation instead focusing on things such as in game fps or other things like that. Aldo you will have to update with less frequency because studio drivers include only stable and tested updates.
Your 1080ti should have those drivers available, you should check on the Nvidia website.
I hope this will help!
