dawkinscm wrote:
Xenocyde wrote:
dawkinscm wrote:

What happened with this? Are you using it for Rife? Have you tried Ensemble?

I got the new PC but have no idea how to setup the deblur model. I reached out to UHD via PM, no answer since May... maybe you can assist me?

Not something I even knew about to be honest. I have a similar setup to you and I was only interested in Ensemble which I have now tested and confirmed that it did make existing Rife artefacts worse.

Ah I see. I haven't tested Ensemble cuz I haven't yet got a 4K screen (minght get a QD-OLED next year, since the 2023 Samsung ones still have some problems). Waiting for a new Rife version, hope it comes out soon. Also, hope UHD sees this comment and reaches out with some instructions to test the deblur model big_smile

dawkinscm wrote:
Xenocyde wrote:
UHD wrote:

If anyone has run this model and shared the results of its performance in scenes with dynamic action I would be very grateful.

I'm getting a new desktop setup with an i7-13700K and RTX 4080 next week, so I can test by then.

What happened with this? Are you using it for Rife? Have you tried Ensemble?

I got the new PC but have no idea how to setup the deblur model. I reached out to UHD via PM, no answer since May... maybe you can assist me?

UHD wrote:
Xenocyde wrote:
UHD wrote:

If anyone has run this model and shared the results of its performance in scenes with dynamic action I would be very grateful.

I'm getting a new desktop setup with an i7-13700K and RTX 4080 next week, so I can test by then.

Super!

Hopefully also the fastest RAM - will come in handy with RIFE in 4K

It's Corsair Vengeance DDR5-5600 but it can be OC-ed to 6000 or even more, I'll see the exact limits.

UHD wrote:

If anyone has run this model and shared the results of its performance in scenes with dynamic action I would be very grateful.

I'm getting a new desktop setup with an i7-13700K and RTX 4080 next week, so I can test by then.

Wow... just wow... looks like it was well worth waiting for UHD's latest developments. I'm all for this new deblurring model, but how can we help exactly? What coding skills are needed? Maybe I can ask some coders I know, but what exactly do I tell them?

Xenocyde wrote:
FenceMan wrote:

Is there anyway to get rid of this thing that pops up everytime I start to play something with RIFE?  Once I close the window everything plays fine.

https://i.postimg.cc/htZWTCB2/Screenshot-2023-05-17-065328.png

It cannot be disabled if you use performance mode, but it only pops up when a certain resolution is not already stored in the timing cache. You need to let the timing build for ~3 minutes every time, don't close the window before it closes automatically. There are only a few timings for movies, but you may encounter specific ones for certain TV series.

zSoi wrote:

Some newbie questions here :

Owner of a rtx3080 12600k, I like the smoothness SVP can get me, but sometimes artifacts are annoying.

What can I expect from this rife ai engine ? is it worth the time to set it up ? I'm watching movies on an ultrawide HDR display 3440x1440, usually x5 frames per sec. (and I don't know if its revelant but my movies are 4k and about 30/50 Gb per movie.

From what I read, RTX3080 is not enough for 4K HDR with RIFE. Even the RTX4090 struggles at more than X3. Also, RIFE V2 seems to be bugged for 4K vids as it lowers the resolution for some reason. But there supposedly is something better than RIFE, we're waiting for UHD to unveil it roll

UHD wrote:
Xenocyde wrote:
UHD wrote:

If anyone is interested in much, much better quality than RIFE interpolation I will write about it soon, here on this forum. Please give me another week, maybe two weeks, until I update my repository on GitHub.

2 weeks are up, UHD! Time to spill the beans, don't keep us boiling here roll

I'll be back with information by the end of April for sure.

I understand you are most likely busy, but May is here.

UHD wrote:

If anyone is interested in much, much better quality than RIFE interpolation I will write about it soon, here on this forum. Please give me another week, maybe two weeks, until I update my repository on GitHub.

2 weeks are up, UHD! Time to spill the beans, don't keep us boiling here roll

Blackfyre wrote:
Sopheus wrote:

Not sure how everyone is able to play on x3 in 4K

Here's your issue.

2.5X should work too I think.

Tested MPC-HC with Avatar 2. 3D load was hitting 80% and after around 1h it jumped to 100% for about 5 seconds. MPV has similar load, max 85%. After that, I decided to test once again with HAGS turned off. 3D load sits at 15% even in MPV! Haven't noticed spikes for the remainder of the movie (around 2 more hours). Can't remember the exact improvements with HAGS off for V1 ... maybe 3D load was 20% lower because I was still getting spikes to 100% although less frequent. V2 is waaay better in this regard.

dawkinscm wrote:

I moved away from MPC (with MadVR) because it was usable with Rife. With MPV I get better performance, almost perfect smoothness, less artefacts and much better interpolated picture quality than with MPC/MadVR so I can never go back. With a 3070 laptop GPU you are always going to have issues but maybe with MPC it will work better for you.

I installed MPC-HC and configured it to run through VapourSynth. Have you tested with this setup? I'm testing now, will report any improvements on the 3D load later.

L.E: Tried a 40 minute episode. Everything seems the same in MPC-HC quality-wise, except the 3D load is 10% lower on average. MPV hits 70%, while MPC-HC barely hits 60%. This could be good news, but need to test further with a long movie, as the frame drops rarely occur in 40 minute episodes with MPV as well.

I switched last night to V2 as per the tutorial posted above. I can't really notice much of a difference in smoothness on my RTX 3070 laptop GPU for 1080p content. Artifact handling seems a bit better maybe.

However, I still get the framerate drops that last around 10 seconds. These usually happen after 1 h or 20-30 minutes during 2h+ movies, very rarely in TV series episodes of ~40 minutes. The SVP index drops from 1 to 0.8-0.7 usually, sometimes lower, and the GPU 3D utilization in task manager jumps from 70-80% to 100% while RAM utilization sits sub 3 GB most of the times. This is happening in MPV by the way. I need to test with MPC-HC.

UHD wrote:

If anyone is interested in much, much better quality than RIFE interpolation I will write about it soon, here on this forum. Please give me another week, maybe two weeks, until I update my repository on GitHub.

Oh? Patiently waiting for your post now...

Criptaike wrote:

Hello, so I've wanted to try the rife engine, the vulkan version works fine, but I sure want the tensorRT version, so when selecting rife after opening a movie, a cmd window pops up and mpc-hc crashes, it just goes to shit and it doesn't really say that there's an error or something, try to search the problem on google and this forum and no luck. Here is a gif of all that happening -> https://imgur.com/a/VHVCvBZ

Any ideas on what's happening? The vulkan one works.

Thanks

Need to wait for 3-4 minutes every time that cmd window pops up.

Mardon85 wrote:

My SVP settings are as follows:

https://i.imgur.com/SBPaBcz.png

What is this user defined option with Tensor RT?

dlr5668 wrote:
dawkinscm wrote:

I've been trying for a while now to get RIFE running smoothly with my 3080 with MPC-BE. I've given up on that so instead I tried using MPV and there was an instant improvement. It still needed a bit of work and I had to learn about MPV config files so that I could configure it to match the scaling used in my MadVR config. Now it looks great and runs very smoothly.

Nvidia should release soon https://www.youtube.com/watch?v=XA-tQpQqD7U Perfect match for Rife

Video Super Resolution support introduced today with the 531.18 drivers.

Xenocyde wrote:

I've been noticing some drops in performance here and there, not sure what triggers this. The SVP index drops to half for several seconds and then it goes back up to full. Could this be a VRAM problem? I have an RTX 3070 mobile with 8 GB, performance boost enabled, but I did not deactivate GPU scheduling.

I disabled GPU scheduling, but still get the occasional drops. Btw, the index drops to ~0.8-0.9, not 0.5 as stated previously, yet it feels stuttery as hell. Also tried with 2 GPU threads instead of 4. Seemed like this solved the problem but I still got a stutter at some point... maybe it reduces the frequency of occurrence.

GPU scheduling did allow me to run SVP RIFE with lower wattage though. This means the GPU cooler is not in full load in overboost mode anymore so less noise. I can even run at full index on balanced mode (lowest power preset on my laptop) for 40-50 minute episodes, but the GPU starts to get too hot over 1+ h and failsafe maximum fan speeds kick in, so this does not really work for longer movies.

The MPV-lazy version lets me adjust these parameters if stutters start to happen:
if interpMulti == 3: maxIpps = 1.0 *
elif interpMulti == 6: maxIpps = 1.0 *

Is there something similar in the SVP implementation?

I've been noticing some drops in performance here and there, not sure what triggers this. The SVP index drops to half for several seconds and then it goes back up to full. Could this be a VRAM problem? I have an RTX 3070 mobile with 8 GB, performance boost enabled, but I did not deactivate GPU scheduling.

Piovasia wrote:

Could someone share me the generate.js and base.py file, please ?  (╯•﹏•╰)

Go to this page and find Chainik's post with red text warnings (towards the top of the page). It has those 2 files.

cemaydnlar wrote:

On tensor rife 4.6 %15 scene change threshold i get lag on fast panning scenes just like on vulkan 4.4 with %10 scene change threshold. My gpu usage is around %50-55. Is there a solution for this ?

We basically need this

constasmile wrote:

Chainik, please try implement new ai scene detection from https://github.com/styler00dollar/VSGAN … detection. I use this fragment from Mad Max for tests how scene detection algorithms working https://drive.google.com/file/d/1f7zZ_A … share_link and this new thing performed almost perfectly.

Maybe Chainik can implement it sooner rather than later.

I reinstalled SVP to test the latest Rife TRT stuff. First thing I notice is that there is only model 4.4. When I play a video in 1080p I get 0.57 out of 1 on the SVP index graph. What am I missing here? I'm back in the exact same spot I was before installing the test pack released on Jan 9. That pack was working fine.

LE: Ah my bad, had to switch to TRT from the first menu.

LLE: What does that dynamic shapes option do? If I leave it on, I see stuttering, even though the SPV index is at 1.

Chainik wrote:

SVP updated, added TensorRT (no need for doing anything manually), fixed (??) possible audio desync after RIFE transcoding (both Vulkan and TRT versions)

Either select the checkbox at install time, or install an additional package called "TensorRT" via SVP menu -> Utilities -> Additional programs and features.

----
It uses "dynamic shapes" by default, i.e. only ONE optimization pass for all resolutions below 2560*1440. And if/when you open a 4K video then it'll do another pass for resolutions from 1440p to 4K.
BUT! this can be a little bit slower than optimization for a specific resolution, like 5% in my case. This is why there's a "performance boost" profile option - it forces its own opt. pass for each resolution.

another finding: model 4.6 needs more than 8G of device memory for 4K res, so if you're on a 8G card and want 4K then choose model 4.4.

----
Also installation scripts fixed so the Vapoursynth Filter must be registered automatically.

Do we need this update if we already have the test pack you posted on Jan 9?

Chainik wrote:

"workspace" is deprecated, should be set to "None"

Ya but what file do I need to edit?

dlr5668 wrote:

edit this line

clip = RIFE(clip=clip, multi=interpMulti, model=RIFEModel.v4_6, backend=Backend.TRT(fp16=True, device_id=GPU, workspace=WS_size, use_cuda_graph=True, num_streams=GPU_t))

I think 2000 is enough for 4k

What file do I edit?