dawkinscm wrote:
aloola wrote:

v15.2 gives me a tiny FPS boost
v14test3 ~150fps vs ~ v15.2 ~158fps, also fixed dynamic shape (Performance boost: off) bug

15.2 changelog:

    Upgraded to TensorRT 10.3.0.

    Fixed performance regression of RIFE and SAFA models starting with vs-mlrt v14.test4. This version may still be slightly slower than vs-mlrt v14.test3 under some conditions, however.

Thanks. I have been waiting for this. On brief testing I don't think I see anything obviously better than 9.2. .

I check for this release every day, you always beat me bringing the news first!
Downloading right now and happy beginning to test, by the way if you willing to test I install a Nigthly build of MPC HC that has a beta version of MPC Video Renderer 0.8.6.2274 (git-2024.08.04-d252d91) x64 i dont what if any change but I like more than the one install on svp list.

dlr5668 wrote:
Insindro wrote:

What are your top 3-5 rife models and miscellaneous settings for scene changes? Just want to get a good idea of what everyone is currently using, thanks!

I like 4.15. If you need 30% faster its 4.15 light. 4.20 is nice for anime. Try these models. I only use v2 for faster seek

These days I prefer 4.18 Nvof MV/blend Adjacent Frame x3, 4k files  downscale 1920*1080 upscale 1440p(Rtx 4070ti)

RickyAstle98 wrote:
scb wrote:
RickyAstle98 wrote:

Line 137
Line 1179
Line 1297

Please do elaborate.. what is the impact of this? What are my options and how does it help? smile

Higher optimization level the longer engine build times but more accurate dynamic tactics!
Level 0 : fastest engine build/performance drop (about -2%)
Level 1 : faster engine build/performance drop (about -1%)
Level 2 : fast engine build
Level 3 : default TRT execution level
Level 4 : DONT but DONT do that plz!
Level 5 : DONT but DONT do that plz!!! (+3% performance)

Very old post, I come back to this and I wonder why you put it like that, only for the longer time to take the builder to finish?  or there is something else what no use a value of 5

Chainik wrote:

> Is there anything we can do to increase NVOF and SVP algorithms strength

yep
smooth.scene.limits.scene
smooth.scene.limits.blocks

the bad thing is that values that work good with resizing to 720p doesn't seems to work with 1080p and 4K

Good to know, only for reference which value it's similar to IC with a value 25% 30% of scene threshold.

dawkinscm wrote:
Chainik wrote:

it's clearly a scene change and SVP do everything right - see f1.jpg before SC and f2.jpg after SC

> what you call "garbage" might simply the AI doing it's job1

see attached rife.jpg
for the 3rd time - if you OK with this - then OK, "SC = disabled" is right for you
and I'll continue thinking that proper SC detection is a must for every frame interpolation algorithm

So I just checked I do get the same before and after frames as shown in f1.jpg and f2.jpg, but for IC.  When I use either SVP or NVOF SC, I get the same issue as shown in Rife.JPG. It would be interesting to see what others are seeing here because I checked and rechecked using SVP SC to be sure,  so something strange is happening here. This might explain why I didn't like the look of what I see when using SVP SC.

For me the frame between F1 and F2 are much better that the rife.jpg and all methods look almost the same.

Blackfyre wrote:

SVP Motion Vectors in those specific time stamps makes the creature ribbons and America Chavez look as though they are frames (even though there are no dropped frames at all), similar to Image Comparison when its set to very low percentages like 6% for example..

Thats exactly what I was trying to say, maybe this helpChainik . for tune up.
The test for the 4.18 it's going well, dont test performance values, only viewing experience. Someone knows which changes /modifications it's in these model?

dawkinscm wrote:

There's also scene in Dr Strange 2 where using SVP to reduce screen size actually helps to remove a movement artefact. The scene when he throws his cloak towards the monster to rescue the girl, there's always a little judder and some artefacts when she returns. But with Decrease Screen size, it's gone.


Same with the cloak and the bike for me that scene its horrible mismatching.

Blackfyre wrote:
Drakko01 wrote:

I pointed out one example of this in Doctor Strange Multiverse of madness, after defender Strange is killed by the Ribboned Creature and America chavez run underneath.

Can you provide the timestamp? I have that in high quality and I can test tomorrow.

3.06 and 3.16 the ribbons of the creature , maybe I misinterpreted and its something else.Thanks for taking the time.

Chainik wrote:

Drakko01

>I still think that image comparison work the best

to be "the best" it must give less false positive scene changes, and less missed scene changes
which is not true

What I try to say it's that i see fewer errors in the reproduction. I pointed out one example of this in Doctor Strange Multiverse of madness, after defender Strange is killed by the Ribboned Creature and America chavez run underneath. before of  update both Svp motion and NVOF give similar errors, and IC dont. After the latest update only SVP motion failed that particular scene.

As dawkinscm say I cant only share my what i see. And I assure you that I am not against svp, from the beginning I preach to anyone who talks to me about its use. I just want to see it improve day by day, since I use it always.

I will like to ear others members results after this update.I still think that image comparison work the best , but i kinda like in some scenes NVOF motion vectors, the improve its subtle but its there.

I totally agree with Dawkinscm, of all the SCD methods, for me image comparison is the one that I think gives the best overall viewing experience. By the way, it's worth mentioning that we have this  new option  not only thanks to the work of Chainik and others. But also because the push and desire of members like Dawkinscm for bringing SVP/Rife to the ultimate potential. Thanks for that!!

dawkinscm wrote:
Chainik wrote:

> please write down here, for laymen to understand, what's the best opion to use
the default one! (which is "SVP motion vectors")

The default one works well with everything except the intro to Hugo where it stutters considerably.

I found other cases when svp motion vectors using blend struggles with a complex scene, for example in Doctor Strange Multiverse of madness, after defender Strange is killed by the Ribboned Creature.

abraxas wrote:

RTX 4070 here. Scaling of 4k content down to -23041260, frames x2 (e.g. mostly 48fps), then upscaling to 4k with MPV, additional filters like adaptive sharpening. Very smooth, very nice, GPU usage max. 80-85%.

In two weeks I am thinking of upgrading to a 4070 ti super, if I do I will share my testings

RAGEdemon wrote:

I don't understand you folks downscaling 4K content.

With RIFE ~4.4/4.6, you can easily get 2X - this will up 24fps to 48fps, but YOU MUST CREATE AND USE A CUSTOM 48Hz @4K resolution! It is buttery smooth.

Higher than 2X gives diminishing returns, and really isn't worth the quality loss from downscaling the image to get there.

First, you assume the screen everyone has can accept 4k@48hz, some of us use a TV.
Second, the loss of quality is overrated, not all of us think the same when it comes to saying that it looks better, too many variables and hardware.

Xenocyde wrote:

Question is which renderer is better with RIFE for mitigating microstutters: VapourSynth or MadVR?

That's where you're wrong, Madvr it's a render, VapourSynth it's a filter, you can use both in combination with Mpchc or MpcBe

Xenocyde wrote:
Drakko01 wrote:

Why dont you try MPC HC+Madvr+Rife for the content and framerate you want , your gpu has room to spare.

Isn't VapourSynth better for RIFE? I'm just looking to get rid of the microstuttering.

I don't understand your question I use Vapoursynth  with mpchc and Rife without the problem.

Xenocyde wrote:
Drakko01 wrote:

Do you unselect avisynth , the other thing its under utilities, set environmet variables for VapourSynth

I removed avisynth altogether, only vapoursynth is enabled. But I think I screwed something up now. After reading your instructions I thought it would be a good idea to uninstall and reinstall vapoursynth via SVP. Then I did SVP menu -> Utilities -> Set environment variables for Vapoursynth and now MPC-BE triggers SPV RIFE but I only have generic 4.4 as model and the TRT drop-down is gone. Halp!

L.E.: Looks like it accidentally uninstalled MPV and TRT when uninstalling Vapoursynth. I installed TRT and MPV again but now MPV would not trigger SVP and MPC-BE displays a warning in attached image over the video image. Think I need to install new TRT.

L.L.E.: OK think I fixed it. Overwritten the vsmlrt-cuda folder and vstrt.dll with the 9.2 version and both MPC-BE and MPV trigger RIFE now. Will test MPC-BE for microstuttering.

Why dont you try MPC HC+Madvr+Rife for the content and framerate you want , your gpu has room to spare.

The Mpc hc Release a New version 2.3 for who want to test

Updated LAV Filters to version 0.79.2-9-g56aab
Updated MediaInfo to version 24.05
Updated MPC Video Renderer to version 0.8.3.2255
Updated libass and various other external libs

Xenocyde wrote:
flowreen91 wrote:
Xenocyde wrote:

how do I force MPC-BE to always start with RIFE?

Go to Options -> External Filters -> move VapourSynth Filter to top and set it as Prefer

I followed the installation steps from here, which shows Avisynth. If I select VaporSynth, the filter does not show up in the play-filter list and it does not trigger SVP at all. I remember something about registering it... I haven't used MPC in almost a year.

Do you unselect avisynth , the other thing its under utilities, set environmet variables for VapourSynth

flowreen91 wrote:
Xenocyde wrote:

how do I force MPC-BE to always start with RIFE?

Hmm maybe Drakko can explain here what to do.

Play a file, automatic its  trigger , go to svp tray icon , Left click on it Select Rife engine, done! next time you play a file Rife its a default, except if you have custom condition set up.

Now for the Sync to display Option , when Aloola sugested to me I change in the frame Syncronization option of MPC BE under EVR Sync Settings.

Xenocyde wrote:
Drakko01 wrote:

Okay now I see, let's take another approach then,, remember me if you use MPC HC or MPV. Do you use in combination with Madvr. Later when I back home I share screenshot of my settings. For most useful information if you use Madvr post you control J stats. and you rendering General settings.


I use MPV so no MadVR from what I know.

Ok, I not much a user of Mpv, I try  MPC-VR in combination of Mpc-BE and had a similar issue with microstutters and Aloola share this tip that was a improvement.

Drakko01 wrote:
aloola wrote:
Drakko01 wrote:

If I use mpc-VR (RTX HDR + RTX SuperRes) with MPC-HC or MPC-BE I don't get the best experience, the quality is very good, but I have a kind of stutter in slow camera movements or certain close ups,  no in panning movements. This is not present when I use Madvr. Maybe i got some settings wrong in MpcVR dunno.
BTW I always talking of Realtime Playback just for clarify.

I think the lack of G-sync/freesync support causing this.
try the options video -> frame synch -> "Sync video to display" in MPC, it may fix the problem

.

Xenocyde wrote:

The issue is not related to a specific scene. The microstuttering happens and when I rewind to check if maybe that was an encoding frame skip, it's not there anymore. I use a 14-year old 37-inch 1080p Samsung TV (yes I know, need to get a new one, been waiting for a good OLED, but just can't pull the trigger). Drivers are the latest 555.99 from Nvidia.

Okay now I see, let's take another approach then,, remember me if you use MPC HC or MPV. Do you use in combination with Madvr. Later when I back home I share screenshot of my settings. For most useful information if you use Madvr post you control J stats. and you rendering General settings.

Xenocyde wrote:

L.E.: Tried 4.17, I can still see microstuttering. Will try V2 as well.

I work better when I see it, can you give me an example video/scene that gives you microstutters?

Also, what display do you use?

Drivers ?

Xenocyde wrote:

There was a time when microstutters did not occur at all, but who even knows what version of RIFE or what drivers I was using back then.


I don't know if this will help you or not, the same thing happened to me, smooth but some scenes have micro stutters, noticeable on faces when the movement is slow, always talking in the same scene, the combination of Rife 4.17, TRT 9.2 and the latest drivers. from nvidia 555.99 fixed for me , at least for now.
Maybe it's a stupid question, but in the nvidia control panel you have the vertical sync option enabled, if not, do it.

Someone can help on how test Vapoursynth R68. Thanks in advance.