I have a Ryzen 9 3900X with 32Gb DDR4 of Ram and a Nvidia Rtx 4070 Super GPU.
I try to make in a 1 minute and 30 seconds video (1920x1080 @ 60 fps) a 4x slow motion using Chronos Fast with the default setting for this model.I only change the slow motion to 4x
Topaz 5.5.0 tells me that he will made at 18-19 fps with a duration about 21 mins. My questions are:
This is a normal time using my actual setup ?
If I change the GPU to a 4080 Super what will be the increased in fps?
If I change the CPU to a Ryzen 9 7900x and the GPU to 4080 Super what will be the increased in fps?
-If I change cpu and gpu and change also the Ram to 64Mb DDR5 what will be the increased
I love the final result of the slow motions but I don’t have so much time to wait for the process to finish
I appreciate your feedback because before spending my money on improving my setup I needed to know if the improvement in time spent is substantial because if it isn’t then it’s not worth improving my setup
close to 20 fps is quite good. when I process full hd video with the 4090 it often won’t be much faster. Currently processing to 1920 x 800 with a processing speed of 21 fps. this is not for slow motion though. Iris medium it is.
It’s tough to say exactly, but all the benchmark speeds with Ryzen 3000 series are significantly slower than even 5000 series with the same GPU. From that I would say get the 7900x before getting a 4080 Super.
Thank you for all your answers. The big question for me and I don’t know if this calculation is possible, is know how much time I will gain if I do the same 4X slow motion with the same video with a better GPU (4080 Super) and a better CPU (maybe Ryzen 9 7900x) and more ram (64Gb DDr5).If I will gain 1 minute or 2 is not worth the cost of upgrade, but if this setup makes the same slow motion in 4 or 5 minutes there is a reason for thinking in spend a lot of money with the upgrade.The slow motion is just an example to compare the time that my actual setup needs to complete the task with a better one
One of my computers has a 7900X and a 4070 ti. I can tell you what speed it runs at for some models and resolutions you ask. I’m not sure how much that will help, but if Imo can give answers with that 4090, we can be sure at 4080 Super will fall somewhere between.
More RAM is not going to make much of a difference. Faster RAM does make a slight difference.
I tested it on a 1 min 30 sec long 1920 * 1080 @ 60 fps video via Chronos fast slow motion 4x with the RTX 4090. The processing speed was aprox. 35 fps. The processing finished in 10 min. + 46 sec. CPU i9 13900K limited to 5.5 Ghz, RAM 32 GB DDR5 @ 6 Ghz. Topaz Video AI 5.5.1.
Looking at the benchmark thread. GPU seems to have little impact on the interpolation models. From what i can see, ram speed+latency and cpu seems to be what makes any difference. I have a rtx 3080, ryzen 7 7800x3d and 2x16gb of ram, with manually set ram timings. I bench mush slower on the AI enchancement/upscale models that people with a 4090. But i bench faster looking at all the interpolation models.
Did a few tests. Input is a 1920x1080p at 60fps video - only having it create 4x slowmo. Chronos fast sits around 24fps, apollo fast around 37fps and apollo around 21fps.
Once again thank you all for your help.I see today the benchmark thread and I try to calculate the time with the base in fps that are shown in the benchmarks and I think that I’am doing some error in calculation.
For example in my example , I have a 1:30 video to work out 1:30=90 seconds, because i have a frame rate of 60 fps I have a total of 5400 frames. Topaz tells me that he is processing @ 18 fps, so if my calculation are correct I need 5400/18 = 300 seconds → 5 minutes.,but the real time was 21 minutes.Where I fail my calculations ?
