Hello, it seems a nice improovement in 4k4x apollo.
2 Likes
here is my 1080p benchmark
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: 13th Gen Intel(R) Core(TM) i9-13900K 63.715 GB
GPU: NVIDIA GeForce RTX 4090 23.59 GB
GPU: NVIDIA GeForce RTX 4090 23.59 GB
Processing Settings: device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 37.39 fps 2X: 21.76 fps 4X: 6.00 fps
Proteus 1X: 33.37 fps 2X: 19.47 fps 4X: 6.47 fps
Gaia 1X: 15.34 fps 2X: 10.57 fps 4X: 6.53 fps
4X Slowmo Apollo: 43.41 fps APFast: 73.53 fps Chronos: 30.43 fps CHFast: 34.88 fps
Topaz Video AI v3.2.3
System Information
OS: Windows v10.2009
CPU: Intel(R) Core(TM) i7-6950X CPU @ 3.00GHz 31.925 GB
GPU: NVIDIA GeForce RTX 4080 15.688 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 27.29 fps 2X: 17.01 fps 4X: 4.68 fps
Proteus 1X: 23.80 fps 2X: 15.59 fps 4X: 4.63 fps
Gaia 1X: 10.34 fps 2X: 7.06 fps 4X: 4.39 fps
4X Slowmo Apollo: 31.92 fps APFast: 67.47 fps Chronos: 22.96 fps CHFast: 24.74 fpsTopaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: AMD Ryzen 9 3950X 16-Core Processor 31.923 GB
GPU: NVIDIA GeForce RTX 2080 Ti 10.782 GB
GPU: NVIDIA GeForce RTX 2080 Ti 10.782 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 15.97 fps 2X: 10.60 fps 4X: 2.98 fps
Proteus 1X: 15.81 fps 2X: 11.09 fps 4X: 3.01 fps
Gaia 1X: 5.06 fps 2X: 3.44 fps 4X: 2.50 fps
4X Slowmo Apollo: 6.87 fps APFast: 51.38 fps Chronos: 10.98 fps CHFast: 19.24 fps
Second run after graphics driver update and reboot then machine settle down:
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: AMD Ryzen 9 3950X 16-Core Processor 31.923 GB
GPU: NVIDIA GeForce RTX 2080 Ti 10.782 GB
GPU: NVIDIA GeForce RTX 2080 Ti 10.782 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 16.41 fps 2X: 10.90 fps 4X: 3.37 fps
Proteus 1X: 16.17 fps 2X: 11.44 fps 4X: 3.30 fps
Gaia 1X: 5.17 fps 2X: 3.53 fps 4X: 2.55 fps
4X Slowmo Apollo: 6.97 fps APFast: 45.34 fps Chronos: 11.20 fps CHFast: 19.53 fps
Results for 1080p and 480p input on a PC running Windows 11:
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: Intel(R) Core(TM) i7-10700F CPU @ 2.90GHz 31.857 GB
GPU: NVIDIA GeForce RTX 4070 11.744 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 16.20 fps 2X: 9.13 fps 4X: 2.49 fps
Proteus 1X: 14.08 fps 2X: 6.72 fps 4X: 2.55 fps
Gaia 1X: 6.01 fps 2X: 4.16 fps 4X: 2.69 fps
4X Slowmo Apollo: 21.89 fps APFast: 38.36 fps Chronos: 12.76 fps CHFast: 18.33 fps
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: Intel(R) Core(TM) i7-10700F CPU @ 2.90GHz 31.857 GB
GPU: NVIDIA GeForce RTX 4070 11.744 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 720x480
Benchmark Results
Artemis 1X: 94.75 fps 2X: 44.97 fps 4X: 12.96 fps
Proteus 1X: 70.96 fps 2X: 36.93 fps 4X: 12.36 fps
Gaia 1X: 36.50 fps 2X: 24.53 fps 4X: 16.37 fps
4X Slowmo Apollo: 107.43 fps APFast: 235.13 fps Chronos: 74.67 fps CHFast: 103.72 fps
Result for 1080p input on MacBook Pro 2019:
Topaz Video AI v3.2.2
System Information
OS: Mac v13.0001
CPU: Intel(R) Core(TM) i9-9880H CPU @ 2.30GHz 16 GB
GPU: AMD Radeon Pro 5500M 3.9844 GB
Processing Settings: device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 1.60 fps 2X: 1.13 fps 4X: 0.45 fps
Proteus 1X: ERR fps 2X: 1.12 fps 4X: 0.47 fps
Gaia 1X: 0.55 fps 2X: 0.38 fps 4X: 0.27 fps
4X Slowmo Apollo: 1.40 fps Chronos: 0.29 fps Chronos Fast: 0.63 fps
I’ll update this post with the results of 1080p and 480p input benchmarks with the latest version of Video AI on the MacBook Pro 2019 as soon as possible.
1 Like
@fmde - Interesting to see some RTX 4070 results but they leave me baffled! Your results for Artemis and Proteus 2x and 4x are less good than my old 3060 ti system yet on Apollo and Chronos, yours are faster than mine, at 480p MUCH faster. The latter is what I would expect but for me, the 4070 results on Artemis and Proteus are very disappointing indeed (I was thinking of getting one). But anyway, here are my results, have I missed something?
Topaz Video AI v3.2.3
System Information
OS: Windows v10.2009
CPU: AMD Ryzen 7 5800X 8-Core Processor 31.948 GB
GPU: NVIDIA GeForce RTX 3060 Ti 7.8613 GB
Processing Settings: device: -2 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 12.44 fps 2X: 9.13 fps 4X: 2.86 fps
Proteus 1X: 12.21 fps 2X: 9.59 fps 4X: 2.73 fps
Gaia 1X: 4.19 fps 2X: 2.88 fps 4X: 1.99 fps
4X Slowmo Apollo: 17.59 fps APFast: 43.17 fps Chronos: 9.76 fps CHFast: 16.24 fps
Topaz Video AI v3.2.3
System Information
OS: Windows v10.2009
CPU: AMD Ryzen 7 5800X 8-Core Processor 31.948 GB
GPU: NVIDIA GeForce RTX 3060 Ti 7.8613 GB
Processing Settings: device: -2 vram: 1 instances: 0
Input Resolution: 720x480
Benchmark Results
Artemis 1X: 71.46 fps 2X: 55.15 fps 4X: 14.11 fps
Proteus 1X: 71.80 fps 2X: 47.54 fps 4X: 13.35 fps
Gaia 1X: 23.77 fps 2X: 17.32 fps 4X: 12.16 fps
4X Slowmo Apollo: 59.97 fps APFast: 144.23 fps Chronos: 51.20 fps CHFast: 68.37 fps
I’m also baffled about the 4070 benchmark. Perhaps a CPU bottleneck? Here’s another 3060 Ti spec for good measure
Topaz Video AI v3.2.3
System Information
OS: Windows v10.2009
CPU: AMD Ryzen 7 5700G with Radeon Graphics 29.852 GB
GPU: AMD Radeon(TM) Graphics 1.9841 GB
GPU: NVIDIA GeForce RTX 3060 Ti 7.8496 GB
Processing Settings: device: 1 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 12.65 fps 2X: 9.23 fps 4X: 3.04 fps
Proteus 1X: 13.84 fps 2X: 9.32 fps 4X: 2.86 fps
Gaia 1X: 4.54 fps 2X: 3.07 fps 4X: 2.15 fps
4X Slowmo Apollo: 19.26 fps APFast: 44.24 fps Chronos: 10.11 fps CHFast: 17.12 fps
Topaz Video AI v3.2.3
System Information
OS: Windows v10.2009
CPU: AMD Ryzen 7 5700G with Radeon Graphics 29.852 GB
GPU: AMD Radeon(TM) Graphics 1.9841 GB
GPU: NVIDIA GeForce RTX 3060 Ti 7.8496 GB
Processing Settings: device: 1 vram: 1 instances: 1
Input Resolution: 720x480
Benchmark Results
Artemis 1X: 70.41 fps 2X: 50.78 fps 4X: 14.75 fps
Proteus 1X: 73.85 fps 2X: 47.62 fps 4X: 14.38 fps
Gaia 1X: 24.68 fps 2X: 18.17 fps 4X: 13.10 fps
4X Slowmo Apollo: 68.16 fps APFast: 145.32 fps Chronos: 54.60 fps CHFast: 69.81 fps
I’m new in that topic and those benchmarks are a little confusing to me. Do I understand it right, that the fps has to be as close or higher as possible to the source fps (which probably in most cases is 24 or 30 fps)? That would mean, that a benchmark fps around 50 is rendering two times of real time? Or a benchmark of 12 need twice of real time?
[quote=“rsvadmin, post:95, topic:42713, full:true”]
I’m new in that topic and those benchmarks are a little confusing to me. Do I understand it right, that the fps has to be as close or higher as possible to the source fps …?[/quote]
In the benchmarks for Artemis, Proteus and Gaia, the more fps the better (if the rendered quality is as good). So for a 24fps video, a benchmark of 48 fps would be 2x real time speed i.e. it should render in half the playing time of the video.
For say a 60 fps video, and the same benchmark of 48 fps, the video would render at slower than real time because in each second, 48 rendered frames is less than 60. It would take longer to render than to play the video at normal speed.
In the Apollos and the Chronos’s, the benchmarks can be confusing because when calculating the rendered fps, Topaz decided to count the new number of frames rendered not the original number, so arguably and IMO, they inflate the result. Depends on your point of view though.
1 Like
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: 13th Gen Intel(R) Core™ i9-13900K 127.75 GB
GPU: NVIDIA GeForce RTX 4090 23.59 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 720x480
Benchmark Results
Artemis 1X: 206.98 fps 2X: 87.20 fps 4X: 26.75 fps
Proteus 1X: 160.52 fps 2X: 82.67 fps 4X: 26.16 fps
Gaia 1X: 84.91 fps 2X: 62.56 fps 4X: 33.19 fps
4X Slowmo Apollo: 223.07 fps APFast: 453.04 fps Chronos: 168.52 fps CHFast: 209.70 fps
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: 13th Gen Intel(R) Core™ i9-13900K 127.75 GB
GPU: NVIDIA GeForce RTX 4090 23.59 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 40.34 fps 2X: 19.29 fps 4X: 5.97 fps
Proteus 1X: 34.61 fps 2X: 16.66 fps 4X: 5.69 fps
Gaia 1X: 15.85 fps 2X: 10.91 fps 4X: 6.13 fps
4X Slowmo Apollo: 38.85 fps APFast: 71.00 fps Chronos: 32.34 fps CHFast: 32.21 fps
I can only repeat that for 2X and especially 4X at higher resolutions, the CPU+RAM speed will be often have more impact than the GPU. That is because the image upscaling is performed in RAM by the CPU before it goes (primarily) to the GPU for AI processing.
Hope that clears up some confusion over certain benchmark results.
Thanks, I hadn’t realised that (or had forgotten) but unfortunately, it doesn’t really clear it up because the benchmarks don’t include RAM speed. They do include CPU though and when comparing my 3060 ti result a few posts up with the 4070 result above it, the CPUs benchmark essentially the same. But without RAM details we can’t compare, and I thought that was a main point of the benchmark threads?
So perhaps RAM speed should be added to the benchmarks?
I asked that first thing when they came out with the benchmarks. They told me it’s surprisingly hard to get RAM speed numbers. So, I think they have the intention, it’s just more work than anyone expected.
1 Like
There is no problem to get memory information in Windows by WIM (for example: class Win32_PhysicalMemory). Only thing make doubt - elevated user rights may required for obtaining information.
Hmm. Scores are very similar to the 4070 Ti
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: 13th Gen Intel(R) Core(TM) i9-13900K 31.787 GB
GPU: NVIDIA GeForce RTX 4080 15.688 GB
GPU: Intel(R) UHD Graphics 0.125 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 27.78 fps 2X: 17.22 fps 4X: 4.76 fps
Proteus 1X: 24.77 fps 2X: 15.96 fps 4X: 5.14 fps
Gaia 1X: 10.38 fps 2X: 6.88 fps 4X: 4.32 fps
4X Slowmo Apollo: 40.45 fps APFast: 70.62 fps Chronos: 23.02 fps CHFast: 31.57 fps
Topaz Video AI Beta v3.2.4.1.b
System Information
OS: Mac v13.0301
CPU: Intel(R) Core(TM) i5-7600K CPU @ 3.80GHz 32 GB
GPU: AMD Radeon RX 480 8 GB
GPU: AMD Radeon Pro 580 8 GB
Processing Settings: device: 2 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 2.88 fps 2X: 1.86 fps 4X: 0.72 fps
Proteus 1X: 2.86 fps 2X: 1.78 fps 4X: 0.74 fps
Gaia 1X: 1.12 fps 2X: 0.77 fps 4X: 0.57 fps
4X Slowmo Apollo: 2.25 fps APFast: 8.00 fps Chronos: 0.45 fps CHFast: 1.13 fps
Topaz Video AI Beta v3.2.4.1.b
System Information
OS: Mac v13.0301
CPU: Intel(R) Core(TM) i5-7600K CPU @ 3.80GHz 32 GB
GPU: AMD Radeon RX 480 8 GB
GPU: AMD Radeon Pro 580 8 GB
Processing Settings: device: 2 vram: 1 instances: 1
Input Resolution: 1280x720
Benchmark Results
Artemis 1X: 6.69 fps 2X: 3.85 fps 4X: 1.46 fps
Proteus 1X: 6.77 fps 2X: 3.99 fps 4X: 1.59 fps
Gaia 1X: 2.44 fps 2X: 1.71 fps 4X: 1.27 fps
4X Slowmo Apollo: 5.40 fps APFast: 19.72 fps Chronos: 1.49 fps CHFast: 2.43 fps
Topaz Video AI Beta v3.2.4.1.b
System Information
OS: Mac v13.0301
CPU: Intel(R) Core(TM) i5-7600K CPU @ 3.80GHz 32 GB
GPU: AMD Radeon RX 480 8 GB
GPU: AMD Radeon Pro 580 8 GB
Processing Settings: device: 2 vram: 1 instances: 1
Input Resolution: 720x480
Benchmark Results
Artemis 1X: 13.22 fps 2X: 7.77 fps 4X: 2.96 fps
Proteus 1X: 13.33 fps 2X: 8.16 fps 4X: 2.92 fps
Gaia 1X: 6.54 fps 2X: 2.52 fps 4X: 2.06 fps
4X Slowmo Apollo: 12.70 fps APFast: 41.39 fps Chronos: 1.49 fps CHFast: 4.30 fps
Topaz Video AI Beta v3.2.4.1.b
System Information
OS: Mac v13.0301
CPU: Intel(R) Core(TM) i5-7600K CPU @ 3.80GHz 32 GB
GPU: AMD Radeon RX 480 8 GB
GPU: AMD Radeon Pro 580 8 GB
Processing Settings: device: 2 vram: 1 instances: 1
Input Resolution: 640x480
Benchmark Results
Artemis 1X: 13.27 fps 2X: 7.49 fps 4X: 2.78 fps
Proteus 1X: 13.32 fps 2X: 7.93 fps 4X: 2.88 fps
Gaia 1X: 6.45 fps 2X: 2.50 fps 4X: 2.07 fps
4X Slowmo Apollo: 12.85 fps APFast: 37.51 fps Chronos: 1.48 fps CHFast: 4.27 fps
Topaz Video AI Beta v3.2.4.1.b
System Information
OS: Mac v13.0301
CPU: Intel(R) Core(TM) i5-7600K CPU @ 3.80GHz 32 GB
GPU: AMD Radeon RX 480 8 GB
GPU: AMD Radeon Pro 580 8 GB
Processing Settings: device: 2 vram: 1 instances: 1
Input Resolution: 480x360
Benchmark Results
Artemis 1X: 13.94 fps 2X: 9.14 fps 4X: 3.35 fps
Proteus 1X: 13.95 fps 2X: 9.19 fps 4X: 3.39 fps
Gaia 1X: 6.15 fps 2X: 4.44 fps 4X: 3.19 fps
4X Slowmo Apollo: 12.52 fps APFast: 41.32 fps Chronos: 2.18 fps CHFast: 4.61 fps
They would need to write a small RAM speed test themselves because with RAM speed, a lot depends on the so called sub-timings. Then there is DDR4 and DDR5, single- double- and quad- channel configurations. Latency is also important. So comparing RAM by just by specs is quite difficult.
Maybe they could add a pure image scaling benchmark where the GPU isnt involved at all. This would give a good indication about CPU + RAM speed in one important area of TVAI.
3 Likes
After NVidia 531.79 driver update today:
Topaz Video AI v3.2.3
System Information
OS: Windows v11.2009
CPU: AMD Ryzen 9 3950X 16-Core Processor 31.923 GB
GPU: NVIDIA GeForce RTX 2080 Ti 10.782 GB
GPU: NVIDIA GeForce RTX 2080 Ti 10.782 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 17.19 fps 2X: 11.32 fps 4X: 3.17 fps
Proteus 1X: 16.82 fps 2X: 10.30 fps 4X: 2.98 fps
Gaia 1X: 5.41 fps 2X: 3.71 fps 4X: 2.66 fps
4X Slowmo Apollo: 7.33 fps APFast: 42.17 fps Chronos: 11.75 fps CHFast: 19.14 fps
Topaz Video AI v3.2.4
System Information
OS: Windows v11.2009
CPU: 12th Gen Intel(R) Core(TM) i7-12700F 31.797 GB
GPU: NVIDIA GeForce RTX 3060 11.845 GB
Processing Settings: device: -2 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 9.60 fps 2X: 7.16 fps 4X: 2.31 fps
Proteus 1X: 9.42 fps 2X: 6.92 fps 4X: 2.32 fps
Gaia 1X: 3.09 fps 2X: 2.12 fps 4X: 1.45 fps
4X Slowmo Apollo: 12.60 fps APFast: 32.08 fps Chronos: 7.43 fps CHFast: 12.12 fps
Topaz Video AI v3.2.4
System Information
OS: Mac v13.0201
CPU: Apple M1 Ultra 64 GB
GPU: Apple M1 Ultra 48 GB
Processing Settings: device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis 1X: 12.36 fps 2X: 7.80 fps 4X: 2.95 fps
Proteus 1X: 12.27 fps 2X: 6.85 fps 4X: 2.28 fps
Gaia 1X: 4.85 fps 2X: 3.03 fps 4X: 2.46 fps
4X Slowmo Apollo: 8.26 fps APFast: 30.49 fps Chronos: 4.13 fps CHFast: 6.10 fps
