Topaz Video AI Performance Alpha 3.2.0.3.a

[suraj]

Hi All,

Here is an early version of app that should improve performance for all machines. The degree of performance increase will vary. This version should also allow for better support running multiple filters and multiple processes.
Please let us know if you have any issues using the app.

Download: Win | Mac | Mac Update

Please upload problem videos and logs here: Submit Files

Changes from 3.2.0.2.a

• Proteus Auto/Estimate works now
• Running single process should be faster for non RTX GPUs (Max Processes = 1)
• GUI updates from 3.1.9 and beyond

Thank you for testing

PS: Since this is an alpha version, it will not interfere with Beta or Release versions on the machine.

[david.123]

Sir, your team is running like a well oiled machine! Cranking out these updates, love it!

Rome is coming along nicely!

[david.123]

Looks like the ‘grey interpolated frames’ issue is still present. Known issue on the dev side of things?

All interpolated frames come out like this:

[suraj]

Is this only happening in the alpha or the release version as well?

[david.123]

Only in the alpha (is in the previous alpha and possibly earlier too).

[david.123]

Was first reported here:

Topaz Video AI Performance Alpha for Windows Nvidia & AMD GPUs

With 360p → 720p (that destroyed D&D cartoon FTW!) I see quite an increase, from around ~34fps in latest v3.1.7 stable, to ~45fps in this alpha. That’s on RTX 3090, 100% memory set, on i9 7940X CPU (fairly old one) and 64GB of RAM. Preview parameters: ffmpeg “-hide_banner” “-nostdin” “-y” “-nostats” “-t” “15” “-ss” “103.736” “-i” “W:/Video/Dungeons & Dragons Episode 19 The Last Illusion.mp4” “-flush_packets” “1” “-sws_flags” “spline+accurate_rnd+full_chroma_int” “-color_trc” “2” “-colorspace” “…

[david.123]

Release version does not have this issue.

[mikmod1]

Only in two previous alpha versions, so far. Testing with 3.2.0.3.a now.

Edit: This alpha has the issue as well.

[xl5 (William)]

I’m finding it hard to keep up with all these Alpha and Beta releases that I’m not sure which one is the best for doing certain tasks I want to complete? I guess the idea is to try something on both and then decide which is better, therefore instead of doing an entire video of (say) 30 minutes, just take a five minute segment and test it on each.

Keep up the wonderful work!

[david.123]

Stick with the release versions if you actually need consistent results, the point of the Alphas and Betas is to test early features and find bugs, if any. Not recommended to use as your daily driver.

[toby.riding]

Topaz Video AI Alpha  v3.2.0.3.a
System Information
OS: Mac v13.0201
CPU: Apple M1 Ultra  64 GB
GPU: Apple M1 Ultra  48 GB
Processing Settings: device: 0 vram: 0.5 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 14.09 fps 	2X: 8.35 fps 	4X: 3.06 fps 	
Proteus		1X: 12.23 fps 	2X: 7.5 fps 	4X: 2.63 fps 	
Gaia		1X: 4.07 fps 	2X: 3.03 fps 	4X: 2.39 fps 	
4X Slowmo		Apollo: 14.23 fps 	Chronos: 3.69 fps 	Chronos Fast: 5.54 fps 	

Slightly faster on most models, slightly slower on the others compared to the previous beta, not even the previous alpha version… just so you know & have the benchmark data with this alpha version. Nothing more nothing less!!

[TPX (Thomas K.)]

Topaz Video AI Alpha  v3.2.0.3.a
System Information
OS: Windows v11.2009
CPU: AMD Ryzen Threadripper 3960X 24-Core Processor   127.88 GB
GPU: AMD Radeon PRO W6800  29.618 GB
Processing Settings: device: -2 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 10.11 fps 	2X: 5.99 fps 	4X: 2.07 fps 	
Proteus		1X: 9.36 fps 	2X: 5.89 fps 	4X: 2.02 fps 	
Gaia		1X: 4.73 fps 	2X: 3.18 fps 	4X: 2.19 fps 	
4X Slowmo		Apollo: 12.19 fps 	Chronos: 5.56 fps 	Chronos Fast: 10.06 fps 	

---

3.2.0.1a

---

3.1.9

[TPX (Thomas K.)]

@suraj could we get a benchmark for Photo A.I. too?

[joachim.ehrhardt]

Topaz Video AI Alpha  v3.2.0.3.a
System Information
OS: Windows v11.2009
CPU: AMD Ryzen 7 5800X 8-Core Processor               31.916 GB
GPU: NVIDIA GeForce RTX 3060  11.845 GB
Processing Settings: device: 0 vram: 1 instances: 0
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 9.06 fps 	2X: 5.84 fps 	4X: 1.31 fps 	
Proteus		1X: 8.29 fps 	2X: 5.53 fps 	4X: 1.37 fps 	
Gaia		1X: 3.14 fps 	2X: 2.12 fps 	4X: 1.38 fps 	
4X Slowmo		Apollo: 6.87 fps 	Chronos: 6.94 fps 	Chronos Fast: 11.43 fps 	

[joachim.ehrhardt]

Why is Proteus Auto 15% slower than Proteus Manual, 4.4 fps versus 5.2 using same video ?

[alanhusband-677420]

“Auto” vs “Manual”.

“Calculating parameters for each frame” vs “fixed parameters set by the user”.

The extra calculations will have additional performance penalty.

[joachim.ehrhardt]

are you sure that “Auto” calculates parameters frame by frame?

[alanhusband-677420]

We know that the parameters change over time based on the video content with “Auto”. So there must be some frame analysis calculations occuring. Otherwise what’s the point of “Auto”?

[JesusSheep]

Topaz Video AI Alpha  v3.2.0.3.a
System Information
OS: Mac v13.0201
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: 2.62 fps 	2X: 1.75 fps 	4X: 0.67 fps 	
Proteus		1X: 2.91 fps 	2X: 1.94 fps 	4X: 0.8 fps 	
Gaia		1X: 1.02 fps 	2X: 0.72 fps 	4X: 0.55 fps 	
4X Slowmo		Apollo: 2.49 fps 	Chronos: 0.3 fps 	Chronos Fast: 0.7 fps