![]() ![]() ![]() Anyways, lots of benchmarks will be revealed. I won't be getting the fully spec'd out model either. I was looking a the fine print to that and it seems its comparison is to the imac which is strange since i rather see the performance from prior m1 chips. ![]() I picked up an openbox Mac Studio base model too so you can see those results too.Īrtemis 1X: 07.07 fps 2X: 05.11 fps 4X: 01.18 fps In Topaz there wasn't a difference in speed between the base and upgraded model and the CPU was slightly out performing the M1 Pro/Max 10 Core anyway so didn't think it was worth the extra money. Gaia 1X: 2.70 fps 2X: 1.66 fps 4X: 1.36 fpsĤX Slowmo Apollo: 7.01 fps Chronos: 2.66 fps Chronos Fast: 4.69 fps Here's my 16" M2 Max (14" is quite a bit slower due to the lower TDP on the system)Īrtemis 1X: 11.28 fps 2X: 6.78 fps 4X: 2.41 fps Gaia 1X: 4.87 fps 2X: 3.16 fps 4X: 2.34 fpsĤX Slowmo Apollo: 8.48 fps Chronos: 4.12 fps Chronos Fast: 6.11 fps Processing Settings: device: 0 vram: 1 instances: 1Īrtemis 1X: 13.02 fps 2X: 7.74 fps 4X: 2.87 fps The M2 Max 38 core is right there with the 48 core model. The Ultra is clearly faster than the other M1 variants now. There's been quite a few changes since my post last year. It is not perfect, I didn’t spend a lot of time on it and I certainly didn’t run model comparisons - it is more as an example. There is vertical judder being caused by bad field matching, and - at least from the clip you gave - given there is no visible interlacing in the fields, you are probably better off deinterlacing outside of Topaz first.Īs an example, this is your clip, processed to remove as much of that crappy interlacing and vertical judder as I can without spending a lot of time on it (and the entire file in case there is issues) with a general upscale and IVTC to 23.976. I see no way that this clip could be imported into Topaz and have the output be good without attempting to correct the input source first. Having said that, just from the clip you gave, its not obvious if the clip contains different framerates as it appears to be hard telecined (or similar) with a very rough interlacing. Its also one of the causes of Topaz Interlaced models not working properly, because these models don’t handle it well - and they probably won’t ever. Its also why VFR content is a pain the butt to make properly progressive. Thanks in advance for your feedback!ĭuplication in cartoons is generally because a lot of animated content has variable frame rates - most typically slow - meaning trying to squash it into a a set framerate often means having to add missing frames to extend it. Please take a look at the Video Roadmap Update for more context on what we’re focusing on right now. Please reach our to our support team if the issue persists.The location of this folder may be found in the preferences menu. If all models fail, please try deleting the.If you are logged in and receiving a watermark on Windows, try deleting the %PROGRAMDATA%\Topaz Labs LLC\Topaz Video AI\models\auth.tpz file while the app is closed.Preset name is limited to max 20 characters.Fixes output resolution not showing correctly when some non-square pixel videos are converted to square pixel.Fixes auth issues for users who run the installer using a different user than used for running TVAI.A new release of Topaz Video AI is now available. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |