こんにちは。
Windows版 ver1.04を使用しています。3DCGAnimeの画像を加工する際に、フォトショップ版フィルター機能およびスタンドアローンTopaz Photo AIのどちらを使用しても、ファイルの読み込みを行い自動処理をさせた時点で、添付した画像のように色味がおかしくなってしまうのですが、原因が分かる方はいますでしょうか？
Hello, I am using Topaz Photo AI for Windows ver 1.0.4
And when I process 3DCGAnime images, whether I use the Photoshop filter function or the standalone Topaz Photo AI, when I load the file and let it process automatically, the color tones are not right as in the attached image. Does anyone know what the cause is?
Could it be your using CMYK rather than RGB? Because CMYK may not be supported.
Not a major problem, but when I installed both Photo AI 1.0.2 and 1.0.4, the icon was not found in the shortcut it put on the desktop. Only a white rectangle appeared. I was able to go to the Program Files directory, create a shortcut from the .exe file, and move it to the desktop manually and it’s fine. Thought it was just a one time glitch, but the same thing happened when the update installed so maybe it’s a bug.
Win11, LR Classic, Photo AI plugin for .nef files. So far so good but my nef edits are not carried through to the dng file. Is this because .xmp files are not being read? I know that I can copy & paste edits but DXO deep prime does not have this problem. Apparently no support for Nikon Z9. Is this correct? I use a Z9 and it seems to work ok. Masking in the sharpening section is poor. Need to be able to add and subtract mask as in Sharpen AI, ideally painting over areas with brush.
Certainly .xmp files will not be read as they are not Topaz XMPs.
You can check if the Z9 is supported either on the LibRAW website or raise a support request at the main website.
Just updated to latest release of Topaz photo 1.0.4 with MBA 2020 M1 (all software up-to-date) dragged a low resolution png file into the program and it inverts the colours - see image
Thanks for the verification.
In my case, this phenomenon occurred with 10241024px images. if I made them 20482048px and then loaded them into TP AI, this phenomenon did not occur and the conversion was successful. I guess the current version of TP AI requires the image to have a certain resolution.
Photo AI Version 1.0.4 is the best yet.
As for added features, I would like to see:
Its Virtual Memory, the OS is using the drive as a very slow ram.
Hi,
I’m getting this issue whilst batch processing on an M1 MacBook Air, any idea what it could be, can provide more info if necessary.
Thanks,
Simon.
1.0.4 has come a long way. Thanks for all of your work, team.
Thanks, again.
There’s a strong color shift no matter what filter I use on my picture, could you check it please?
Same behavior when processing batches of about 100-150 .jpg files using autopilot. Win 11, Ryzen 5800X, RTX 3080, 32 GB RAM. Sometimes it would complete the batch, some other times I have to open Photo AI again and resume from the last file it left off.
This seems to be the same range where this fails…Some programs (Lightroom as an example) will let the user set a preference range for disk swapping… Maybe there’s a built-in range in PAI that needs to be tweaked, and/or moved to the Preference menu so the user can set that range… That 43+G is disk swapping not RAM…
if you quit PAI, then restart, an update button appears on the upper right corner if updates are available. click there to start the update if desired.
Thanks. In fact I see that the Z9 is now supported in the LibRAW 0.21(Beta 1). Not sure how to proceed.
I guess they would not put the beta in the production but you can raise a support request asking that Z9 support is hurried along.
For me (I have only Photo AI) it works amazingly well in a workflow with Photoshop. The best this that it get constantly better. This will be their only photo software in the future. But you will have to maybe get use to a new workflow.
Obviously this is an extreme example but you can see that Autopilot selects the Standard Sharpen model which is not the correct one:
If I select the Motion Blur model instead and crank it right up it does do some amazing recovery:
Again this is only a controlled TEST image and you would never expect 100% recovery but, at least, you would expect that the correct model is selected.
Also the current Masking is not suitable without being able to manually adjust the mask.
Yes. I’m also waiting for that. I solve it by by setting the sensitivity to cover all interesting areas and then in Photoshop masking the areas that i want to exclude.
