AI Enhancer > how to use it

[Börje]

I'd like to test drive the new AI enhancer feature biut I'm not sure how to make it work.
I see that there is a new Multi-Pass pass called "AiEnhancer", but activating it makes no difference in the final render (with multipass enabled).

I guess I need to activate it somewhere else? The help unfortunately gives me no clue...

>>> SORRY. I have overseen an existing post which is right about this feature

[maru]

We have recently updated our help docs:

Here is how to use the enhancer - https://support.chaos.com/hc/en-us/articles/36244594382353-How-to-use-AI-Enhancer-in-Corona-for-Cinema-4D

It is also included in the documentation - https://docs.chaos.com/display/CRC4D/Corona+VFB+2.0#CoronaVFB2.0-CoronaVFB2.0Buttons

I hope all is clear now!

[Börje]

Thanks!
Yes I found the documentation shortly after my post. But nevertheless dpoing exactly what is documented I don't see the "enhance image" button.
But I have just opened a support ticket with a screen recording of my workflow as suggested by Tom.

[bnji]

Hi there,
I'm glad you also found the documentation.
Here's a suggested workflow.
1. Add the AI Enhancer multipass on your scene project
2. Render out your image
3. Once your render is complete, on the VFB 2.0 click on the AIEnhancer button (this will switch to the Collaboration tab)
4. Upload your image to the Chaos Collaboration cloud from there
5. Once the image is uploaded, open the image on the Collaboration cloud (once there, click on the image)
6. Once the image is open on the collaboration cloud, click on the Enhance button and wait for the result
7. Once the process completes, you can either switch between versions or even compare them.
I hope this helps.
Looking forward to hearing from you.
Kind regards.

[maru]

Continuing in support ticket #388434.

[celmar]

thank you, Borje!  in fact, here, when the render is finished, the ai button (sort of magic sign), stay greyed, unavailable???

[romullus]

Ai enhancer is only available when render resolution is within 512-2048 px range.

[tuami]

Hello,
we’re using Cinema 4D’s Take System to render multiple images in a single batch. Currently, the next rendering process starts immediately after the previous one finishes, which prevents us from using the AI Enhancer between individual renders.

Is there a possible workaround to avoid this issue? Alternatively, is there a way to apply the AI Enhancer to the rendered images afterward?

[xxxsalvo]

Hello,
we’re using Cinema 4D’s Take System to render multiple images in a single batch. Currently, the next rendering process starts immediately after the previous one finishes, which prevents us from using the AI Enhancer between individual renders.

Is there a possible workaround to avoid this issue? Alternatively, is there a way to apply the AI Enhancer to the rendered images afterward?

Same question here

[Nejc Kilar]

Hello,
we’re using Cinema 4D’s Take System to render multiple images in a single batch. Currently, the next rendering process starts immediately after the previous one finishes, which prevents us from using the AI Enhancer between individual renders.

Is there a possible workaround to avoid this issue? Alternatively, is there a way to apply the AI Enhancer to the rendered images afterward?

Same question here

I haven't tried this myself so I'm essentially just thinking out loud :)

I'd try saving my renders as CXRs (which should work with Takes), then once all the rendering is done I'd bring those back into the VFB and upload them to Cloud Collab that way.

That means there's still some manual work to be done but the rendering would be automated.

[xxxsalvo]

Hello,
we’re using Cinema 4D’s Take System to render multiple images in a single batch. Currently, the next rendering process starts immediately after the previous one finishes, which prevents us from using the AI Enhancer between individual renders.

Is there a possible workaround to avoid this issue? Alternatively, is there a way to apply the AI Enhancer to the rendered images afterward?

Same question here

I haven't tried this myself so I'm essentially just thinking out loud :)

I'd try saving my renders as CXRs (which should work with Takes), then once all the rendering is done I'd bring those back into the VFB and upload them to Cloud Collab that way.

That means there's still some manual work to be done but the rendering would be automated.

Thanks Nejc for your contribution!
I'll try for sure and I'll let you know. Hoping that in future these steps will be automated :)

[d4l1]

Thanks for the tip with the CXRs and re-uploading it via VFB into the Cloud. At least it sounds like that would work. :)

For our end, while testing the new AI Enhancer it seems like it does a good job with images that are 2K resolution max. I can upload our default image size of ~ 4K, but the AI really is making things worse when it comes to humans. Only did a few tests, but it works best with 2K images. This would result in a workflow like: Rendering the default size of 4K as usual. Then do a second round of rendering, now with a 2K resolution and CXR file format. Then the AI magic happens and afterwards we merge both images, the base rendering in 4K and the upscaled + enhanced JPG inside PS. Even though these are a few steps it may removes the annoying process of adding 2D cutout people at all. :)

[TomG]

I am not sure the AIEnhance multi pass is stored in the CXR, I seem to recall having to do this to work with the AI enhancer, which is use Continue from File and run 1 pass, to generate that multi-pass element, and then it could be submitted. Was a while ago that I tried it though, during development so things could have changed, but just as a heads up.

[tuami]

Hey, I found a workaround to use Takes along with the AI Enhancer afterwards. Maybe this solution will help you — just make sure not to close C4D in the process.