garageCube

[KelleyBean]

A very basic question about how MadMapper handles image/media resolution...is there a default size at which MM imports video or stills? I ask because it seems that no matter what size I import images in the default appears to be 1920 x 1080, even if the image itself is only 100px x 100px and regardless of projector settings. Is there some sort of upsampling happening here, and if so is it better to enlarge smaller images, work to a size that corresponds to final output, or to use bigger images and downsample them?

In short, I'm trying to composite a slideshow of images within a larger animation. The images are stills that are taken on site, saved to an image folder, and then loaded into MM as they are collected. I'm trying to determine what the native res should be for these images to set my camera accordingly.

[caneron]

MadMapper imports all images at the right resolution (see the input view), and outputs all images at the right aspect within your projector frame (I.e., max height width @ projector resolution).The output resolution won't match your input resolution unless you've already set your images (in bridge or something) to be max height 1080, max width 1920 (or whatever your projector resolution is). You can't output higher resolution than there are pixels on the projector, and similarly you can't scale images higher than their default resolution and expect it to be clean. In your example, a 100px x 100px image would default to 1080x1080 if your projector is 1920x1080.

Unless you are needing flexibility from image to image for mapping (not layout, but literally mapping), I recommend that you create your slideshow in another program - like Premiere or After Effects.

If for some reason you need the exact resolution matched, you can copy the quad points' input positions from the input view to the quad points output positions in the output view. But I don't think that will result in what you're asking for.

If your question is about literally: how does MadMapper interpolate scaling beyond 100% - I think it's like any other program. I don't think it's upscaling the resolution as a GPU process (like logical upscale in AE) or anything.

[KelleyBean]

I think that's a great answer! Thank you! Just to re-state and make sure I understand...I've included a screen shot to make things a little clearer...

The final project is a "wall" of animated elements (in the screenshot, the large b/w animation) with a number of smaller quads that will contain portrait images (the smaller square color images). This piece would be at a festival, and there would be an on-site photo booth that would send the photos to a folder, which would then in turn populate the photo booth images in these quads. So, people would be encouraged to stick around, watch the piece change and wait for their image to pop up. Ideally, there would be two people operating the piece: one to run the camera, and another to work MadMapper.

So, if I read you correctly, if the projector resolution was 1920 x 1080, I'd want to shoot at least at 1080 x 1080 to ensure good image quality, even if they are downsampled. Does downsampling larger images take a lot of processing power, i.e. do I make any gains by trying to keep the image sizes as small as possible or does it not matter?

Also, if I can sneak another question in here...I have noticed that new images added to image folders do not populate if they are already running in MM. Would live updates to folder content require a Syphon-based object or am I missing something?

Thanks so much for your help!

Kelley