Optimize Editing UI Performance

I think there are some optimizations that are available that can substantially improve the rate at which I can edit images. A few notes:

  1. It seems the image caching can be made more effective. If I click to an image, sometimes the image updates 2-3 times before the final image appears. If I immediately click over to another image, and then click back, it seems that previous image wasn’t cached, and again I go through 2-3 updates before the image appears. Why isn’t the cache populated immediately once the image is rendered?

  2. Why doesn’t Photolab perform this processing during idle time on images in front or in back of the current image, so when I click over to the next image, the render is already completed? PL could theoretically process previews for a whole directory while I go get a coffee and I would be able to run through a whole folder and do cropping and straightening much much faster.

  3. Why does the image have to “regenerate” when I switch from straightening (rotation) to cropping? When I switch, the image disappears for a second before it reappears. This seems unnecessary.

  4. Once the image has been fully rendered, and I grab a control (like selective tone or exposure), why do I get a low resolution preview while I’m dragging the slider? Can’t the previous render be used for the adjustment preview?

  5. Why do local adjustments require a “full preview”? (I have no doubt there is a good reason.) But why can’t I start to draw my mask (or draw the gradient) while the full preview is being generated? Often, I’m just trying to adjust some lighting variation on the edges of the frame, and I can draw the gradient even without the full preview completed.

There are others, no doubt, but these are the slow-downs that are biting me at the moment.

For the record, most of my shots are from a 50MP camera. And my machine is an 8 Core Ryzen 9 with an RTX3070 GPU and 8GB VRAM and 40GB of system RAM. Yes, high megapixels, but Definitely higher than average machine horsepower…

Improvements in this area can help everyone…

I absolutely agree with you. Sometimes I switch forward and backward between two or three images in the filmstrip to compare them. And always PL starts to rebuild the image again and again to show it on the screen. And my machine is an Intel(R) Core™ i9-9900K CPU @ 3.60GHz with 64 GB RAM, so that should not be the problem.

1 Like

Even if some of these bitmaps aren’t cached to disk there’s no reason they couldn’t be cached in available RAM.

But what will happen when you need every available byte for exporting using DeepPRIME?

That’s only during export. And I had no problems using DeepPRIME exporting 4 50MP images simultaneously when I had 16GB. So I don’t think it’s a problem.

Why are you always so negative on my ideas? You don’t have to comment on every post?

Seriously you put down every one of my suggestions. I don’t get it. (I seriously thought this idea was Joanna proof. When I posted I wondered will Joanna find some reason to poo-poo this idea. And I thought. No. Everyone benefits from improved performance. I guess I was wrong.)

I’m sorry if it appears that way. I have this weird brain that seems to always think “but what if?” It’s something I used to get big money for as a software design consultant.

And it’s not just you. All great ideas need to stand up to argument, even mine (when ì get them) :wink:

1 Like

Oh and to seriously answer your question, this kind of problem has been solved for decades in computing. Just add a garbage collection process (or low memory event handler) to dispose of the oldest cached images from RAM when memory is starting to get low.

A great idea.

But then I just had a thought… how are you setting the preview quality?

Kind of irrelevant. However the preview is determined it should cached.