Gigapixel AI v4.7

Batch processing … I tried to process several png files setting the height to 2160 px in auto mode. It would process a few then crash to the desktop. When I loaded several files again, the “saved” data showed 2159 instead of 2160. Changed it to 2160 and the same file crashed.

Processing that file not in batch mode worked fine.

@svetlina_design If you are using Windows, go to the Start (program list)> Topaz Labs LLC > Gigapixel AI and run it. If it is 4.7.1 then your Icon link is pointing to the older version In Topaz Labs. Just drag the icon from the new install to the task bar or desktop (where you normally run from).

In v4.4.6 we could drag a folder of pictures into GP, in v4.7.1 I get the ‘File Type Not Supported’. Is it possible to enable this action again?

1 Like

@eric The new Gigapixel AI 4.7.1 crashes every time I follow these steps:

  1. Open image A by dragging it into the application window
  2. Make adjustments to the settings
  3. Close image A (either after saving or not saving)
  4. Open image B by dragging it into the application window
  5. This results in the application closing without notice.

This is on an iMac using OS X Mojave. I have tried reinstalling the application and restarting the computer, however the crashes persist.

2 Likes

When I save a file with height 2160 and load another file the sticky setting in height sometimes is 2159 so I have to change it back to 2160.

I checked and the time to process my test image with the various different settings and the results are identical to version 4.6.

Timing Test Results

Also the built-in calibration routine still sets both CPU and GPU to On if I’m using High or Medium memory allocation. With Low memory allocation it selects CPU only.

This bug is appeared specially when PhotoShop is opened during processing. Set “File → Preferences → Enable discrete GPU” as “No”, the problem will be solved.

Still selects both during calibration …

image

Processing 2400x1600 JPG 1.2x,

  • Both Active … - 1:05 minutes
  • GPU/Medium … - 1:04 minutes
  • OpenVINO … - 2:28 minutes

image

So, in summary, although it chooses both, it uses GPU processing because it is faster. It would be nice if the preferences actually reflected the correct benchmark medium.

That’s not a solution. It takes forever to process with the GPU disabled. Beside, I don’t have Adobe products on my computer, except for the pdf reader in my browser.

Please go to Help-> Graphics info, press Copy and paste the information here.

Also let me know if you have more than one GPU on your PC.

One graphics card

Application & Version: Topaz Gigapixel AI Version 4.7.1

Operating System: Windows 10 (10.0)

Graphics Hardware: GeForce GTX 1080 Ti/PCIe/SSE2

OpenGL Driver: 3.3.0 NVIDIA 445.87

CPU RAM: 65483 MB

Video RAM: 11264 MB

Preview Limit: 8000 Pixels

Hi, I really cannot see you having any issues with that configuration. I have a PC with 16GB RAM and a GTX1050/4GB and have no issues. Have you tried going to Preferences and clicking Use recommend settings then letting it recalibrate as that sometimes makes a difference. Also I would select GPU only with medium memory utilisation after that as it seems to work better.

I have the same problem. Drag and drop an image, process it, no problem. Drag and drop a second image, program shuts down without any error message. Windows 10 i9-9900 64GB RTX2080ti

1 Like

Doing the calibration didn’t help.

At first, I was thinking it was only the files that were being scaled down (the files in the batch group that were larger in height than my target height of 2160 px.) But found out that wasn’t the case. So now I need to see if I can spot a difference in the png format. Sometimes it fails on the 2nd file in the batch, sometimes the 9th.

I’m going to concentrate on testing various image file types and sizes.

1 Like

I found the crashing happens only when face refinement is on, in my case anyway.

I can confirm that here as well, hadn’t realized it.

So, to create the problem: drag an image. Turn face refinement on. Close the image. Drag another image.

After some testing, I’m 99.9% sure it’s something to do with the png format in some of the files. Some were screen captures, some from one of my cameras, and some generated via software (Terragen or Lightwave renders.)

I generally store my images in png format, but over the years I’ve used more apps, like Photoshop, Paint.net, Photoline, etc.

Anyway, I grabbed a bunch (400) of image files(all png) and tried a AI Gig batch to a height of 2160. Some failed causing AI Gig to crash. So I went back to the source (the 400 files) and batch converted them to .bmp’s us an old utility called “Bulk Image Converter”. I then used the same utility to convert them back to png’s and fed those to AI Gig and it processed all 400 without error.

The problem with that converter I was using, it didn’t show if it used compress or not with png files. So I tried the same test with Affinity photo in batch mode and AI Gig liked those too.

Anyway, I have a workflow that works for me now.

1 Like

I’m having troubles with blurry output images on 4.7. Never had this problem before. Maybe I’ll come back to 4.4.5.

Is this a different behaviour with the new version or is it what I experience and recorded here: - YouTube

Exactly the same. In the preview it show me a sharp image, but the output is half blurry and half sharp

https://i.imgur.com/hWlNpsA.jpg