read_region from level=0 takes too long

[reyvateil]

Context

Issue type (bug report or feature request): bug
Operating system (e.g. Fedora 24, Mac OS 10.11, Windows 10): Debian 10
Platform (e.g. 64-bit x86, 32-bit ARM): x86-64
OpenSlide Python version (openslide.__version__): 1.1.2
OpenSlide version (openslide.__library_version__): 3.4.1
Slide format (e.g. SVS, NDPI, MRXS): MRXS

Details

Slides at level=0 are roughly 200,000 x 100,000 pixels. When trying to read_region 512x512 tiles from level=0 every tile takes anywhere between 40s to 2min to read. In contrast, when trying to read 512x512 tiles from level=1 (100,000 x 50,000 px) it only takes less than a second.

Is such a jump in extraction time expected?

[bgilbert]

No, that isn't expected. If you're still having this problem, please post enough info to allow us to reproduce this, or at least post the relevant section of your Python code.

[YuyanRuan]

Same here. I guess there may be due to the nature of the wsi itself. But not sure what's the problem

[sj9515]

Having similar problems a year later. I am opening the slide, reading the region and saving it as a PNG image. Depending on the slide, it takes between 30 seconds to almost 3 minutes at times. Any suggestions?