How to speed up raster drawing

[emd55]

Hello,

I have a large (~25 GB) raster image that I am trying to view on a map, and it is extremely slow to draw.

I just want to see a 'course' version on the map to check that it's showing up correctly, and was trying to make the drawing more course than the data (30*30m) but can't find how to do that without actually changing the image resolution. I changed the symbology so that everything is one color and zoomed into a small area, allowed it to draw within that extent, then zoomed out slightly and repeat, but this process is extremely slow (over an hour so far to draw 1/20th of the full extent).

I am about to give up on viewing it on the map for now. It is using a lot of computer power, and the location where this image is stored does not have much/any(?) more space (but the computer hard drive does - which I think is where Arc needs space to do processing?).

Is there a way to fix this?

[JayantaPoddar]

Building pyramids should help. It would get stored (require some additional space) in the same workspace as the Raster data.

Also I hope the Raster data is stored in a local drive (non Onedrive). Keeping the Raster on a network drive or external Disk might adversely impact the performance.

There are other considerations to look at, like dedicated GPU, RAM, available free space, Drive type (SSD/HDD), etc.

[DuncanHornby]

A couple of top tips is ensure your raster has the lowest bit level possible, especially if your values are whole integers. If your raster has values with decimal points then it should be a float 32. If your raster holds only values of say 0 to 15 then you could store the raster as a 4-bit raster which would crush a raster down to size.

If your raster format supports compression ensure that is turned on to. Recently I was working on a project that was creating 100GB size rasters and it was because the raster calculator does not honour compression for a TIF format. By running it through the Copy Raster tool and ensuring LZW compression was on these rasters crushed down to 10GB!

So with sensible bit depths and what @JayantaPoddar  recommends you will get better performance.