If the light cache was calculated at say 640x480 but the the final rendering was 1280x960. Would using the lower res maps for higher res image give predictable result?
Thanks
Many thanks for the information Svetlozar I never knew that, thats going to save me plenty of time!! If only I realized this a few years ago…
Cheers
So we can bake the LC with a little resolution (10x6) and re-use the map for the final resolution.
But, the question is : bake the LC with a little resolution will be more faster than bake the LC on a final resolution ? or we do not have benefice to do this ?
There is no differences in the render-time between LC calculated on different resolutions , otherwise it won’t be resolution independent 
According to my tests there is a very very small difference when using different resolution - it is not noticeable in the render but it is there.
10x6 resolution = 269kb LC map
1000x563 = 269kb LC
10000x5630 = 281kb
The fired rays between the second and the third are also slightly different - but I haven’t found any differences in the final render.
OK thanks for that I naturally assumed it would be much quicker to calculate smaller maps, but as you say that’s not the case, I don’t feel so daft after all!
Yes this is normal because everything works much faster on smaller resolutions - but it is because almost everything is resolution dependent, which is not the case with LC.
It would be fantastic if there is a way to calculate everything on low-res and then just do a final render on high-res with the same quality - but unfortunately there is no such thing as free lunch.
We have to pay either with the quality or with the render-time.