Questions about tile-serving

[Priv.-Doz. Dr. M. Weihrauch]

Dear community:

I am currently experimenting with an own C# tile serving script
integrating OpenSlide to serve in a deepzoom format.
I therefore have some basic questions:

1. If I am thinking about that correctly, the steps are briefly:
- A deepzoom tile is requested, e. g. zoom level 14, x1, y2
(image/14/1_2.jpeg), overlap=0, tile size 256
- I get the total width / height of the full resolution image (zoom
level 0 of OpenSlide). With that, I can calculate my deepzoom pyramid,
with the width and the height of each zoom level and therefore also the
number of tiles - fortunately, I already know how to do that.
- E. g. with zoom level 14, let's assume, I calculated the layer size on
that zoom level 14 with 16,321 pixels width, 9,821 height.
- If I am not mistaken, I would now have a look in OpenSlide, if that
specific zoom level according to the dimensions exists in the original
image, i. e. one of the levels of the original image has also
approximately 16,321 pixels in width. If yes, I extract my e. g. 256x256
directly. If not, then I take the next bigger zoom level, extract
512x512 and resize to 256x256?
- Are the zoom levels for most openslide images dividing by 2 and 4 for
the next zoom level or are there some other exotic formats?

2. If the deepzoom overlap is 1, does that mean that I have to add 1
pixel on the right and on the bottom border of the tile or 1 pixel to
each side (given that this is not a tile from the image border)?

Thanks!!

Martin