map.view.zoom explanation for 2D maps

[cbashioum]

(extracted from CMAPI google Group Post https://groups.google.com/forum/#!topic/common-map-widget-api/DjyIA0kUO1s)

I believe you are referring to https://wiki.openstreetmap.org/wiki/Slippy_map_tilenames#Resolution_and_Scale - is that correct? There is also a discussion on zoom levels at https://wiki.openstreetmap.org/wiki/Zoom_levels

Here is a question for the those on this list – would it be reasonable for us to just reference the OpenStreetMap wiki pages above in our explanation of the range in map.view.zoom (and perhaps in other “zoom” payload objects in other places in the spec)? Would this be a problem for anyone?

When we reference it, we can reference it as an example, or we can reference it as “the way it SHOULD be done”. Thoughts??

From: common-map-widget-api@googlegroups.com [mailto:common-map-widget-api@googlegroups.com] On Behalf Of Anders Alstad
Sent: Wednesday, June 15, 2016 5:06 AM
To: Common Map Widget API common-map-widget-api@googlegroups.com
Subject: zoom range and "closest zoom level"

Hi,

I see that the documentation for "range" in "map.view.zoom" channel has been updated, but it is still unclear. For a 2D map, what is the relationship between, "range", "distance in meters from the map" and the "closest zoom level"?

E.g. the OpenStreetMap wiki documents the relationship between the most used zoom levels, scales (depending on dpi) and resolutions (horizontal tile size in m/px). Do "range" in "map.view.zoom" channel represent one of these numbers (level, scale or resolution)? If not, please define how to calculate the "closest zoom level" from a given "range".

Regards,
Anders. A.

[eric-briscoe]

http://icaci.org/files/documents/ICC_proceedings/ICC2009/html/nonref/1_6.pdf

Describes calculations fro determining scale from a camera above ground. Use this as a reference to resolve issue

[truff]

The conversion is not possible to do without specifying additional information about the camera (e.g., focal length or field of view). I suspect that most people just want it to "look like Google Earth" though. GE appears to use a field of view of 60 degrees. I suggest we pick a value for field of view and document it.

[eric-briscoe]

An example of javaScript conversion code between range, scale, and zoom level can be found here:

https://jsfiddle.net/e82t0xzc/11/

@truff @cbashioum Please let me know if you agree with this approach and we can determine how to add to CMAPI spec

[truff]

My gitHub notifications were going to an old e-mail address so I'm just now seeing this. Here are my comments on the fiddle:

The fiddle first converts zoom level to scale, then scale to camera height above surface. There is no 'correct' scientific or mathematical way to map zoom level to scale. That mapping is arbitrary. We don't care what mapping a given 2D CMAPI compliant map uses. Their map toolkit or map application code will provide them a means of getting map scale regardless of whether or not they even use a map zoom level. Ultimately any 2D map needs to convert scale to height above surface in order to work with CMAPI. So the only two functions in the fiddle that matter are scaleFromRange() and rangeFromScale(). These simply use a magic number to perform a linear transformation (scaling) between the two numbers. The real question is how was that magic number determined?

It could be determined empirically by comparing, for example, Google Earth at a given camera height to any 2D map at a given scale level, then seeing what works. If you do it this way you are unknowingly basing your conversion (i.e., your determination of the correct "magic number") on GE's field of view (which is fixed at 60 degrees). You could also determine the conversion using trigonometry. But if you try this you will quickly realize you need to know the field of view (or, equivalently, focal length) in order to generate the conversion equation.

So what does all of this mean for the CMAPI spec? We could just tell everyone to use that formula with the magic number, or we could document the field of view that CMAPI uses and allow developers to create their own conversion routines from first principles. I think the latter option is a better way to write a spec.

[truff]

Maybe it makes more sense if the CMAPI API accepts a map scale for the zoom parameter instead of height above surface. Then 3D maps can convert to height above by using their map's field of view setting and we don't have to worry about specifying one in CMAPI.