specify spatial reference for ExportToPNG

Hi,
We're using ExportToPNG() to export an image from a data frame, but we're also separately and "manually" writing out coordinates to HTML image maps. (By "manually", I mean we're not using a builtin package tool such as "ExportToHTMLIMageMap().")

Our problem is that the two don't align; the projection is apparently different. E.g. North America is "taller" (more stretched) in the exported HTML coordinates. Here is a screenshot in browser; the black outline shows the HTML image map coordinates:

Attachment 20995

When we export the coordinates to HTML, we're doing an explicit projection (Project_management()), specifying the spatial reference of the data frame (df.spatialReference).

But when we export the image using ExportToPNG(), there seems to be no way to specify a projection or spatial reference. Does anybody know how ExportToPNG() decides what projection to use, or how we can influence it? In ExportToLayoutGeoTIFF() you can supply a data frame for the spatial reference, but not in ExportToPNG().

Ideas we have currently...

1. Set it in the environment, e.g.: arcpy.env.outputCoordinateSystem = df.spatialReference
2. Tell ExportToPNG() to output a georeferenced world file, then try to examine the world file to see what "pixel scale information and real-world coordinate information" it contains (how do we examine it?)
3. Before ExportToPNG(), do an explicit Project_management() with the data (as we do before the HTML coordinate export), allowing us to specify the spatial reference. Repoint layers that use this data to use the new, projected data. I believe I was told that data frames (data view) show data in geographic coordinates, not projected coordinates, so I don't know if this would have any effect, or is even coherent.

Any suggestions would be appreciated.

Attached Thumbnails