d3d9: Add DF16 support

[Henri Verbeet]

2009/6/7 Frank Richter <frank.richter at gmail.com>:
> On 07.06.2009 19:35, Henri Verbeet wrote:
>> 2009/6/7 Frank Richter <frank.richter at gmail.com>:
>>> As far as I could gather DF16 is the "ATI way" of getting a renderable
>>> 16 bit depth texture.
>> Without knowing much about the actual format, DF16 implies this should
>> be a floating point format, similar to the ones provided by
>> ARB_depth_buffer_float.
>
> Maybe... but it seems that format is solely intended for use on render
> target textures, not any download or upload... so not sure if the data
> type would matter in practice. Also, I didn't find a 16-bit float depth
> texture format for OpenGL so far.
>
Even if the format isn't lockable, you can still use the data with a
shader. Typically float formats aren't normalized, so you can have
values outside the traditional [0.0,1.0] range. If there's no specific
extension for this format you could use GL_DEPTH_COMPONENT32F as
internal format and GL_HALF_FLOAT_ARB as type, although that would
waste some memory, of course. Is there any specific application that
needs this format?