Transparency Tests For FS9 Only
I have seen many discussions about transparency and it is hard not to
be confused by this topic in FS9 or FSX.
MS FS9 SDK's do not clarify anything on this topic but strangely only
seem to confuse the matter.
When to use glass_t.bmp versus glass.bmp and what does that _T really
do?
And, what about makemdl's options concerning "HasReflectMap" and
"HasLightMap" have to do with this aside from adding more confusion.
I am using variations of a glass.bmp texture for these tests.
The primary color for all textures is always darkgray at 55,55,55.
I vary the alpha channel to change appearances and transparency.
So, the fact is that you can use either glass_t.bmp or glass.bmp but
you must treat them differently in gmax.
Glass_t.bmp requires that you set opacity in gmax at your desired
transparency level. I use 50% opacity for the glass_t.bmp side of the
test.
To test the Glass.bmp (no _t), I use 100% opacity and vary the alpha
channel darkness to get more or less transparency.
I use a specular level setting of 37 for both tests to get sun glare.
Here is the setup:
Glass.bmp Opac % Alpha ---- Glass_t.bmp Opac % Alpha
--------- ----- ----- ---- ----------- ----- ------
glass1 - - 100 - - 200 ---- glass1_t - - 55 - - 230
glass2 - - 100 - - 150 ---- glass1_t - - 55 - - 200
glass3 - - 100 - - 125 ---- glass1_t - - 55 - - 170
glass4 - - 100 - - 100 ---- glass1_t - - 55 - - 140
glass5 - - 100 - - 075 ---- glass1_t - - 55 - - 110
This test shows that glass.bmp (no _t) uses the alpha channel
effectively to control transparency.
It also shows that specular setting of 37 operates equally well to
yield sun glare with both techniques.
So, what is the glass_t technique doing differently?
The effect of using the _t and darkening the alpha channel seems to add
more reflective glossiness. Check Holiday Gift Guides and Michaels Ad.In doing that, it effectively decreases
the appearance of transparency as if the glass is becoming more
mirror-like as the alpha channel goes darker. It's an interesting
effect depending on your needs. However, you can retain much of that
mirror-like look while increasing transparency if you go back to gmax
and reduce opacity to say, 40% instead of 50%.
I also found that makemdl export option noted above ("HasReflectMap")
have some bearing on these results.
By unchecking that option, it eliminates the glossy reflection; the alpha channel is then used to
control the transparency beyond the 50% opacity stated in gmax. I
recommend that you leave the reflect option turned on.
In summary, the glass.bmp using the alpha channel to control
transparency is the effective and recommended basic approach.
However if your project requires a glossy-like or mirror-like, transparent
window or surface, you can use the glass_t.bmp approach in conjunction
with a lowered opacity setting for the material in gmax.
Keep in mind that a glass_t.bmp (or any use of _t) with 100% opacity in
gmax and a dark alpha channel will get you chrome.
I hope this test clears up some things, and helps spur some thoughts
and experimentation for you.
I am including the gmax source and bmps for you should you wish to run your own tests.
I have seen many discussions about transparency and it is hard not to
be confused by this topic in FS9 or FSX.
MS FS9 SDK's do not clarify anything on this topic but strangely only
seem to confuse the matter.
When to use glass_t.bmp versus glass.bmp and what does that _T really
do?
And, what about makemdl's options concerning "HasReflectMap" and
"HasLightMap" have to do with this aside from adding more confusion.
I am using variations of a glass.bmp texture for these tests.
The primary color for all textures is always darkgray at 55,55,55.
I vary the alpha channel to change appearances and transparency.
So, the fact is that you can use either glass_t.bmp or glass.bmp but
you must treat them differently in gmax.
Glass_t.bmp requires that you set opacity in gmax at your desired
transparency level. I use 50% opacity for the glass_t.bmp side of the
test.
To test the Glass.bmp (no _t), I use 100% opacity and vary the alpha
channel darkness to get more or less transparency.
I use a specular level setting of 37 for both tests to get sun glare.
Here is the setup:
Glass.bmp Opac % Alpha ---- Glass_t.bmp Opac % Alpha
--------- ----- ----- ---- ----------- ----- ------
glass1 - - 100 - - 200 ---- glass1_t - - 55 - - 230
glass2 - - 100 - - 150 ---- glass1_t - - 55 - - 200
glass3 - - 100 - - 125 ---- glass1_t - - 55 - - 170
glass4 - - 100 - - 100 ---- glass1_t - - 55 - - 140
glass5 - - 100 - - 075 ---- glass1_t - - 55 - - 110
This test shows that glass.bmp (no _t) uses the alpha channel
effectively to control transparency.
It also shows that specular setting of 37 operates equally well to
yield sun glare with both techniques.
So, what is the glass_t technique doing differently?
The effect of using the _t and darkening the alpha channel seems to add
more reflective glossiness. Check Holiday Gift Guides and Michaels Ad.In doing that, it effectively decreases
the appearance of transparency as if the glass is becoming more
mirror-like as the alpha channel goes darker. It's an interesting
effect depending on your needs. However, you can retain much of that
mirror-like look while increasing transparency if you go back to gmax
and reduce opacity to say, 40% instead of 50%.
I also found that makemdl export option noted above ("HasReflectMap")
have some bearing on these results.
By unchecking that option, it eliminates the glossy reflection; the alpha channel is then used to
control the transparency beyond the 50% opacity stated in gmax. I
recommend that you leave the reflect option turned on.
In summary, the glass.bmp using the alpha channel to control
transparency is the effective and recommended basic approach.
However if your project requires a glossy-like or mirror-like, transparent
window or surface, you can use the glass_t.bmp approach in conjunction
with a lowered opacity setting for the material in gmax.
Keep in mind that a glass_t.bmp (or any use of _t) with 100% opacity in
gmax and a dark alpha channel will get you chrome.
I hope this test clears up some things, and helps spur some thoughts
and experimentation for you.
I am including the gmax source and bmps for you should you wish to run your own tests.