-
Which the release of FS2020 we see an explosition of activity on the forun and of course we are very happy to see this. But having all questions about FS2020 in one forum becomes a bit messy. So therefore we would like to ask you all to use the following guidelines when posting your questions:
- Tag FS2020 specific questions with the MSFS2020 tag.
- Questions about making 3D assets can be posted in the 3D asset design forum. Either post them in the subforum of the modelling tool you use or in the general forum if they are general.
- Questions about aircraft design can be posted in the Aircraft design forum
- Questions about airport design can be posted in the FS2020 airport design forum. Once airport development tools have been updated for FS2020 you can post tool speciifc questions in the subforums of those tools as well of course.
- Questions about terrain design can be posted in the FS2020 terrain design forum.
- Questions about SimConnect can be posted in the SimConnect forum.
Any other question that is not specific to an aspect of development or tool can be posted in the General chat forum.
By following these guidelines we make sure that the forums remain easy to read for everybody and also that the right people can find your post to answer it.
Still learning
- Thread starter Francois
- Start date
Francois
FSDevConf team
- Messages
- 280
- Country
Hmmm.... some more news on this project.
I now shot screenies of the 'finished' render of the flying school model, and started applying them on a 'gray' model. So taking the original textures as a photo of an entire side.
(Re) applying that 1 photo means having to break it up in pieces and apply to the various faces again.
What I found is that the resulted render is very fuzzy and by long not as crisp as the original textures applied. And yes, I used all png format, not jpg loss.
So I am not sure about this method...... the regular 'original' textures applied on the initial model look much better! Yes, I will need to take all those windows out and apply a photo on them instead...... hmmmmmm...
This is the originally rendered model:
And this is the resulting screenshot of it (one side)...
Hmmmm... I'll need to change the lighting to begin with. It is too dark indeed. I still like method '1' better. Maybe I am doing something wrong, but it looks as if the screenshot lacks detail.
I now shot screenies of the 'finished' render of the flying school model, and started applying them on a 'gray' model. So taking the original textures as a photo of an entire side.
(Re) applying that 1 photo means having to break it up in pieces and apply to the various faces again.
What I found is that the resulted render is very fuzzy and by long not as crisp as the original textures applied. And yes, I used all png format, not jpg loss.
So I am not sure about this method...... the regular 'original' textures applied on the initial model look much better! Yes, I will need to take all those windows out and apply a photo on them instead...... hmmmmmm...
This is the originally rendered model:
And this is the resulting screenshot of it (one side)...
Hmmmm... I'll need to change the lighting to begin with. It is too dark indeed. I still like method '1' better. Maybe I am doing something wrong, but it looks as if the screenshot lacks detail.
[EDITED]
Looks good, Francois !
Admittedly, it seems likely that one might notice a change in 'detail' or 'sharpness' after the textures are "customized" by a rendering engine.
Also, a reduction of 'detail' or 'sharpness' is likely to occur if the original larger total linear 'span' of individual mapped/tiled texture pixel information is distributed over a smaller span of texture image size in pixels... as information on an entire side of a building is being written out via either of the "export a 2D Graphic image file from Sketchup" routines I described above.
[END_EDIT]
BTW: One might try the latter 'alternate' routine allowing one to to "export a 2D Graphic image file from Sketchup that is much larger than one's Windows Desktop display resolution setting" at:
http://www.fsdeveloper.com/forum/showpost.php?p=210002&postcount=45
FYI: It is possible that some computer hardware configurations with more system RAM and VRAM may allow even higher Windows Desktop display resolution settings; this may allow Sketchup to initially submit larger pixel spans in the:
Sketchup Menu > File > Export 2D Graphic > Export Image Options > "Use view size" = 'checked' / Image Size - Height = "#" of pixels dialog box
With such very large Windows Desktop resolutions (and with all Sketchup Toolbars "hidden" to allow a larger workspace area), the submitted initial 'Image Size - Height = "#" pixels' value can then be maximized before multiplying it by "4" to save into a bigger image file.
That image file can then be separately processed with the (more precise) re-sample routines in an external graphics utility to minimize loss of detail in one's intended texture.
Also, in a external graphics utility, "rendered" textures can then be sharpened, and contrast/brightness/gamma/saturation/hue/intensity etc. all adjusted.
[EDITED]
And as described previously, Sketchup offers the option to 'auto-update' custom imported (or copied 'default') texture image materials after saved interactive graphical edits, photoreal or other special rendering etc. are applied to one's mapped file:
http://www.sketchupartists.org/tuto...p/use-image-editing-software-inside-sketchup/
[END_EDIT]
Remembering that there are pros and cons to having a custom scenery object appear "too sharply defined" compared to other FS scenery objects, one might actually see a greater measure of "realism" in an object that looks slightly "less well defined"... as it allows a greater perception of 'distance' and naturally-occurring extents of 'atmospheric haze' to be visualized when viewing one's scenery object in FS at run time.
NOTE: A zoomed-in examination of default and 3rd party textures in a graphics utility quickly reveals that most FS textures are <purposely ?> not super-sharp... and yet they still blend into the overall FS scenery well.
[EDITED]
PS:
Although I suppose one could subdivide ones final 'rendered composite texture file' into individual (ex: square 64x64, 128x128, 256x256, 512x512, or 1024x1024 pixel) texture file "segments", IIUC, one would achieve better performance with a somewhat more complex model having numerous vertices at various sloped angles to render the multiple Roofs and Dormers etc. by mapping the same 1-piece texture material file onto the faces of a particular side of the model.
Ones goal (wherever possible), IMHO, is to reduce the total aggregate of draw calls from all scenery objects in areas adjacent to low-altitude / on-ground activity at an aerodrome, by using ex: a 'singe texture file' a minimum number of times on the face(s) of a building 'side' being textured.
Thus, MCX can apply the "batched texture draw" function attribute (aka "DrawcallBatching") to a models texture mapping... so that at run time in FS, it is rendered efficiently and file I/O is minimized. :spushpin:
This makes for better overall performance in FS, and the reduced FPS hit 'savings' allows "wiggle room" in ones total demand on the FS rendering engine for 'expenditures' on other things like complex aircraft / other scenery objects / detailed land class and/or autogen objects / AI traffic / animations / clouds etc.
"Drawcalls
For the performance of an object in Flight Simulator the amount of drawcalls used by the object is an important factor. ModelConverterX has a tool to minimize the amount of drawcalls. But what is a drawcall?
Simply said a drawcall is a group of triangles that use the same material settings (colours and textures) when the graphical card draws your object. Every change in a material setting means that the graphics card has to set its state again and that is a relatively expensive operations for the performance.
So to optimize performance you can best make sure you have as little different materials as possible. Even one small parameter that is different in the material results in a new drawcall already. Another why to reduce them is to not use colours on your triangles, instead give a piece of unused texture that colour and map it on the triangle."
"DrawcallBatching
Determines whether the object will have drawcall batching (or) working LODs."
http://www.fsdeveloper.com/wiki/index.php?title=ModelConverterX#Drawcalls
http://www.fsdeveloper.com/wiki/index.php?title=ModelConverterX
"DrawcallBatching specifies whether drawcall batching or LODs have preference when exporting. When this value is true drawcall batching will also work for objects with LODs (but as a downside the LODs no longer work). With the default value of false you get drawcall batching for objects without LODs, but for objects with LODs the LODs still work."
http://www.fsdeveloper.com/forum/showthread.php?t=64006&highlight=DrawcallBatching
"Models with LODs will still not batch drawcalls, if you want that you need to set the DrawcallBatching property in the ModelConverterX exporter settings to true. In that case drawcall batching has priority over LODs (but the LODs of your object will no longer work). "
http://www.fsdeveloper.com/forum/showthread.php?t=63392&highlight=DrawcallBatching
[END_EDIT]
GaryGB
Looks good, Francois !
Admittedly, it seems likely that one might notice a change in 'detail' or 'sharpness' after the textures are "customized" by a rendering engine.
Also, a reduction of 'detail' or 'sharpness' is likely to occur if the original larger total linear 'span' of individual mapped/tiled texture pixel information is distributed over a smaller span of texture image size in pixels... as information on an entire side of a building is being written out via either of the "export a 2D Graphic image file from Sketchup" routines I described above.
[END_EDIT]
BTW: One might try the latter 'alternate' routine allowing one to to "export a 2D Graphic image file from Sketchup that is much larger than one's Windows Desktop display resolution setting" at:
http://www.fsdeveloper.com/forum/showpost.php?p=210002&postcount=45
FYI: It is possible that some computer hardware configurations with more system RAM and VRAM may allow even higher Windows Desktop display resolution settings; this may allow Sketchup to initially submit larger pixel spans in the:
Sketchup Menu > File > Export 2D Graphic > Export Image Options > "Use view size" = 'checked' / Image Size - Height = "#" of pixels dialog box
With such very large Windows Desktop resolutions (and with all Sketchup Toolbars "hidden" to allow a larger workspace area), the submitted initial 'Image Size - Height = "#" pixels' value can then be maximized before multiplying it by "4" to save into a bigger image file.
That image file can then be separately processed with the (more precise) re-sample routines in an external graphics utility to minimize loss of detail in one's intended texture.
Also, in a external graphics utility, "rendered" textures can then be sharpened, and contrast/brightness/gamma/saturation/hue/intensity etc. all adjusted.
[EDITED]
And as described previously, Sketchup offers the option to 'auto-update' custom imported (or copied 'default') texture image materials after saved interactive graphical edits, photoreal or other special rendering etc. are applied to one's mapped file:
http://www.sketchupartists.org/tuto...p/use-image-editing-software-inside-sketchup/
[END_EDIT]
Remembering that there are pros and cons to having a custom scenery object appear "too sharply defined" compared to other FS scenery objects, one might actually see a greater measure of "realism" in an object that looks slightly "less well defined"... as it allows a greater perception of 'distance' and naturally-occurring extents of 'atmospheric haze' to be visualized when viewing one's scenery object in FS at run time.
NOTE: A zoomed-in examination of default and 3rd party textures in a graphics utility quickly reveals that most FS textures are <purposely ?> not super-sharp... and yet they still blend into the overall FS scenery well.
[EDITED]
PS:
Although I suppose one could subdivide ones final 'rendered composite texture file' into individual (ex: square 64x64, 128x128, 256x256, 512x512, or 1024x1024 pixel) texture file "segments", IIUC, one would achieve better performance with a somewhat more complex model having numerous vertices at various sloped angles to render the multiple Roofs and Dormers etc. by mapping the same 1-piece texture material file onto the faces of a particular side of the model.
Ones goal (wherever possible), IMHO, is to reduce the total aggregate of draw calls from all scenery objects in areas adjacent to low-altitude / on-ground activity at an aerodrome, by using ex: a 'singe texture file' a minimum number of times on the face(s) of a building 'side' being textured.
Thus, MCX can apply the "batched texture draw" function attribute (aka "DrawcallBatching") to a models texture mapping... so that at run time in FS, it is rendered efficiently and file I/O is minimized. :spushpin:
This makes for better overall performance in FS, and the reduced FPS hit 'savings' allows "wiggle room" in ones total demand on the FS rendering engine for 'expenditures' on other things like complex aircraft / other scenery objects / detailed land class and/or autogen objects / AI traffic / animations / clouds etc.
"Drawcalls
For the performance of an object in Flight Simulator the amount of drawcalls used by the object is an important factor. ModelConverterX has a tool to minimize the amount of drawcalls. But what is a drawcall?
Simply said a drawcall is a group of triangles that use the same material settings (colours and textures) when the graphical card draws your object. Every change in a material setting means that the graphics card has to set its state again and that is a relatively expensive operations for the performance.
So to optimize performance you can best make sure you have as little different materials as possible. Even one small parameter that is different in the material results in a new drawcall already. Another why to reduce them is to not use colours on your triangles, instead give a piece of unused texture that colour and map it on the triangle."
"DrawcallBatching
Determines whether the object will have drawcall batching (or) working LODs."
http://www.fsdeveloper.com/wiki/index.php?title=ModelConverterX#Drawcalls
http://www.fsdeveloper.com/wiki/index.php?title=ModelConverterX
"DrawcallBatching specifies whether drawcall batching or LODs have preference when exporting. When this value is true drawcall batching will also work for objects with LODs (but as a downside the LODs no longer work). With the default value of false you get drawcall batching for objects without LODs, but for objects with LODs the LODs still work."
http://www.fsdeveloper.com/forum/showthread.php?t=64006&highlight=DrawcallBatching
"Models with LODs will still not batch drawcalls, if you want that you need to set the DrawcallBatching property in the ModelConverterX exporter settings to true. In that case drawcall batching has priority over LODs (but the LODs of your object will no longer work). "
http://www.fsdeveloper.com/forum/showthread.php?t=63392&highlight=DrawcallBatching
[END_EDIT]
GaryGB
Last edited:
- Messages
- 5,214
Hi,
I am afraid, François, that you have become addicted to this Sketchup thing, as I have.
I am working on a ferry and noticed the following: whatever you do in sketchup, the conversion to FSX is a major "pita" regarding textures. MCX still has some flaws but that might not be the fault of MCX but the way you transform those textures to FSX.
You use a jpeg for showing things in GSU, I suppose and you convert those jpegs to bmp's or to dds textures. But resizing them and converting them makes you lose detail and changes the colors. If on an object you have several textures of different sizes, converting them will give you different shadings dependent on the initial size of the texture.You will have to give them all the same kind of compression and rework the textures after conversion.
Just so that you know it because that might also be what happened when rendering your GSU textures to FSX via MCX.
I am afraid, François, that you have become addicted to this Sketchup thing, as I have.
I am working on a ferry and noticed the following: whatever you do in sketchup, the conversion to FSX is a major "pita" regarding textures. MCX still has some flaws but that might not be the fault of MCX but the way you transform those textures to FSX.
You use a jpeg for showing things in GSU, I suppose and you convert those jpegs to bmp's or to dds textures. But resizing them and converting them makes you lose detail and changes the colors. If on an object you have several textures of different sizes, converting them will give you different shadings dependent on the initial size of the texture.You will have to give them all the same kind of compression and rework the textures after conversion.
Just so that you know it because that might also be what happened when rendering your GSU textures to FSX via MCX.
- Messages
- 34,350
- Country
Hi,
Resizing textures will of course affect their resolution and that will affect the quality of the texture a bit as well. But in general you will only resize them a bit. And a developer who knows he is modelling for FS, will make sure the textures he uses in SketchUp are the right sizes for FS (so no resizing is needed).
Converting to DDS will affect the quality a little as well, but the compression does not make a big difference usually. I would almost dare to say that the original compression on the JPG is likely to have more effect than the DDS compression. Colours will change a little indeed, but you won't see it if you R, G or B value is 1 off.
I think if you start modelling knowing what FS needs, just like you would do in GMax, FSDS or any other tool, there are not so many issues with converting SketchUp output to FS.
Resizing textures will of course affect their resolution and that will affect the quality of the texture a bit as well. But in general you will only resize them a bit. And a developer who knows he is modelling for FS, will make sure the textures he uses in SketchUp are the right sizes for FS (so no resizing is needed).
Converting to DDS will affect the quality a little as well, but the compression does not make a big difference usually. I would almost dare to say that the original compression on the JPG is likely to have more effect than the DDS compression. Colours will change a little indeed, but you won't see it if you R, G or B value is 1 off.
I think if you start modelling knowing what FS needs, just like you would do in GMax, FSDS or any other tool, there are not so many issues with converting SketchUp output to FS.
Francois
FSDevConf team
- Messages
- 280
- Country
As the initial title says: still learning
I uses my own photos so far. They are RAW files (Nikon NEF format), which I change into .png. I then cut the parts I need on 1023 or 512 size (for the larger areas), all still in .png so as not to lose anything through conversion to jpg.
So once I have made 1024x1024 tiles with my png textures and painted the GSU model, I use export to make a dae file for MCX.
Then WHAT do I do? It converts the textures to dds for me, right?
I uses my own photos so far. They are RAW files (Nikon NEF format), which I change into .png. I then cut the parts I need on 1023 or 512 size (for the larger areas), all still in .png so as not to lose anything through conversion to jpg.
So once I have made 1024x1024 tiles with my png textures and painted the GSU model, I use export to make a dae file for MCX.
Then WHAT do I do? It converts the textures to dds for me, right?
- Messages
- 5,214
Shoo Arno,
You are supposed to be on your holiday or I am misinformed!
I used photoreal textures to start with but I needed many and not all of them with the same resolution. So I used those in GSU just to know what I was doing and then adapt, resize and convert the jpegs/bmp's to dds and this is what happens afterwards: (see attachment).
So I have to edit the dds again and match the colors once more to the (in this case) night colors used for the other textures.
PS and by the way: the rump of the ship should be almost black (because the night textures are black 7.7.10) but they are not.
You are supposed to be on your holiday or I am misinformed!
I used photoreal textures to start with but I needed many and not all of them with the same resolution. So I used those in GSU just to know what I was doing and then adapt, resize and convert the jpegs/bmp's to dds and this is what happens afterwards: (see attachment).
So I have to edit the dds again and match the colors once more to the (in this case) night colors used for the other textures.
PS and by the way: the rump of the ship should be almost black (because the night textures are black 7.7.10) but they are not.
Last edited:
Hi Francois:
Since IMHO there has been a recurring topic here in the Sketchup forum regarding ways to use 'multiple textures' with models by texturing 'multiple faces', I thought it might be interesting to look at some options for applying textures (and/or make and 're-apply' composite textures) to multiple faces all at once.
I was also curious for my own projects as to what info might be out there to even better facilitate this workflow; so I searched and found a few links that explain some 'tricks and traps' with the process:
Can I import a photographic texture to multiple faces at one time?
http://www.google.com/support/forum/p/sketchup/thread?tid=1ff1f0390efa2733&hl=en
Make unique texture from multiple faces
http://forums.sketchucation.com/viewtopic.php?f=40&t=24519
SketchUp / Texturing > Adding Colors and Textures > Using Shift and Ctrl keys
http://courses.washington.edu/udpddp/exercises/tu52.shtml
...And for any pipes yet to be <tu-be ?> textured:
Wrapping an Image Around a Cylinder
http://www.youtube.com/watch?v=ILnIbp509cY&feature=related
"Pipe Tool" (a freeware plug-in alternative to payware "3SKENG"; currently there is no manual included, so please note the following):
"You need to put the file "Pipe Tool v4.0.rb" along with the .png files in the ..\plugins folder. You cannot put the folder in the ..\plugins directory."
http://forums.sketchucation.com/viewtopic.php?t=34361
Hope this info might be of help with the Sketchup multi-face texturing process !
PS: Oh, what the heck...while we're at it, lets have some fun seeing how fast "some" payware plug-ins can make the modeling process (play the video) !
http://www.youtube.com/watch?v=FcpGwlfu3vU
http://www.smustard.com/script/ProfileBuilder
GaryGB
Last edited:
I create all my textures in GIMP, generally 32 bit bmp. I import into GSU as texture and use GSU to get the UVs set. After importing my GSU model into MCX, I just replace all the GSU textures with my originals in the material editor. Then use the mass texture converter to switch them to dds. As long as I don't skew the texture when positioning it on a face in GSU, the UVs will be correct for my master textures.
I haven't yet tried using image from a render as a material, but I am playing with it for night texture. Problem is I used a tiled texture on some faces so I can't just replace it if I want the night lighting not to tile.
scott s.
.
I haven't yet tried using image from a render as a material, but I am playing with it for night texture. Problem is I used a tiled texture on some faces so I can't just replace it if I want the night lighting not to tile.
scott s.
.
Francois
FSDevConf team
- Messages
- 280
- Country
PS: Oh, what the heck...while we're at it, lets have some fun seeing how fast "some" payware plug-ins can make the modeling process (play the video) !
http://www.smustard.com/script/ProfileBuilder
GaryGB
NOT FAIR !!!!!
But that's not being used for FSX, is it ? *grin*
NOT FAIR !!!!!
But that's not being used for FSX, is it ? *grin*
Most of the examples shown were, of course, so complex they would be FPS-killing "vertex nightmares" if used in FSX.
But it's fun to see how certain tasks for creating less complex models might be made easier.
GaryGB
Last edited: