#Map problem in net rend
5 messages in this thread
Bob:
The master/slave configurations seem to be consistent. I render to the same
file type (TGA), using the same bitmaps, and have tried rendering both to
fields and full frames. I have also put the master machine in the slave mode
and got the same type of problem (again, if I render from the screen on the
same master machine, the image looks OK).
Within one hour, I'm going to upload the two images–titled "Map in Net Rend 1"
and "Map in Net Rend 2"–to the H/W section of the MM forum to give you more
info.
Thanks for help. PD.
Paul:
Upload your images to the ASHOWCASE forum if you would as the space here in
AMMEDIA is running low. I think your responses are starting to point to bitmap
resolution although this doesn't necessarily explain the difference in images
which you are getting.
When I setup my network render slaves I run a little batch file which copies
all of the pertinent 3DS files including the 3DS.SET file and the 3DSNET.SET
file from my file server. I then edit the 3DSNET.SET file manually to give the
node a unique address. Anyway, this automatically gives me all identical
3DS.SET variables. I mention this as a way to almost guarantee that it is not
an errant variable in one of your machines 3DS.SET or 3DSNET.SET files.
There is going to be a time delay as Anne will have to merge the images you
will upload but I'll check them out when they are available.
– Bob Ritger
Bob –
NR1-1050.GIF and NR2-1050.GIF are released into the Hardware Library. I'll
leave them there for a couple of weeks since they are an active topic of
discussion here and then move them to ASHOW or perhaps Paul can delete them (or
tell me to) when the problem is solved.
Anne
Dear Bob:
Sorry about the duplicate message–I'm somewhat new to this process.
To answer some of your questions: the ground object material is tiled over a
large area. The bitmap resolution, while not extremely high, should be enough.
It looks OK if rendered from the screen on the master machine but blurry when
rendered in the slave mode even on the same master machine. This, I think,
rules out the configuration inconsistency scenario.
One thing which comes to mind is that the original model has been imported from
AutoCAD where it had been modeled in milimeters. To aid in makin changes, I
had to set the units to: DECIMAL/IN/1=1000. In effect the original objects
were much larger, so to speak, than those that I added in 3DS, one of which is
the ground object that has the map problem. (I don't think this is a helpful
lead.)
One other thing, which may be related to this problem somehow: I noticed that
network renderings from any viewport other than the camera come out
incorrectly–the whole picture looks zoomed in with only a small portion of the
original object(s) visible. This is not as serious a problem as that with the
map since I don't render entire animations from non-camera viewports but
perhaps it will suggest some related configuration error to you?
Thanks. PD.
PS: The two images are now in HW secition–hope the resolution is good enough.
Also, I could upload the 3ds.set and 3dsnet.set files if you wish.
Paul:
I checked out your images, very strange indeed! However, I think I can narrow
down your problem:
It appears to me that there is a layer of semi-transparent glass between the
viewer and the site's bitmap, true? It could be that the transparency settings
are causing your problems. If I have guessed right about the glass then you
should see identical images if you render the same views without the glass.
The bitmap that you have used should not exhibit any problems due to resolution
until you reach a very acute angle and then only if it is low-res. Your posted
views show that resolution can be discounted as a cause. There is a strip of
glass missing in your images towards the right of the image. Through this gap
the bitmapped areas appear identical in both images. This leads me to believe
that it is the semi-transparent glass causing the problem. Check into this and
let me know what you find out.
Wild structure by the way! Great model, where is the project?
– Bob Ritger