field problem in 3dsr4
26 messages in this thread
Hi anyone. I'm having a strange problem here with my field-rendering
in 3dsr4. As I render (from 0-300 frames), the field problem gets
progressively noticeable. For example, by the time it is rendering frame 100,
one field is correctly rendering frame 100 while the other field looks like
it's rendering frame 50 or something! Anyhow, it's really off. The night
before I rendered another 300-frame animation in field-mode perfectly fine.
But this is another scene I was working on.. by the way, the camera motion in
this "problem-scene" was MERGED (imported) from another mesh.. Maybe does
merging/camera animations have something to do with messing up the
field-rendering? Just a though. Also, if I try to render the frame (for
example frame 100) individually (Single), it renders fine! Hmm!
I _just_ received r4 so I'm not sure if it's the new release or
something that I'm just not doing right. I've been doing field rendering
pretty heavily with no problems in r3 so I'm pretty sure I'm doing everything
right. Thanks for any help.. this is getting frustrating. I'm aware of the
new Keyframe-Field feature and it is OFF currently.
— Steve
If this is all happening on one machine (not across a network), it would be the
first time we've ever heard of field-order changing in mid-stream. What kind of
video output system are you using? Targa->analog video? PAR? Abekas? Is this
a reproducible problem or is it random?
– G
>> If this is all happening on one machine (not across a network), it would be
the
>> first time we've ever heard of field-order changing in mid-stream. What kind
of
>> video output system are you using? Targa->analog video? PAR? Abekas? Is
>> this a reproducible problem or is it random?
Gary, I'm not sure how to describe the problem well enough to you. I have been
rendering over the network but the problem can be reproduced on one machine.
When you say "field-order" changing that means going from FIELD-ORDER=1 to 0 or
something right? Well, I'm pretty sure if that's what you mean, this is NOT a
field-order changing problem. The fields appear to be SEVERAL frames apart!
One field is in frame 100 while the other is in frame 50! BTW, I'm using a
Targa 64+ card and that's it. I've had zero problems with fields in 3ds3
(except for that flat-reflection problem, but at least that was known). It
seems pretty hopeless. I'm not sure if uploading the mesh will do anything for
you because Murphy seems to be saying that it will work fine over there. But
if you want I can certainly try doing that too.
Here's the problem outlined one more time:
– I render starting from frame 0-300
– The first frame is perfectly okay.
– as we progress in frames, the field difference becomes larger and larger
– until about frame 100 we are at field differences of about 50 frames! (I
mean, it's pretty obvious)
– when I go to investigate further, I try to render a single frame at 100 ..
and it does the same field problem thing
– but the thing is when I scale my time segment from 300 to say, 600.. then
scale it back down to 300.. and try to render a single frame and it appears to
WORK!
– so, last night I render all 300 frames thinking that it works after the
scaling, and I wake up this morning and look at frame 100 and it's screwed up
again!
here's the killer:
– this was all working fine in the previous mesh I was working on just a few
days ago (using 3dsr4 as well and ALL the same 3ds.set, etc settings).. the
ONLY difference is that I IMPORTED (File/Merge) the camera animation, the
camera object, the lights, and some other things.. I remember there being a
"problem" with the File/Merge on another occasion in a previous 3ds release..
maybe there is still a problem?
— Steve, thanks ever so much.. if you can solve this problem for me.. thanks
even more!
Hi Steve,
<< – until about frame 100 we are at field differences of about 50 frames! (I
mean, it's pretty obvious)
– when I go to investigate further, I try to render a single frame at 100
.. and it does the same field problem thing>>
I would like to see a flic demonstrating this problem. Can you upload one to
the libs with "For Jonas" in the Keywords please?
<< previous merge issue.. >>
We would need a reproducable case. From your message it seems that Murphy
rendered the *same* file and it did not have this problem? Have you
tested/duplicated this behaviour on any other machine?
jonas[adesk]
>> I would like to see a flic demonstrating this problem. Can you upload one
to the
>> libs with "For Jonas" in the Keywords please?
Sure.. I'll upload a portion of it in a sec..
>> We would need a reproducable case. From your message it seems that Murphy
>> rendered the *same* file and it did not have this problem? Have you
>> tested/duplicated this behaviour on any other machine?
Murphy? I was talking about Murphy as in "Murphy's Law!" Sorry for being so
misleading. Yes, this behaviour happens on all computers here.. I'm beginning
to consign this to a possible bug?
— Steve
Hi Steve,
<< Yes, this behaviour happens on all computers here.. I'm beginning to
consign this to a possible bug? >>
Then upload the 3ds file too… I'd like to see what it does on my machine..
jonas[desk]
Hi Steve,
<< ..upload in a sec.. >>
Had my *wizop* search the upload area and he has not found any uploads yet from
you..
staying tuned..
jonas[adesk]
I'll go ahead and upload the flic.. it's called TABLE.ZIP.. I'm still working
on reproducing the problem with a simple mesh. I'll upload that when I finally
get it to do it again.. but if I can't get it simplified, I'll just send the
entire mesh to you. But in the meantime, at least you can get an idea of
what's happening with the sample flic I uploaded..
The flic is compiled from about 7 tga files. It was rendered over the network
with most of the computers going to NULL. They all have the FIELD-ORDER=1 and
it is a SLOW animation.
— Steve, thanks
FYI, the flic really isn't going to be able to tell Jonas anything useful. You
might as well keep trying to get a 100% reproducible case, simplified as much
as possible, to him in .prj form.
– G
Okay, Gary.. well, it's uploaded anyways. It will at least show what I'm
trying to convey in words. It shows pretty graphically the field differences.
Also, I think I've finally got the problem narrowed down to a reproducible
case! I will upload the mesh as soon as I can. Sorry it's so big but the
problem only seems to "work" with the entire mesh! Also, to reproduce the
problem, render frame 100 in fields mode to TARGA, single. That ought to do it
for ya.. I'll upload it in a sec and call it 3DSTABLE.ZIP
— Steve
Include a very detailed .txt file explaining the steps necessary for Jonas to
reproduce the problem, along with your machine configuration etc.
– G
Oops.. forgot to include the text file.. I'll upload that right away now.
It'll be called 3DSTABLE.txt — I'll also include the "for jonas" keywords..
thanks for reminding me.
— Steve
PMJI, but are you using any IPAS routines like Magic? That could be your
problem…
Dave Mennenoh
Nope, no IPAS's..
— Steve
Did you get a chance to look at that prj file yet? I don't know if you've
noticed, but I also uploaded a text file (search for "for jonas") describing
the problem exactly. I've done another mesh since the last time we talked and
I'm getting the exact same results. It seems to only happen when I have a
scene with auto-reflection maps (both flat and the other)..
anything else you need for this problem?
— Steve
Hi Steve,
<< Did you get a chance to look at that prj file yet? >>
I'm sorry but I could not get to it late thursday. I'm currently using a PC
that does not have 3DS loaded. I'll be back at my machine on the 29th. and I'll
down load it then.
jonas[adesk]
okay, I'll be waiting for your reply..
— Steve, thanks
Have you been able to take a look at the mesh I sent you yet? Just in case you
forgot, it's the field-rendering problem in r4..
— Steve
Hi Steve,
My machine has been taken apart at the request of my wife so the
carpets can be shampooed and my computer room must be moved from one
end of the house to the other. I'm hoping to have these chores
finished by evening so I can rebuild and then I can check out all of
the project files in the libs waiting my persusal.
Thanks for the patience. It's funny how my wife thinks that my vacations aren't
for work. 🙂
jonas[adesk]
Understandable.. <g>.. I'll be patient for a few more days.. All I can hope is
that the problem does indeed happen with that mesh I sent you on your machine!
— Steve
First, I really don't think that we're looking at a bug here. We didn't change
ANY of the frame/field interpolation code between r3 and r4, and although there
was an OLD File/Merge bug in r2, that was fixed in r3 over a year ago (and
we've had no report of any problems besides yours).
Please provide Jonas with a simplified 100% reproducible case in .3ds or .prj
format, and then he'll start tracking it down on his machine. We'll take it
from there.
– G
>>the first time
I did report a couple o' weeks ago I had a problem w/field switching in r4 on a
single machine; it was set up on network render (one machine), escaped out of,
and cleaned/restarted. At the restart point the field order switched. Hasn't
happened again, and have no clue as to why. Was sorry it happened, though; it
was in a sequence using flame, so you can imagine the rerender time ;^).
John Stetzer
JWS
Field switching could possibly happen with an analog (Targa-type) device, but I
can't see any way for it to happen on a digital device. Plus, that doesn't
sound like the problem that this guy was having… it wasn't a field switching
problem — it was a completely wrong frame getting placed at the wrong point.
– G
>>analog (Targa-type) device
PAR
>> that doesn't sound like the problem that this guy was having
Upon rechecking the message, your right; next time I'll just have to be a
little more thorough in my reading. Sorry.
John Stetzer
JWS
are you using a par? software ver 34b or 30b if so its the par not
3ds. ther is a field problem with the par , it reverses the fields.
sherwood
>>software ver 34b or 30b
Yes, it on the PAR, but I'm using 1.27 (I think). Hasn't happened again, so I
count myself as fortunate :^).
John Stetzer
JWS