Found 42 posts.
Search results Show results as topic list.
Technical Discussion » XSI 2 Houdini
- jrwillette
- 49 posts
- Offline
Who would have know that camera would be so popular… I know what you mean though. I used one that Denis Gauthier at a52 had created while I was working there and it was great, when I left it was so frustrating so I finally had to take the time to make one (with a few extra features of course). One of the first things I did everywhere I worked was install that cam file, so I figured this way maybe it would already be there when I arrive at my next gig
Technical Discussion » "Auto-Link" with HScript?
- jrwillette
- 49 posts
- Offline
hoknamahnYes, this is unfortunate. It would be great if there were a way to lock those parameters, It is very frustrating to find out that you've broken x amount of toggle references because you clicked the toggles without realizing it would delete the channel reference. This is also related to the fact that you cannot keyframe a toggle… this would make a good RFE.
And bad too that if the parameter, having reference “!!ch()!!” to switch by hand, that it this reference
Technical Discussion » XSI 2 Houdini
- jrwillette
- 49 posts
- Offline
Do any converters spit out channels that preserve keyframing?
Every one I've ever seen do raw channel dumps. Usually it doesn't matter to me since I'm never changing the animation that comes into Houdini so I've never really thought to look for a better solution. I always figured that each program was using slightly (or greatly) different math to define channels so it would be very difficult to get an exact match.
Every one I've ever seen do raw channel dumps. Usually it doesn't matter to me since I'm never changing the animation that comes into Houdini so I've never really thought to look for a better solution. I always figured that each program was using slightly (or greatly) different math to define channels so it would be very difficult to get an exact match.
Technical Discussion » XSI 2 Houdini
- jrwillette
- 49 posts
- Offline
If XSI can export RIB you can use Martian Labs [martian-labs.com] GLUE to import it into Houdini.
At one time I had written a conversion utility (in awk) that went from Lightwave to Prisms… but I don't think that would do you any good.
At one time I had written a conversion utility (in awk) that went from Lightwave to Prisms… but I don't think that would do you any good.
Technical Discussion » "Auto-Link" with HScript?
- jrwillette
- 49 posts
- Offline
I think that method doesn't work with toggles because you can't keyframe a toggle, have you tried opparm?
This is where the syntax always messes with me ( I don't know what !! means but it seems to be very important), let me know if this workspparm /obj/myobj mytoggle ( ‘!!ch(“mychannel”)!!’ )
I haven't tested this out, but that is what opscript spits back for a toggle that has been previously Auto-Linked from using Operator Type Properties.
In your example above, it didn't look like you were trying to Auto-Link a toggle, so I assumed it was a parameter that could be keyframed.
This is where the syntax always messes with me ( I don't know what !! means but it seems to be very important), let me know if this workspparm /obj/myobj mytoggle ( ‘!!ch(“mychannel”)!!’ )
I haven't tested this out, but that is what opscript spits back for a toggle that has been previously Auto-Linked from using Operator Type Properties.
In your example above, it didn't look like you were trying to Auto-Link a toggle, so I assumed it was a parameter that could be keyframed.
Technical Discussion » "Auto-Link" with HScript?
- jrwillette
- 49 posts
- Offline
I don't think you can do an “Auto-Link”, but you can script it. You have to add a channel then a key:
chadd -t 0 0 /obj/cam1/tx
chkey -t 0 -F ‘ch(“scale”)’ /obj/cam1/tx
This would add a channel to your tx parameter of your camera, and channel reference it's own scale parameter (which I do all the time :wink: ). Of course any relative or absolute channel reference will work (or any other expression).
chadd -t 0 0 /obj/cam1/tx
chkey -t 0 -F ‘ch(“scale”)’ /obj/cam1/tx
This would add a channel to your tx parameter of your camera, and channel reference it's own scale parameter (which I do all the time :wink: ). Of course any relative or absolute channel reference will work (or any other expression).
Houdini Lounge » Saving geometry with colors and texture
- jrwillette
- 49 posts
- Offline
jeffI don't know if the .obj format supports point colors, but the Houdini implimentation of it does not, I think it uses the .mtl files for colors. See my post above.
obj supports point colors and texture uv's so…
I tried saving a .obj with colored points then reading it back in… no colors.
Houdini Lounge » Camera Settings - Help needed
- jrwillette
- 49 posts
- Offline
JohnnyBGoodeNo problem, I learned a lot of that making the Camera Creation Script. Glad to help.
Thanks so much for solving so many issues I have been having!!
So, just to check I've got this right - as it was a 2/3 camera, I assume if I put 19.9333333 into the apeture parameter, then tap the focal length I was shooting at (for instance 120mm) into the focal lenth parameter it should match up?
Yes, it looks like you got it. In theory it should match up, but there are many other factors involved; the accuracy of the CCD (the image may only take up a portion of the CCD so the 2/3" might not be correct), Lens curvature, among other things. It should get you close though.
keltuzarSadly, I cannot take credit for those handles, that is all SESI. Many of the stuff that my camera network does can be seen while interacting with the camera in the viewport. If you right click on the camera in the viewport you can turn on frustum and focal length, and other cool stuff. My camera just allows you to see those things all the time, but not interact with them, it's all tied to the exsisting controls on the default camera.
Btw how did u achieve to have whose fancy little handles when u look thru your cam?
JohnnyBGoodeNot sure how to do that, I think you can add custom items to the defaults menu, but that would be about it.
Maybe you could add something to it that would help with the kind of ‘real-world’ lens issues i've been having?
I guess one could make a Digital Asset that encompassed a camera, then you could add a whole slew of things. That is a project for another day.
stevenongThanks for the reassuring BTW: forgot you thank you for welcoming me back to the Houdini Community on another thread. I never left Houdini, but I just dropped of the list for a while… Now that I'm working freelance I figured it was a good way to get my name out there again
I'm on the same page as Jeff. I've always rendered everything with the same aspect ratio as the plate.
Houdini Lounge » Camera Settings - Help needed
- jrwillette
- 49 posts
- Offline
ops: Sorry, if I came over a little strong Dave, I pulled an all-nighter last night and am a little on edge. Just re-read my post, seemed a little harsh, not my intention.
Always glad to offer a little insight though. I worked with a guy for several years that was particularily anal about this kind of thing, used to drive me crazy (love ya Denis), and a lot of it rubbed off on me… I'm sure he'd be proud :'{
I was just talking to someone else on that subject and he said a lot of people render that way… not sure if they know something I don't, but it just don't seem right.
Take Care
Always glad to offer a little insight though. I worked with a guy for several years that was particularily anal about this kind of thing, used to drive me crazy (love ya Denis), and a lot of it rubbed off on me… I'm sure he'd be proud :'{
I was just talking to someone else on that subject and he said a lot of people render that way… not sure if they know something I don't, but it just don't seem right.
Take Care
Houdini Lounge » Camera Settings - Help needed
- jrwillette
- 49 posts
- Offline
Alright! That confirms that my expression works(576/720)*(16/9)=1.422222Thanks!
But… I must disagree with rendering at square pixels when your plate is anamorphic. In order to get the best match to your photography your CG should be treated in the same manner as your source material. Anti-aliasing is going to be different if rendered anamorphic or square pixels, and you want to keep from scale filtering as much as possible to ensure the integrity of your image.
Most of my work has been in the Commercial Industry NTSC .9 aspect ratio, and for over 12 years I've never met anyone who would consider rendering their images with an aspect ratio of 1 then scale it down in Y to match the plate. Also, when I was working at Sony, we rendered anamorphic aspect to match the plate… it is the right thing to do.
But… I must disagree with rendering at square pixels when your plate is anamorphic. In order to get the best match to your photography your CG should be treated in the same manner as your source material. Anti-aliasing is going to be different if rendered anamorphic or square pixels, and you want to keep from scale filtering as much as possible to ensure the integrity of your image.
Most of my work has been in the Commercial Industry NTSC .9 aspect ratio, and for over 12 years I've never met anyone who would consider rendering their images with an aspect ratio of 1 then scale it down in Y to match the plate. Also, when I was working at Sony, we rendered anamorphic aspect to match the plate… it is the right thing to do.
Houdini Lounge » Camera Settings - Help needed
- jrwillette
- 49 posts
- Offline
Hey! thanks for plugging my Camera script
As for your camera issue, you need to plug your film back width into the aperture parameter on your Houdini camera. If your focal length is in millimeters then your aperture has to be as well. Not sure if the 2/3“ is a truely accurate measurment of the film back width, but it may get you close. 2/3” = 19.9333333 millimeters. Look through your camera manual to see if it has any info on the film back width (or something similar to that)… or try to track a shot using a good match moving software and it should spit out the proper values for you.
As for your pixel aspect ratio, it's going to be a ratio between your resolution and your screen aspect ratio, something like576/720)*(16/9)I could totally have the order of the division messed up (I'm funny that way), if it doesn't look right play around with a variation on that.
BTW: Did you know you can get conversions using google? Try the following in a google search.2/3 inch in mm
Also… if you like my Camera Creation Script, rate it up!
As for your camera issue, you need to plug your film back width into the aperture parameter on your Houdini camera. If your focal length is in millimeters then your aperture has to be as well. Not sure if the 2/3“ is a truely accurate measurment of the film back width, but it may get you close. 2/3” = 19.9333333 millimeters. Look through your camera manual to see if it has any info on the film back width (or something similar to that)… or try to track a shot using a good match moving software and it should spit out the proper values for you.
As for your pixel aspect ratio, it's going to be a ratio between your resolution and your screen aspect ratio, something like576/720)*(16/9)I could totally have the order of the division messed up (I'm funny that way), if it doesn't look right play around with a variation on that.
BTW: Did you know you can get conversions using google? Try the following in a google search.2/3 inch in mm
Also… if you like my Camera Creation Script, rate it up!
Houdini Lounge » Saving geometry with colors and texture
- jrwillette
- 49 posts
- Offline
What do you mean? Why would anyone want to export from Houdini to another software package? Isn't it supposed to be the other way around?
Sorry, I don't think .obj format supports colors, their is usually a .mtl file needed that describs the surface attributes for .obj files and Houdini neither reads or writes that file.
Here is what the docs have to say on the subject.
For further information look under Help Formats>Geometry Formats Supported by Houdini
Maybe the information you need is in there, looks like the Lightwave format supports color.
Sorry, I don't think .obj format supports colors, their is usually a .mtl file needed that describs the surface attributes for .obj files and Houdini neither reads or writes that file.
Here is what the docs have to say on the subject.
.obj … Uses gwavefront - polygonsonly. Only texture and pointnormal attribute (groups are kept on read)
For further information look under Help Formats>Geometry Formats Supported by Houdini
Maybe the information you need is in there, looks like the Lightwave format supports color.
Technical Discussion » Kaleidoscope
- jrwillette
- 49 posts
- Offline
BTW: I just had to try the parellel object test, it works better if you put a border on the “mirror” objects so you can see the edge of it. Even with reflect bounc of 10 it took a really long time to render. It really makes you appreciate the speed of light. Light is fast… really fast. The good part is, it doesn't seem to eat up memory like I thought it would, since it is really only accumulating a single value. If there is a limit, you don't want to try to find it.
I also put together a quick Kaleidoscope and it seemed to work just fine with 4-6 bounces, depends on the framing. Make sure that the image/objects you are reflecting are close to the end of the tube or the results will not look right. A good distance is around 1/4 the width of one if the mirrors, that gives a nice tiny bit of parallax without getting too disjointed. If you go too far the parallax is so great that the kaleidoscope loses it's effect.
I also put together a quick Kaleidoscope and it seemed to work just fine with 4-6 bounces, depends on the framing. Make sure that the image/objects you are reflecting are close to the end of the tube or the results will not look right. A good distance is around 1/4 the width of one if the mirrors, that gives a nice tiny bit of parallax without getting too disjointed. If you go too far the parallax is so great that the kaleidoscope loses it's effect.
Technical Discussion » Kaleidoscope
- jrwillette
- 49 posts
- Offline
The reason you are not seeing a change is most likely that all your rays are escaping, so the extra bounces are redeundant and don't add much to the render time (if anything). I'm not sure if there is a limit to it, if you are interested in finding out the limit, or just experimenting, you could place 2 reflective objects parellel to each other with your camera in the middle (pointing directly at one of the surfaces). Ensure that you can see the edges of the objects in the render, perhaps by placing them inside a large constant shaded sphere. Crank up the reflect bounce and get real comfortable, or go on vacation or something. When you get back from vacation (hope you had a great time!), if Houdini hasn't run out of memory, there may or may not be some rendered pixels.
If the render isn't long enough, throw in a bunch of depth of field with the anti-aliasing cranked way up
If the render isn't long enough, throw in a bunch of depth of field with the anti-aliasing cranked way up
Technical Discussion » Bug: Hscript "while" loop structure
- jrwillette
- 49 posts
- Offline
I duplicated your code, no spaces, and ran it in both Hscript and in Houdini's textport and it ran just fine both ways. I know you said it isn't always consistant so it may be difficult to figure out.
Questions/Things to look at:
Are you on Windows or Linux? I ran my test on windows ops:
What text editor are you using? Perhaps there are hidden characters in your code? Try writing it with different editors.
How are you running the script? source myscript.cmd?
If you want, zip/gzip an example that isn't working and mail it to me jeff@mister2.net and I'll take a look and see if I have the same problem.
It's important to zip it up so the email process doesn't mess with the file.
Questions/Things to look at:
Are you on Windows or Linux? I ran my test on windows ops:
What text editor are you using? Perhaps there are hidden characters in your code? Try writing it with different editors.
How are you running the script? source myscript.cmd?
If you want, zip/gzip an example that isn't working and mail it to me jeff@mister2.net and I'll take a look and see if I have the same problem.
It's important to zip it up so the email process doesn't mess with the file.
Houdini Lounge » some xsi-user questions
- jrwillette
- 49 posts
- Offline
I know this thread is about a month old but I just noticed the question about the water in a52's Cadillac Turbulence ad. I worked on that spot, so thought I could clear things up a bit. Yes, it was done in Houdini, but the water effects we did in CG were mainly on the trailing wakes of the cars and shockwaves. A lot of brilliant composite work was done to enhance shots like the Breakthtough and Tire Vortex, and most of the elements when the wake overtakes the car (where the car stops) was compositing. The only time particles were used, from what I can remember, was on the stopping shot, but vary sparingly and that may have been a softbody deformation of the wake overtaking the car. Pretty much any water that breaks from the surface of the wakes and all that cool mid-air slo-mo stuff is 2D.
The wake effects were done with trailed geometry based on the profile of the car, then displaced both in SOPs (vex) and in the render. It was some basic noise operations that evolved both down the length of the wake and in time.
The Shockwave was done via a combination of a vex distortion for the central vortex and some expanding torus shapes (take that real flow!). Leading edge and cresting effects were done in compositing.
That's about the skinny of it.
That spot can still be seen on a52's website [a52.com] by following their “Classics” link and searching for Cadillac.
The wake effects were done with trailed geometry based on the profile of the car, then displaced both in SOPs (vex) and in the render. It was some basic noise operations that evolved both down the length of the wake and in time.
The Shockwave was done via a combination of a vex distortion for the central vortex and some expanding torus shapes (take that real flow!). Leading edge and cresting effects were done in compositing.
That's about the skinny of it.
That spot can still be seen on a52's website [a52.com] by following their “Classics” link and searching for Cadillac.
Technical Discussion » Sliders in a custom panel
- jrwillette
- 49 posts
- Offline
Yes, Digital Assets are much more useful, but I wouldn't go as far to say that Custom Panels shouldn't be used. There are a few cases where custom panels are prefferable to a Digital Asset. Custom panels are a great way to create your own button bar, useful for running scripts or commands that you don't want to tie to a hotkey. You can create several buttons in a row and it takes up much less space than a parameter pane.
Houdini Lounge » Bridging
- jrwillette
- 49 posts
- Offline
Yes, PolyLoft is the way to go.
About the Facet SOP; The reason that Consolidate Fast is good enough, is becuase it is the best choice. It should be “Consolidate Newer Faster” and “Consolidate Older Slower”… basically the “Slow” version is an old method of consolidate included for backward compatibility… apparentley many people were very upset when the new version came out and demanded that they keep in that old version just to confuse everybody into thinking it must be better because it takes longer to compute
It does say something to that effect if you read the Parameter Help… but who the heck does that? I've been using Houdini for years thinking… “Well I better use slow, just in case.” :twisted:
I'm better now.
About the Facet SOP; The reason that Consolidate Fast is good enough, is becuase it is the best choice. It should be “Consolidate Newer Faster” and “Consolidate Older Slower”… basically the “Slow” version is an old method of consolidate included for backward compatibility… apparentley many people were very upset when the new version came out and demanded that they keep in that old version just to confuse everybody into thinking it must be better because it takes longer to compute
It does say something to that effect if you read the Parameter Help… but who the heck does that? I've been using Houdini for years thinking… “Well I better use slow, just in case.” :twisted:
I'm better now.
Houdini Lounge » Braided leather belt
- jrwillette
- 49 posts
- Offline
I would probably model one repeatable section with curves, use a sweep sop to place skinned cross-sections on the curves, ensure that the points at both ends are planer then duplicate it out. Then model the ends. If the position of the ends were based on the MIN and MAX of the belt you could instantly make a belt of any length.
Standard bone or lattice techniques could be used to deform/animate it. You could even creep the geometry onto a surface.
Standard bone or lattice techniques could be used to deform/animate it. You could even creep the geometry onto a surface.
Technical Discussion » Nearest Primitive to Point; expression/technique?
- jrwillette
- 49 posts
- Offline
Nice!
Thanks, I read the help on that before but I guess I just stopped reading after it says “…specified sop's primitive.”
Thank you very much, that's exactly what I needed.
I thought I was going crazy becuase I was sure I had done that before.
Thanks, I read the help on that before but I guess I just stopped reading after it says “…specified sop's primitive.”
Thank you very much, that's exactly what I needed.
I thought I was going crazy becuase I was sure I had done that before.
-
- Quick Links