Singing Potatoes
Friday, 5 September 2003
Technology to the rescue!

A little while ago, I sent in a bug report for some 3D software I use. On Wednesday night, I got an email from one of the programmers saying that he couldn't see the bug I was talking about; was it still happening in the most recent version?

I remembered that hidden on their site was a video tutorial showing how to create hair braids; they demonstrated the process by combining a "movie" of the screen with a narration of what was being done, which I thought was really effective.

I looked around for freeware programs to create such a video, and to my surprise found that Microsoft's Windows Media Encoder 9, which includes this capability, could be downloaded for free (because, of course, it only encodes to Microsoft's proprietary format). So I told WME to capture just the 3D app's window, demonstrated the bug as I explained where the problem was occurring, and sent it to the programmer — who then saw what I was talking about, and the bugfix should be in the next release. Hooray for technology! (And it was only 361K for 51 seconds of 567x549 video with sound, which I have to admit is pretty impressive.)

Unfortunately, the experience reminded me that I really hate my speaking voice (to my ears, I sound like a cross between Harold Ramis and David Schwimmer: nasal, yet weasel-like), and that I really suck at extemporaneous speaking (I re-recorded it several times until I didn't sound like a complete imbecile; maybe I should have written a script).

Posted by godfrey (link)
Comments
Silly me, I went to the link thinking I was going to learn about a new hairstyle.

Worlds apart we walk, you and I.

Sid
Indeed! Well, I suppose the technique could be used to design new hair styles!

Like Yoda sometimes speak, we do!

Which bug was it? Curious...
If you adjust the bias properties for a point, then select another point by clicking on a spline (not the point itself), the properties panel continues to show the bias values for the previously selected point.

Which really sucks when you're putting rounded bevels on the edges of a mechanical or architectural model...

Interestingly, I also submitted what I thought was a bug in which bias values weren't properly reflected across the X axis in mirror mode, but:

Sorry but mirror mode and bias wasn't ever designed to work. That it did
anything at all was a bug.

Which seems kind of silly to me — why wouldn't you want the bias values to be reflected on the opposite side, if the whole point of mirror mode is to let you keep a symmetrical model?

Mostly 'cause bias is mathematically dependent on the location of the CP, so if you copied bias values you'd get strange results--I think that in most cases, the gamma can be flipped neg/pos and it fixes it, but there may be odd cases where that doesn't work.

Yeah, looking at your example png, I bet negation would work for the gamma...
I realize that it's a tough one. If you have two separate splines (say, running vertically down a face), the InMagnitude handle is probably going to be on the same side of a point on both splines; however, if you're working with two points on a symmetrical spline which crosses the axis, the "innermost" handle of a point will be the InMagnitude on one side, and OutMagnitude on the other.

Man, now that's some 21st century technology; sending a mini-film to demo your software bug.
Innit, though?