In this animation the deformation function (just a simple sine function) is
offset 180 degrees from letter to letter. This is most obvious between the "u"
and the "l." The animation is cyclic.

Have a great day everyone!

Kind regards,
Dave Blandston

Post a reply to this message

Attachments:
Download 'lavatext.mp4.dat' (641 KB)

I like it-- even though my name isn't Paula :-P

Height_fields, I assume?

The interesting thing from my point of view is that the downloaded animation
actually plays in Windows Media Player (on my Win 10 machine); this is the FIRST
time that a file from POV-Ray's web portal, when downloaded, actually does so!
In the past (in Win 7 and Win XP) I always had to use a different media player,
I think because of the additional '.dat' suffix that has always been added to
such files, and which I could never 'fix'. AVI, MP4, they all had the same
problem.

But that has now changed (for whatever reason)!

I'm curious to know which app you assembled your animation frames in. Currently,
I use the old but still-available VirtualDub2 MOD, although I haven't yet
uploaded one of my current POV-Ray MP4 videos since purchasing my Win 10
machine.

Post a reply to this message

"Kenneth" <kdw### [at] gmailcom> wrote:
> Height_fields, I assume?
>
> I'm curious to know which app you assembled your animation frames in.


No height_fields, just CSG boxes and cylinders and such.

I used ffmpeg to make the .mp4 file using this command line:

ffmpeg -y -r 60 -start_number 0 -i Frame%03d.png -crf 17 -c:v libx264 -pix_fmt
yuv420p -r 60 LavaText.mp4

I don't actually know what some of those switches even mean but the end result
is a high quality 60 FPS .mp4 file. I'm glad it worked for you! I know exactly
what you're talking about regarding the past problems with animation files and I
agree, it's great that it works now.

Have a fantastic day!

Kind regards,
Dave Blandston

Post a reply to this message

"Dave Blandston" <IsN### [at] protonmailch> wrote:
>
> No height_fields, just CSG boxes and cylinders and such.
>

That's interesting. My initial guess about height_fields was based on some
artifacts that I see, on the 'sides' of the letters U, L and A. Like moire
patterns, but more linear; not quite the the 'look' of possible AA artifacts, or
even due to the vagaries of MPEG compression-- more like the  effect when the
side of a height_field is completely vertical (and when the HF does not have a
high-enough triangle resolution).

The sine-wave traveling through the letters has be intrigued. If the letters are
just CSG objects, how did you 'squash and stretch' them? My guess would be:
using functions for initially creating the (many?) 'simple' object shapes, then
adding an animated sine function on top of that. OR, by 'slicing' CSG objects
into lots of thin sections, and maybe scaling the slices in x using the various
phases of a sine function?

The artifacts I see make me think of thin sections combined into full letters.
All guesswork, of course ;-)

Actually, my very FIRST impression was that the letters were bent by an animated
sine-wave 'camera normal'-- but the background is static, so that's not the
trick.

Post a reply to this message

"Kenneth" <kdw### [at] gmailcom> wrote:

> The sine-wave traveling through the letters has be intrigued. If the letters are
> just CSG objects, how did you 'squash and stretch' them? My guess would be:
> using functions for initially creating the (many?) 'simple' object shapes, then
> adding an animated sine function on top of that. OR, by 'slicing' CSG objects
> into lots of thin sections, and maybe scaling the slices in x using the various
> phases of a sine function?

Very good question, Brother Kenneth.

Having gone through all of Dave's Bordered Characters (BC)code, I can say that
he could - and would - move a mountain using only a teaspoon.  However, I think
all of that is pretty complicated and would require a full reworking of the BC
code to pull off.

So, I'm gonna guess that it's a very skillfully applied transform matrix.

Post a reply to this message

"Kenneth" <kdw### [at] gmailcom> wrote:
> The artifacts I see make me think of thin sections combined into full letters.
> All guesswork, of course ;-)

Howdy Kenneth,

You guessed very well! You are correct that the characters are made of thin
slices. I'm not nearly as adept at mathematics as I would like to be and am
therefore limited to the less elegant and sophisticated methods. The moire
patterns are caused by the union of the differing-width slices. One fine detail,
however, is that the slices are not scaled in the x dimension, but rather each
slice is intersected from a character that is created at the correct size for
that particular deformation function value. If the slices were simply scaled,
the inner parts would not match up. This is also what allows for the vertical
deformation.

The next font variation in the set (#38) applies the same technique to the z
dimension as well. The effect is stunning in my opinion but requires radiosity
to really see what's going on. A character would be sliced up into (number of
vertical slices) * (number of depth slices) blocks which results in very slow
rendering times. An animation should be very cool indeed but due to the render
time required I haven't done it. Maybe a single word ("POV" would do) could be
animated at relatively low resolution in a couple weeks.

Here are the demo images of the font variations in case you want to see what #38
looks like:
https://news.povray.org/povray.binaries.images/thread/%3Cweb.59f0b9385b47c265c581d2c50%40news.povray.org%3E/?ttop=43635
9&toff=500

I've added a few more characters (@, cross, arrows) since the last "Border
Characters" posting so if anyone's interested I'll post an update.

Have a wonderful day!

Kind regards,
Dave Blandston

Post a reply to this message

"Bald Eagle" <cre### [at] netscapenet> wrote:
> So, I'm gonna guess that it's a very skillfully applied transform matrix.

What??? There's an easier way???

Post a reply to this message

"Dave Blandston" <IsN### [at] protonmailch> wrote:
> "Bald Eagle" <cre### [at] netscapenet> wrote:
> > So, I'm gonna guess that it's a very skillfully applied transform matrix.
>
> What??? There's an easier way???

Ha!
Don't we wish.

It was late and I was working from memory.

And you did go ahead and "do it the hard way", didn't you...   :D
"Each and every pixel in this animation was artisinally hand-crafted from
photons generated with power from a custom-made foot-pedal operated
generator...."

I was probably mixing matrix transforms and Mike Williams' non-linear scaling of
isosurfaces in my head.  A matrix won't use the proper interpretation of 'x',
since it's outside of the function parser.
I suppose you _could_ make an object pattern function and do it the isosurface
way, but that would introduce the added complication of texturing all of the
separate parts of the lettering...

"the slices are not scaled in the x dimension, but rather each
slice is intersected from a character that is created at the correct size for
that particular deformation function value. If the slices were simply scaled,
the inner parts would not match up. This is also what allows for the vertical
deformation."

Ah, going back and watching your animation on 'loop', I understand what that
means and what you're doing.  Very clever, and it gives a very nice effect -
that goopy, squishy effect that the vertical deformation achieves.

You do nice work, Dave   :)

Post a reply to this message

"Bald Eagle" <cre### [at] netscapenet> wrote:
> Ha!
> Don't we wish.

Yep that was a big project and a lot of work but also lots of fun. Thanks for
the compliment sir!

Kind regards,
Dave Blandston

Post a reply to this message