Cheetah3D 6.0 is out - What's next?

Hi Martin!

You are bringing excellent news to us!

I am glad to hear that beside the upcoming 6.3 we are having even another 7 flagship well on the way.

Keep up the good work!

Martin said:
Hi,
yes, my original plans for the v6.x development changed quite a bit. Sorry for that. I should better not comment about future plans anymore. So I can't disappoint anyone.:wink:

The feature set of v7 was highly motivated by the wish list forum. So I'm sure that many of you will get at least some wishes fulfilled.

Since the v7 feature set was pretty ambitious I'm behind my original timetable although I already started very early with the v7 development this time. So v7 will still take quite some months. But it's definitely worth the wait.

v6.3 which will be released pretty soon and will offer approx. 30 bug fixes and minor improvements.

Kind regards
Martin
Click to expand...
 
podperson said:
  1. Typically, vector motion blur is a near zero overhead on rendering time. (All it involves is making fairly simple calculations that are largely needed just to assemble the scene and storing the results in a buffer.)
  2. Pixel Motion Blur requires you to render extra frames, so the increased rendering time is far greater for Pixel Motion Blur.
  3. I am free of Adobe and trying to stay that way.

Hey I'd love both, but if I can only have one I'd pick volumetrics. While they might be spiritually similar, the implementations seem to be quite different (e.g. 3D Studio Max had decent volumetrics long before it got SSS support; OTOH Cycles has SSS support and no volumetrics).
Click to expand...

If you have seen fast motion blur it's probably not that accurate. Sample based motion blur is probably the best, and is pretty slow. And it essentially produces extra frames as well, but compresses them down. Any other type of motion blur (vector) is similar to what can be done in comp pretty easily. You can often even export the motion (velocity) data, and generate it in post utilizing the data. I've done motion graphics in the advertising industry, and we almost always simulate motion effects in post. It's just significantly easier. And it's fairly easy to use other packages (non-adobe) to the same effect. All I know is I'd never render using motion blur. (except maybe a still)

And Blender has a patch with volumetrics from this years Summer of Code if you want to try it out. I'm sure graphicall has a build.
 
Martin said:
Hi,
yes, my original plans for the v6.x development changed quite a bit. Sorry for that. I should better not comment about future plans anymore. So I can't disappoint anyone.:wink:

The feature set of v7 was highly motivated by the wish list forum. So I'm sure that many of you will get at least some wishes fulfilled.

Since the v7 feature set was pretty ambitious I'm behind my original timetable although I already started very early with the v7 development this time. So v7 will still take quite some months. But it's definitely worth the wait.

v6.3 which will be released pretty soon and will offer approx. 30 bug fixes and minor improvements.

Kind regards
Martin
Click to expand...

No no. Your open communication is always appreciated. I'd prefer you be outspoken. If you don't meet the goals you communicating to us like this is what we all appreciate.
 
Yeah. I remember using C4D when it was cheap, and you could even get it free from magazines. Now it's thousands of dollars. They used to be a fun, and friendly company. Now they cater to studios of all sizes, and charge ridiculous sums of money. Their modular approach seems nice until you realize you can't just upgrade the core, but you have to upgrade all your modules, or they won't work anymore. The MSA is an improvement, but I'm not a fan of them anymore. Amazing software...but they just lost me with their pricing strategy. People complain about Adobe, but I love the subscription deal. While I do wish they allowed me to subscribe and literally own the package...the monthly fees are low and easily planned with budget and taxes. Anyone who reads the fine print of most software should know you never really own it. Most software agreements actually read as you leasing the right to use the software, and that the right can be withdrawn at the software companies discretion. Small companies like Cheetah produce the only software that doesn't have agreements like that.
 
Yeah, the pricing structure and the "feature roulette" made me stop with them. In fact, I just loaded my old v10.5 today to see if it would work on OS 10.8.5 and it does. I think they're up to v15 now.

To get the features I already have I'd have to pay about $3000 for in the new structure. To "downgrade" to their Prime package it would cost about a grand. And, I'd be LOSING features. Nutty...

I will keep an eye on Cheetah. I really like it, but some of the features I'm looking for are lacking at the moment (which I've mentioned before).

The dynamics stuff is really neat, though. Especially at Cheetah 3D's price.

Bob
 
podperson said:
I think motion blur is vital for rendering anything that moves, including particle and dynamics animations. There is a workaround, but it's barely practical for stills and unbearable for animations. So I think it's as big a deal as Volumetrics and NLA.

(I'd love to see SSS, and it has been promised for some time, but it's much less important.)

I said a while back that C3D could give up on trying to differentiate itself with its renderer (and just plug in Cycles, Lux, or whatever) and I'd be perfectly happy with that. I don't know how other people feel, but I think that C3D's renderer has pretty much fallen behind the curve and is never going to catch up. On the other hand, C3D's UI is so far ahead of the competition that this is where it should concentrate, so another option is to support external renderers (and thus get SSS, Volumetrics, motion blur, etc. for "free") and concentrate on things like NLA, rigging, modeling, UV-mapping, and so forth, that C3D can do better (or at least more simply) than any other program.
Click to expand...

An additional renderer capable of new features would be welcome, as much as I love the ease of use and quality of Cheetah 3D's renderer.

Agreed about the interface; Cheetah 3D's interface is phenomenal from my experience with other 3D software over the past 15 years or so. Few programs are so easy to get the hang of with a tool layout so much fun to interact with.
Also thanks to the clever icons and camera adjustment icon/ tools. Just my opinion, but it's tough for me to consider more powerful CGI programs when Cheetah 3D continues to have such a user friendly set up.

----Quick Wishlist v7x
Faster response time with real time playback during higher polygon counts would be my biggest wish with version 7x, though this is an editor as mentioned, not a game engine like Unity. :cool:

Martin said:
Hi,
yes, my original plans for the v6.x development changed quite a bit. Sorry for that. I should better not comment about future plans anymore. So I can't disappoint anyone.:wink:

The feature set of v7 was highly motivated by the wish list forum. So I'm sure that many of you will get at least some wishes fulfilled.

Since the v7 feature set was pretty ambitious I'm behind my original timetable although I already started very early with the v7 development this time. So v7 will still take quite some months. But it's definitely worth the wait.

v6.3 which will be released pretty soon and will offer approx. 30 bug fixes and minor improvements.

Kind regards
Martin
Click to expand...

Any hints on what the minor improvements might be?

Martin, I love the particles system; great job. Will there ever be the option to input the Start End times as frames (Start at frame 5, end at frame 842) instead of Times (start at 3.47 seconds for example)?

Sorry, I should put that in the wish list when I get a chance.

Looking forward to 6.3 and the remaining 6x versions.
 
Nichod said:
If you have seen fast motion blur it's probably not that accurate. Sample based motion blur is probably the best, and is pretty slow. And it essentially produces extra frames as well, but compresses them down. Any other type of motion blur (vector) is similar to what can be done in comp pretty easily. You can often even export the motion (velocity) data, and generate it in post utilizing the data. I've done motion graphics in the advertising industry, and we almost always simulate motion effects in post. It's just significantly easier. And it's fairly easy to use other packages (non-adobe) to the same effect. All I know is I'd never render using motion blur. (except maybe a still)
Click to expand...

I'm actually mainly concerned with stills but to put a simple counter-example ElectricImage was famous for its vector motion blur (at the time a trade secret). It was good enough for ILM to use in numerous movies (including several by George Lucas, James Cameron, and Steven Spielberg). Vector motion blur actually works better than sampling for many purposes — e.g. a point light in the distance will be sampled as a bunch of dots, whereas vector motion blur will actually smear it.

Because motion blur is an attempt to simulate both photographic artifacts AND perceptual artifacts, it's as much an art as a science, and there is no one perfect solution, but our only option right now is sampling, which is in many cases both inferior and in all cases the most time- and resource- intensive option.

And Blender has a patch with volumetrics from this years Summer of Code if you want to try it out. I'm sure graphicall has a build.
Click to expand...

Blender has vector motion blur implemented exactly as you describe above and some form of volumetrics or other for years (iirc Blender's volumetrics were, historically, awful).
 
podperson said:
I'm actually mainly concerned with stills but to put a simple counter-example ElectricImage was famous for its vector motion blur (at the time a trade secret). It was good enough for ILM to use in numerous movies (including several by George Lucas, James Cameron, and Steven Spielberg). Vector motion blur actually works better than sampling for many purposes — e.g. a point light in the distance will be sampled as a bunch of dots, whereas vector motion blur will actually smear it.

Because motion blur is an attempt to simulate both photographic artifacts AND perceptual artifacts, it's as much an art as a science, and there is no one perfect solution, but our only option right now is sampling, which is in many cases both inferior and in all cases the most time- and resource- intensive option.
Click to expand...

Again easily achieved in post. Nuke has a vector motion blur node that produces phenomenal results. After Effects can do the same thing. And vector motion blur does sample, but in a slightly simpler way, it actually grabs the velocity vector (direction of the motion) of elements in the scene, and focuses sampling/blurring on those areas. So it does actually sample. Brute force sampling actually samples all elements, so you get accurate blurring in reflections, etc. it is slow, but definitely not inferior. Some recentish implementations are pretty quick..like Renderman's 3D motion blur which is much more accurate than any we have described, because it actually considers depth in the scene referencing the camera.

podperson said:
Blender has vector motion blur implemented exactly as you describe above and some form of volumetrics or other for years (iirc Blender's volumetrics were, historically, awful).
Click to expand...

I was referring to cycles volumetrics being implemented.
 
And vector motion blur does sample, but in a slightly simpler way, it actually grabs the velocity vector (direction of the motion) of elements in the scene, and focuses sampling/blurring on those areas.
Click to expand...

Let's just play a game where we assume that we both know how this stuff works ;-). The point is no system works perfectly (want a great example where vector motion blur has problems? reflections).

In general, sampling an overcranked animation is by far the least render-time efficient option, and it's highly debatable which approach produces the best results. My view is that if well-implemented vector motion blur is good enough for ILM and Pixar, it's probably good enough for me.

Cycles, incidentally, gives you perfect motion blur for free. (Edit: well, upon playing with it, it's theoretically perfect but apparently got some hideous bugs :) ).
 
Last edited:
The best approach is usually somewhere in the middle ground, as always. A low level of multi-frame rendering for samples, smeared together by the vector motion blur usually looks great.

Another area where vector blur fails is where movement occurs behind other, thin objects, i.e. a baseball moving behind a chain link fence. The smearing tends to be non-selective.

Outputting a motion vector pass and controlling the motion blur in post would be another, highly desirable, approach.
 
You must log in or register to reply here.
Back
Top