Hi Marco,

> 'cause i tested that those artifacts are also captured by both the commands
> getViewport.renderToClipboard() and getViewport.render(), two commands that
> would be very useful for creating presentation images of the 3d models.

Yes, that's to be expected because both of those methods use the viewport realtime display
mechanism. They are meant to just capture the display and are not really intended for
making final presentation graphics.

For better quality presentation images, the PDF or AI exports will work much better, those are
the functions that are more oriented towards making a high quality graphic result and not
only primarily focused on speed like the viewport display is. Or also the other main way to get
high quality renders is to export to a rendering program.


> Anyway, do you think that the next V4 version, will integrate an optional better display
> mode, in case of using a discrete video card, like those offered by Nvidia Quadro series ? <....>

I don't know yet what will be done in this area for V4. I have a few ideas on how to improve
that particular display artifact but it has to be done quite carefully in order to not degrade
the performance of the display engine. So any changes have to be tested quite a bit to make
sure it doesn't come at too high of a performance cost. That type of stuff involves quite
a bit of work and it's difficult for me to predict when such things will be completed. It's very
important for the display to be quick so that it can be interactive, the display engine is just
overall much more focused on speed and not on trying to make a presentation type graphic.

Any future improvements in this area will probably involve calculations done by the CPU
before stuff is sent to the graphics card, and so probably won't be dependent on any
specific video card.

- Michael

Hi Michael,

now it's all clear. Thank you for your comprehensive answer!

I understand your technical point of view...you prefer to be independent from any particular hardware requirements,
and do all the tasks leveraging on the cpu rather than any type of gpu.

From what I've read here and there on the web, i'v seen that there are two main approaches to display 3d graphics models in realtime.

Someone prefer to squeeze the gpu up to its limits, while others
say that the tasks that a cpu can accomplish are more accurate and reach better results.

I'm sure that whatever approach you will adopt for V4, it will be a good compromise between speed and quality,
(also winking at a better antialiasing quality :)

Have a good work,

Marco.