Very very slow development of new version of Corona, why?

[Juraj]

That kind of chart would be very interesting across industry specialization. Judging by ArtStation, for something like Concept Art, Octane would be super high.

I am suprised Vray kept such dominance, but this chart is not only within Archviz/AEC, but also within CGArchitect and its mainly US community. 976 responses also seems on rather low end of scale.

[frv]

I don't know of any other survey at this scale. It does say a lot.
I do feel it aligns with the situation in The Netherlands. Most studios that do Archviz still use traditional render engines like Vray. It's very hard to change to something else too. It's not just the engine but alle the assets and resources, plugins and platforms like Max that come in to play as well. Twinmotiom and Lumion are more or less in-house solutions.

[BigAl3D]

I chuckled when I saw this thread. In the past, I have complained that I thought Corona was upgrading full versions too fast. Ha. I think Tom makes a great point about a maturing application becoming a product with refinements, speed boosts and the occasional major feature. I mean, how many groundbreaking features does Microsoft add to Word these days?

In the old days, major versions of 3D apps always boasted about render speed increases. I don't see that much with Corona right now. Mainly relying on the number of cores on the system being used. If the features stayed the same, but render speed was increased by 50% or more, that is a game-changer for most. Corona is already great and artists are making amazing images now, but if they can do the same thing a lot faster? Boom.

[TomG]

Increasing speed follows the same curve as software maturing - big gains are possible early on, but this reduces over time. After all, you still have to calculate how light bounces around in a scene, which takes certain mathematical calculations, which take time.

You can liken it to top speed in cars, can't verify the accuracy of this graph but it'll do as example:
https://www.sharpsightlabs.com/wp-content/uploads/2014/12/data-analysis-example_scatterplot_fastest-speed-by-year_ggplot2_600x350.png

Looking at the averaged curve not specific individual points, to go to 1.5 times speed from 100 to 150 took about 15 years, from 1955 to 1960. To go to 1.5 speed again (225), it took 40 years, from 1960 to 2000. To go to 1.5 again... well, we haven't got there yet, and may not, because there is friction and all sorts of physics and the car flipping over as it can't stay on the ground any more and all sorts of other constraints.

So while we are always looking for ways to improve speed, those are now smaller increments that are possible - because despite our slogan, rendering is not magic, but mathematics :) Just in case anyone starts impatiently waiting for 50% speed boosts!

[Cinemike]

So while we are always looking for ways to improve speed, those are now smaller increments that are possible - because despite our slogan, rendering is not magic, but mathematics :) Just in case anyone starts impatiently waiting for 50% speed boosts!

I wish you could link Corona's render speed to inflation ;)

[BigAl3D]

I have Matt Damon working on it now... using mathematics.