Weekly Update 2018-07-01
Another week down, another week closer. So at the moment we are still on track to hit our next scheduled release. As is generally the case, we’ll closely monitor the situation as we drawn down on the final hour. Right now I’d estimate we are at about 65% completion with a week to go.
All of the scenes for this distribution have been written, it’s a matter of finishing up the renders. We were on track and doing quite well, but a last minute process change caused us to go back and re-render much of what we already had done. Let me explain.
We like to build the scene as we envision it. We stage the environment, put everything in place, pose the characters, adjust the physics a little bit, set the lighting an the fire off the render. In general it works quite well and much of what we’ve put together has come out looking rather fantastic. But then an inconsistency made itself known.
While working on some of the modeling for the new Patreon banners and reward images, Fecal noticed that these individual renders had a superior quality to what was being produced for the game. There was a clarity in the characters, a realism that just wasn’t being captured in the game renders. Ultimately we decided it was a matter of resource allocation.
We run Iray using a 95% convergence setting. Simplistically, this means that when Iray feels like the image has reached a threshold of 95% ‘real’ it terminates the render and stops any further enhancement. When rendering an entire scene, it could run for twenty or thirty minutes before reaching this threshold or hitting one of the max settings that may mitigate the convergence process.
Well when we render a single character with a three light rig, the convergence happens in a matter of minutes because Iray doesn’t have to worry about measure light bouncing off of anything except the character. There is no environment to deal with so Iray is able to focus on just achieving the highest possible quality for a single element, the person being rendered. Naturally, we needed to take advantage of this.
Rather than rendering whole scenes at the same time to achieve a uniform look and minimize the amount of post production work required, we decided it was best to break each scene into groups of elements and render them in separate passes. A scene containing two characters talking in an office for example, would be accomplished in three passes. One for each character, and one for background itself. It actually doesn’t cost any anything in actual render time. Each pass renders faster than the passes combined, but the combined time equates to a little less than what the whole scene would have.
Of course now we have to factor in post-production time to assemble the image layers in photoshop and make some minor changes to channels and contrast to make everything look whole again but the visual payoff is worth the extra effort. Now each primary element in a scene gains the full force of Iray’s render ability and the upscale to quality has been excellent.
I cannot honestly think of anything else we can do to enhance the visual quality of this game any further. It is one area of the project in which I think we have plateaued. All that is left is mastering shaders and materials and we will have made this thing as visually real as possible. At least with Daz.
That is all I have for the moment. Look for some mini updates throughout the week as we get closer to release. Production is a very fluid thing and we want to keep you apprised of any and all developments along the way. The wait is nearly over friends.
We cannot wait to share it with you!
We at Ianvs Salute You!