Hidden Line Rendering - Do these options add render time if no such objects exist

[Kevin McAllister]

It seems like the "Display Surface Hatches" and "Display Text and Markers" options are enabled by default with hidden line rendering in Sheet Layer Viewports. Do these settings add render time if no surface hatches, text or markers are present in the design layers associated with the viewport e.g.. do they generate a rendering pass regardless of whether its needed or not?

Kevin

 

[PVA - Admin]

I am 80% sure that they do. but @Dave Donleywould know better than I.

[Kevin McAllister]

Thanks Jim. I suspected they might based on the little bit of testing I did here. Do you remember my batch rendering problems? Those items are turned on in most of those viewports. They're not something I turned on myself so either on is their default or they are turned on when bringing files forward from an older version that didn't have this feature....

Kevin

[PVA - Admin]

Pretty sure they're on by default. I didn't alter that setting when I tested it in the beta, hopefully even if that was an issue before it would be corrected by release. I know quite a few backend renderworks upgrades took place this year that won't get much fanfare.

[Kevin McAllister]

A couple more questions about rendering. I just did an experiment rendering the same viewports a few times in a row. Then I restarted VW and rendered it again. I used a stopwatch to time each try. My times ranged from 4:32 all the way up to 6:10 for the same section viewport. The one right after restart was the fastest. What else would cause the rendering time to vary? Why do viewports with hidden line in the foreground often flash (refresh) between linework on / linework off at regular intervals while rendering? It also seems like a lot of processor cycles go to teapots (one on the viewport, one in the bottom right hand corner of the drawing window) and timers (bottom right hand corner).

Kevin

[PVA - Admin]

The first render I do after a restart seems to vary wildly from concurrent renders. I've seen this in some applications but not others, as well as in certain render engines in things like C4D where you can choose from a selection, I do not know if it is a common ocurrance in rendering sectors or not. I discount the first render entirely when comparing rendering times for this reason. The last time I did time trials (which I think was 2016 SP0 with Caustics) I got consistent results with the 2nd and all subsequent renderings within a margin of error of about 5%