Creating a good benchmark scene that actually does a good job of testing a variety of Terragen features and giving a representative value for a broad range of computer hardware is more challenging than you might think. Not so much in the sense of actually building the scene, but as far as determining how to build it, what to put it in it, what render settings, etc. Combine that with other constraints or desires, such as including object files (size of download then becomes a concern, usage rights, compressibility, etc.). Basically there are a lot of factors that are not so obvious and are particular to benchmarking, as opposed to just making a "nice looking scene".
That being said, we could potentially open it up to submissions, scenes that people were happy to share and did not mind being redistributed for the purposes of benchmarking. But we would have to take any submitted scenes and evaluate them for suitability, likely adjusting settings (especially render settings), and adding or removing things as-needed. It might be less work, it might not be, hard to say. But we might get a prettier end-result scene.
Anyway, we have not spent a ton of time on the benchmark of late, but it's on the list of things to tackle in the next month or two. Once I do dive into it, it will not take weeks, so you needn't worry it's taking a lot of time away from other things. If there is genuine interest from multiple people in submitting scenes for possible use as a benchmark, I'll discuss with Matt and see if he's amenable to that, and if so what general criteria we'd want submissions to adhere to. But we'd want to see at least a couple of people throwing their hat in on submissions to take that route I think. If not we'll just do the scene internally, it will be simple, but will showcase a reasonable variety of TG functions, and will accomplish the other more technical goals.
- Oshyan