eromancer
patreon
December Progress Update!
🕑 Added 2016-12-01 17:13:16 +0000 UTCHello guys!
Thank you all for your contributions this month! Prepare thine asses for another progress update! I have a lot of topics to go over, so I’m going to get right into it.
New Team Member / Promotional Video
Perhaps you noticed the little clip above? It’s a work-in-progress sample from the new promotional video done by our newest team member! He’s an all-around audio and production badass that I go way back with, and I’ve been working on getting him involved with the project for a while now. From here on out he will take the reins when it comes to audio production, a thing that’s been severely neglected outside soundtrack work and licensing. Having experience in graphic design and video editing, he will also be working on promotion and promotional content (such as the above), another thing I’ve been completely slacking on in favor of giving attention to development matters. That brings the count to four members, all made possible solely by your contributions!
Artwork PDF Update
The v0.051 artwork PDF has (finally) been posted! The update includes the 35 missing pages from Neon’s new outfit, and improvements/fixes to the all of the Fleshpod artwork (these will be added to the game in the next release). I chose to hold off on the PDF update because I wanted to add in the improved shading techniques that I’ve been using in recent artwork to the Fleshpod artwork. Apologies for that, but $5 patrons and up can now find it here!
New Character Design (Work In Progress)
Completely unexpected, I had inspiration strike and spent a couple days modeling the face / working out the general design for a new character that will make her appearance in probably V0.07. The Inner Circle got a small preview of my work so far, but once she’s more complete I’ll reveal her to the world (in more ways than one I suppose)!
New 3D Workstation / Rendering Rig!
I’m going to geek out on this one for a bit. If you recall, I spent a lot of time in October developing a software based solution to my character artwork rendering problems, which ultimately resulted a full custom Render Manager for Daz Studio. This was extremely successful, and epitomized for me the meaning of work smarter, not harder. As much as you may not like to hear it, my personal workload this past month has heavily involved further optimization. This time however it was mostly hardware related, and as such I took the plunge and built a top of the line 3D workstation / rendering rig. This has been a long time coming, since up till now I’ve been using a mid-range gaming PC from 2012(?) for development. Up till now however it’s been completely out of our price range. This was by far the biggest expense since the start of the game, and I want to thank you guys for giving us the financial backing to make it possible!
This was no small endeavor, as these kinds of systems aren’t standard by any means, and it took a lot of research and searching for parts in order to reduce cost (to give you an idea, the system would have cost $6500+ retail in parts alone ). Admittedly, it was a lot more work than I thought going into it, and the build went anything but smoothly due to difficulties in acquiring parts, but early benchmarking and practical tests suggest it was a worthwhile move. So, what does this behemoth do for me you may ask? In short, it not only has the potential to cut a lot of time off of overall development, but will allow for more adventurous undertakings (more animations, etc). In not-so-short terms, check below:
- Extremely improved rendering times. The rendering engine used for MATM’s character artwork is CPU-based, and the 22 cores provide 44 threads against my old system’s 4 threads. The result is roughly 6-10 times faster rendering based on early tests. This means half hour batches (which most are), can be rendered in just a few minutes. One practical test showed rendering the diffuse pass from one of the recent Ven scenes to take 1 min 30 seconds, down from 9 minutes 48 seconds on the old system. This is further improved by additional software optimizations mentioned below.
- Greatly improved hard drive speeds. The hard drives for this thing have a 7x faster read speed and 3x faster write speed than standard SATA solid state drives. Since the larger 3D scenes take 5 minutes to open on my current system, this will be a welcome change. Unfortunately, these are the last parts I’m waiting on, so I haven’t been able to benchmark them yet.
- Double the SSD space for working drives, double the RAM for binge multitasking in PS/Daz. My current SATA SSDs are tiny 250 GB drives. The new ones, while not huge, are 512 GB. Since these drives have to house both project data and 3D asset libraries to see their benefit it should help a lot, meaning I won’t have to spend as much time cleaning house and shuffling data.
- Improved editing performance, especially in Photoshop. Testing revealed that Photoshop is highly parallelized. This means I should get a huge multicore boost and hopefully no longer have to deal with painfully laggy editing in PS. Daz Studio, while less parallelized, has a certain (stupid) core render engine related feature that make projects painfully slow for up to a few minutes after opening a large scene. In early testing this was completely resolved on the new system.
- - Extensibility beyond MATM. This machine will likely be my workstation for years to come, and I designed it with scalability in mind since I intend for future projects to be more demanding. Once the time comes, I can double the processing power to 44 cores / 88 threads by converting it into a dual processor system, and/or add additional graphics cards to convert it into a GPU rendering rig assuming future projects utilize a GPU-based rendering engine (which they probably will).
As someone who’s never worked in visual effects professionally, forgive me for flipping my shit over it, but it’s a pretty huge deal for me :D. Assuming some of you are also into this kind of thing, I’ve uploaded some early benchmark results comparing my old system to the new one .
More Development Software Optimization
Yes, I’m sure you’re bored of hearing about it, but since I spent time doing it I feel like I should report on it. As I noted in last month’s progress update, I wanted to overhaul my Daz Studio material presets database based on findings during Rendering Manager development that suggested I could optimize load times during batch rendering. That optimization is all complete, yielding near instant load times.
Having put so much $$$ into a new rendering system, I felt it important to further investigate improvements that could be made on the software side of things. Comprehensive testing of render settings has shown that some simple tweaks can improve render times by up to an additional 40% with virtually non-noticeable quality loss (completely non-noticeable once downscaled for in game graphics).
AltairPL’s Battle Engine Overhaul
APL is working on overhauling the battle engine to better accommodate features we've either been piling on since V0.03, have had to set aside until something like this could be accomplished, or have planned for the future (things like double teaming).
He started out the month experimenting with various ways to attack this large undertaking, and has begun by starting with what are called 'Actions'. These are basically data structures that handle info for any action a character or enemy makes during battle (attacking/skills/item usage and so on). One of APL's goals was to make the battle system entirely concurrent, meaning that theoretically actions could take place simultaneously. While we don't plan on changing how the active time system behaves from the player's perspective, doing this will allow for a lot of flexibility and power with behind the scenes stuff. It will also make the engine way more inherently stable, whereas currently stability is something that needs to be addressed on a case by case basis.
We’re glad to say he’s accomplished the foundation for this, as shown in this example screenshot. Note the multiple Vens doing stuff simultaneously and the game not crashing.
Other notable improvements that fall outside the back-end stuff include the addition of armor indicators to the battle scene. We went back and forth over a few possible designs, but for now are going with the one loosely shown in this example GIF. Basically, when an actor takes damage that damages their armor (not all does mind you) the armor indicator will become fully opaque and the section that's been damaged (future outfits may have multiple breakable sections) will flash. He also edited Neon's heavy frame indicator to hopefully better suggest that the top is invulnerable.
TK’s Map Work
TK has mostly finished up his Access Tunnels revamp, of which you can check out some of the work-in-progress processing jobs here. I haven’t had time to process his latest maps yet, but we’ll definitely be seeing the entire revamp in the next release! In other news, he’s recently been experimenting with ideas to make his life easier for working with urban environments (hint hint).
Thanks again for your support this month! Sorry I don’t have more to show for V0.06 yet, but we’re definitely looking forward to tackling the new content (me especially after having been in optimization hell for so long). Anyways, here’s hoping it goes smoothly! Till next time!
- Eromancer
Comments
New releases aren't really tied to the monthly schedule. They'll be released as they are finished (something the new format gives us the flexibility to do now is to release them anytime, not just at the beginning of the month).
Robert Dant
So is there an actual schedule that shows the plan to release a demo each month?
I love the video. But I just like to say the intro video has a very Battlefield 3 / BF 4 feel to it...
Odriew
With the exception of what sounds like several exciting updates coming through in the battle system, my favorite part of this was the $4.25 shipping on a 4000 dollar item.
