Poly Haven
patreon
Our Vision for 2023
🕑 Added 2022-11-08 04:11:21 +0000 UTCPoly Haven has grown a lot this year, thanks in no small part to your continued support on Patreon and through our new Blender add-on <3
I've been thinking a lot about how to move forward and figure out the best ways to produce better free and public domain 3D content - and I'm excited to announce some of these plans to gather feedback from you before implementing them!
First of all, about our current funding goal on the website:
The current goal of hiring a full time texture artist was already somewhat implemented earlier this year when we hired Dario, and we've also started hiring contractors to produce texture scans both independently and as part of our pipeline for some time.
This progress bar also only includes Patreon donations and corporate sponsorships, yet these are not our only income sources now. Our other sources (the add-on sales and bespoke licenses) have grown significantly this year, though they are too unpredictable to be fairly represented in such a progress bar without being misleading, or enticing us to commit to something we might not be able to sustain long-term.
For this reason I've decided to remove the over-simplified and potentially misleading progress bar and instead communicate directly here and in our regular dev logs about what our future goals are :)
So what are our goals?
As you might have noticed, our publishing schedule has become almost a daily thing, and I suspect within the next few months we'll be publishing an asset every single day of the week including weekends, and potentially begin doing multiple assets per day.
This makes me eager to push our workflows further towards improving quality, and focus more on the usability of our assets.
Here are some of my ideas on how we can improve, but I'm keen to hear your feedback on what you'd like to see from us in the future:
1. Improve the overall value of all assets by producing them in themed collections. We've already been trying to do this to some extent (e.g. with the Smuggler's Cove and Shed projects), but we'd like to put more focus on this and plan regular travel to specific environments for scanning and reference capture. We'll make a dedicated post about our first trip soon.
2. HDRIs: Include light measurements and calibrate intensities to real-world values; include capture of grey and chrome spheres along side the color chart, with a consistently defined workflow to allow further verification and reliability of data; Improve standards of how we handle and remove lens flares, particularly in bright sunlight or night environments.
3. Textures/materials: Explore vector displacement in cases where regular single-dimension displacement doesn't work well (overhangs and spikey things); Train one of our staff as a licensed drone pilot to capture more aerial scans; Emphasize our current standards of color accuracy and calibration, perhaps communicating this better on the site; Provide more precise scale measurements of the surface size.
4. Models: Provide LoDs for future models - technology like Nanite in UE5 is not very widely available outside of Unreal, until then LoDs are maybe still the best way to optimize assets even for offline rendering; Provide our assets natively for Unreal to avoid nasty FBX issues; Improve FBX support and investigate USD again; Rig all future assets for animation and easier control and placement.
Along with improving quality of HDRIs, textures and models, we plan to start looking into new asset types such as atlases, imperfections, and procedural generation tools (geometry nodes!).
Our focus is still on creating assets for realistic environment art, we don't plan to expand into stylized content as there are far too many different styles to cater to.
We also intend to stick to photo-based workflows and avoid procedural materials or rendered HDRIs. Our 3D models may still utilize procedural materials, but are typically based on a foundation of photographic content.
As for how to implement these ideas, we have the budget to hire two new artists next year. We're looking for local (South African) artists for these roles as we plan to create a studio and shared office space to facilitate deeper collaboration especially when it comes to developing new workflows for new asset types.
With all that said, let me know what you think! Do you have some other ideas? Or feedback about our plans above?
Comments
S C
All this is absolutely fantastic. Although not sure I would give up on procedural materials quite yet. I still find them really useful when scanned feels off. And I can count on your procedural materials being a cut above others I acquire.
Poly Haven
Once we have a drone I do plan to play around with these and see if there's a good way to do them :) Realistically I think we'll need a custom drone with a special camera rig to do them properly, but we have to start somewhere.
Danny
Are drone HDRIs still on the table? Even just a few high quality ones would be amazing
Poly Haven
We're definitely considering timelapse HDRIs, though it is quite a niche use case and a huge amount of work, so I'm not sure it's the best use of our resources right now. The audio samples to go with the environments is a nice idea though.
Darryl L Kilzer
Thank you for all the great work! Just keep doing what you do, it's awesome!
Mircea Kitsune
I have a question: Is it possible you could someday support HDRI videos for animation? Those are very hard to find especially under a free license. They would allow better quality animations, for instance by having people walking around or water moving or trees swaying in the wind, compared to static images where everything's frozen in time and there's no associated audio. Thank you.
Poly Haven
A system for variations of assets has definitely been on our minds lately :)
Jan Krassnigg
"LoDs for future models" - that would be amazing! Love the models, but for real-time rendering they can quickly bring a PC to it's knees. Also some of them take so much memory, that using a lower LOD would generally be sufficient for many games.
Thierry Mayrand
Have you guys thought of altered models, e.g. cut or aging fruit? Thanks for all your hard work
Terry Buttocks
I did not realise that there would be a lot of extra overhead, and I can see that it does not make much sense to include these. I sometimes composite elements, esp the sky into final images - The Sky Is On Fire has some excellent backplate images.
Poly Haven
Can I ask what your use case for backplates is? These take a significant amount of extra work to create (also requiring a 2nd person to be present for the shoot, with another camera), but only around 1% users will download at least 1 backplate along with the HDRI.
Terry Buttocks
Fantastic!! You guys ROCK!!! Thank you so much for all you are doing. One humble suggestion: would it be possible to include some more backplates with the new HDRI's please?
