Leakoo
yiffalicious from patreon
yiffalicious patreon

Interaction preparations - Memory & load optimization

🕑 Added 2022-01-20 05:13:39 +0000 UTC
Interaction preparations - Memory & load optimization

Comments

FurryVNE Team

This doesn't have anything to do with the amount of parts you can have in a character. If you're exhausting the limits you could try asking for help in the forum or on discord and we can help sort you out because you should be able to create almost anything despite the limits if you're using the editor correctly. All the best. - odes

Chad Saunders

Hey guys gotta say I’m with the other guys this time with the load times it can become a problem loading other people’s scenes I imagine but when creating I don’t mind waiting a little longer if necessary I can see you’re trying to strike a good balance and speed things up with minimal quality loss but if it ever comes to a point where you have to sacrifice one or the other in my opinion I’m willing to wait longer to have my models load at their best quality That being said will these updates only affect the size occupied on the v ram or will it also change the amount of assets we can add before uploading a character to the cloud because I’ve got multiple characters I’ve never been able to publish because of the size limitations?

rugarou

interesting... so personally, I dont know about the others, but "load in" time doesnt bother me much if everything runs smoothly after the fact. massive mod-lists in bethesda games taught me paitience. Im with Austin on this. however the only issue with leaving texture compression to the end user is ... they wont... old hands like Craket and Blaky know to compress and optimize thier creations/textures. Newer users arent going to do this. the problem with making easy to use tools are they are easy to use. this inevitably leads to inexperienced users using the tools.. lol... so its good to draw new users, but then you also get all the problems associated with it. learning curve. having a server for textures etc is going to use bandwidth, a lot of it, and thats going to be expensive, so at that point you are also accumulating an expense and not just an item. it also implements another potential "fail point" with multiple parts that can fail in its pathway. tho I guess ultimately its not much different than the interactions server, it will just be another VM youll have to run alongside the interactions server assuming you guys plan to keep that concept around.

FurryVNE Team

You can already change resolution in preferences, and for parts/garments you can pick a specific resolution to use for the part.

Austin

Was going to say what someone already said of making the compression step as part of the publishing process. The server solution is trying to solve a problem that you don't yet know actually exists for end users. It introduces a lot of complexity and cost without knowing the payout. Make it simple for now and expand on it later. Another interesting idea might be offering a way for creators to define the resolution of the textures during the character creation process. I don't expect every character needs the absolute highest resolution for everything. It does however introduce potential problems with uv mapping accuracy which some features I imagine are very heavily reliant on.

Craket

Good working there. But I must say. Waiting 20 seconds for an 8 character scene to download is not that bad. If you are considering user experience, with so much technical heavy lifting a pc can do, you the have to implement stats per character/scene that tells you "Characters in scene" - "Scene mb size" - "Scene Length". That way, in terms of user experience, people can read into if something is heavy to run or not. Same should be applied to a character "Parts&garments used" - "Character mb Size" etc. Because by the end of the day, you might not be able to dodge making anything faster, cus we will just mess things up and go hard on it. The more optimized you make it, the more "room" for us to be hard on the app. Which is good, but not too necessary. Just give the people the information if something is going to be though to download/open.

FurryVNE Team

Your ideas were among the top I had. The first one you already identified a problem with - the fast load format is prone to changes. So you'd have to do it periodically. As for the 2nd solution - it's definitely doable. But also at the cost of user experience as well as introduced complexity. Things like internet upload speed, multiple users trying to upload to the same character at once, errors in the uploaded file not to mention intentional abuse are all things we can circumvent with a dedicated server. Not to mention the user experience would be better, too.

Andrew

Why not just make baking the fast-load version of a character one of the steps of uploading? Then only the creator will have to wait through that. Or if they have to be redone periodically (say, for new software versions), perhaps you could have the first n clients that try to load it get the slow version, build the fast one, and then automatically upload their results. Once the server receives a few identical results, it can cache it and serve it to everybody following. Offload your computing on the users!

FurryVNE Team

https://gpuopen.com/compressonator/

would you be willing to share with us what the 3rd party CLI's name is? the results are great and I'd love to see what else they've worked on.

You and your team could take the whole year trying to optimize these changes, and I’d gladly wait. It may not seem like much, but I just want to say thank you for the effort that you all are putting in for this.


More Creators