Posted April 24, 2022 by OzoneWipeout
This week’s focus was on getting the save/load system closer to a reliable state and preparing other parts for the upcoming tech demo. I believe to have fixed the most common and game breaking bugs in the save/load system. There are still a couple more rare and less consequential bugs that I am aware of but haven’t put much time into fixing yet. I don’t consider those a huge priority because I have a fairly good idea what is not working and just don’t know why it is happening yet. For example there is a chance for an entire chunk to be empty and I assume this is related to a wrongly marked chunk for saving when its contents have already been removed causing it to save the chunk with no content inside. This is a less severe bug than the others I’ve fixed where only a few items were missing or the save/load procedure stopped working entirely. If a chunk is completely empty it is very likely to not have saved the data entirely and should most likely be related to a small error in when it was saved. This will also not break the system entirely as it seems to only affect that one single chunk. This does of course need fixing but is rare and allows testing of the system and the demo even if this scenario occurs. I’ve done extensive testing this week to hopefully have found and fixed all frequent and game breaking bugs related to the save/load system.
Additionally I’ve done a little bit of experimentation on various options to improve the rendering in the game. Since all my objects in the game are dynamic aka they can be moved by the player or by internal systems, many of the more common rendering optimizations (static batching, baked occlusion culling etc) can not be applied to my game. The difference in options for dynamic objects and static objects in games are massive. I’ve always made sure to keep a close eye on the rendering performance and did these experiments and testings frequently in the past to ensure I don’t run into problems in the long run with the massive amount of dynamic objects planned to be in the game. This has also always been a major argument to not go for high fidelity graphics because balancing that with the fully procedural world is if at all possible with the current average hardware incredibly hard to accomplish. However, since I did decide to increase the graphical fidelity of the game a bit in the past weeks and will increase them even further in the future I need to constantly test those limitations again and potentially add new systems to help lower the rendering impact of hardware. Combined with the mod support in the game that brings the uncertainty about the objects ending up in the game via mods (people could add photorealistic models to the game with their mods) preparing objects in the engine editor beforehand is not possible. All of those systems need to be dynamically applied to all the objects in the game to also allow them to work for those modded items. After the upcoming tech demo I will be implementing a dynamic occlusion culling asset made by an established asset store publisher for which my initial limited testing this week looked promising. There are a few more time consuming optimization systems I will be implementing in the future too but I’ll keep the information on those for another devlog.
I’m almost done with the preparations for the testing of the third tech demo and should hopefully be able to contact those interested in a few days.
Wishlist on Steam
Join the Discord
That’s all I got this week. I hope you are all doing well. Stay safe out there and I’ll see you next week.