- Aug 26, 2018
- 728
- 7,029
I usually priotize quality over filesize, so I chose to ignore mod size complaining, it's meant to be an extra enhancement and enhance the overall quality level of animations so i would say to them "deal with it or don't use it"This is a question for me.
The simple answer is: Because of complaints and comfort/laziness.
There were one limit (max. 16 MB/s) on other NLT HD-Mods.
People start complaining. It's too large, please spilt it up.
I don't have premium-account, which means, if a file is not downloaded within 10 days, it gets deleted.
People start complaining again and I was uploading again and again.
This is the reason why I choose AV1 as codec.
8MB/s quality-wise is equal to 12-16 MB/s of VP9.
A smaller size mean less time uploading.
It also looks a lil' bit worse ingame anyway.
My goal in TON and TGO was max compatibility; but thats the goal anymore in SotS.
AV1 should run on 6-8yo CPUs/GPUs. If not... well that's unfortunate.
FYI: I have a "4K+ No-Compromises-Edition" of every HD-Mod on my server (just in case).
I have a personal cloud-server; but this is for work/science projects only.
Next option are premium-accounts from file-hosters.
But I have all the hardware/software- and labour-costs and should pay in addition to that for free services on a pirate-site, where nobody wants to pay anything anyway.
Sorry. This sounds harsher as it should. No bad feelings on my part. I'm just based/realalistic here.
Do as I do: ignore irrelevant criticism
AV1 encoding makes sense but i don't know the performance impact compared to h264 or VP9 for low-end PCs
gofile and pixeldrain are fine for your uses. you can also use Mega but only for a small quota
For me setting a CRF value is still the best way because it lets the encoding algorithm choosing the optimal bitrate when needed even though I don't think our disagreement has much impact on the final result.
I'd agree on this one, these animations definitely need a face-reconstruction model (iris or rhea for topaz video AI)I'm not sure about the need for such extreme actions.) As for me, what was in 16mb was the best... not even 12 or 8... Exactly 16. Yes, such settings remove some details, like Amira's on her clothes, but there The hair also became more soapy in the new 2k.
Even if you try to be close to ideal, there will still be a more soapy stain in some area, or a part will be erased. And so the faces look great, but in the clothes those details are not so visible..
There, in your new comparison with Julia, she is wearing green shorts, and the new settings have partially blurred her pattern.. That is, there are also some nuances even in clothes..
IMO, this workflow seems far too complex for the small details it could add, as well as increasing the risk of bugs and artifacts.
That being said, it's important to remember that the complexity of enhancement for this game is absolutely tedious, and I share don Kilot's pain just looking at the original animations.
Very poor encoding, plastic character models, inconsistent artifacts between scenes, poorly realized character movements sometimes worsen by interpolation, etc.
Last edited: