Update (Patreon)

Good day everyone, time for a quick update post!

I mainly wanted to talk about commissions and the new model but I also have an announcement:

Travel and dailies

I'll be visiting my cousin for a celebration on the 27th of June and will be back on the 2nd of July. So depending on the time of arrival I'll be abroad for around 6 days. I'll try to finish as many commissions as possible before setting off and I'll soon post another poll for the week of June 23-29, so that I can create and schedule the daily posts in advance.

As the 30th of June and 1st and 2nd of July I still won't be able to create content (I'm not sure if I'll be back in the morning or at night on the 2nd), I'll pick characters from the same poll to create the dailies for those days. For the rest of the week I'll use the results from the poll that'll be releasing on the 25th of June.

Sorry if that sounds confusing I'll try to sum it up:- June 11 poll -> June 16-22 dailies.- Soon upcoming poll (either June 13 or 14) -> June 23 - July 2 dailies- June 25 poll -> July 3-6 dailies- July 2 poll might be postponed to July 3 depending on my arrival -> July 7-13 dailies.

And everything back to normal afterwards.

I would do a recommendation gathering post but it takes some time to gather everyone's recommendations and put it in a poll, and I have start working on the dailies ASAP in order to finish them by the time of me leaving. So I'll just do a recommendation post for the next poll.

Commissions

While still got some backlog I'm still working on them faster than I'm getting new ones, so hopefully soon enough I'll be back on schedule.On a quick note, I do get a few "non-conventional" commissions which do take a longer time to complete and I have to balance my time working on those and starting other commissions as well in order to maintain a fair waiting time for everyone, so I'm sorry if it takes a bit longer to finish some of your commissions!

New model

As I test and tweak the new model I'm trying to hopefully do my daily content in this model ASAP as it would help in a lot of ways moving forward. When I do however, I'll keep my old model on the PC with all the resources so anyone can commission me works in the default style as well if they prefer that one better.

Some of the benefits I found as of now using the new model:- Better prompt following, more pose variety- Better lighting/coloring- Less mistakes that need to be edited post generation- ~90% back compatible with resources made for the older model- Better style retention as of now (needs more testing, but looks promising)

Did some expression testing (Don't ask why Krul, I just picked a random recently used character :D). Every image was generated using the same prompts and generated once in a batch of 4. So basically I generated 4 images, picked the best one and used that.

"Expressionless" - neutral expression. My default style has a larger eye size in general, I've noticed that with multiple characters as well. I should be able to replicate that, I'll look into it.

Various common expressions (default style) - I was a little surprised about these results as the "smug, excited, smirk" expressions look almost the same in my default style. I usually combine these with half-closed eyes some added blush, at times I mix in some other expressions to better bring out the desired result, but for these tests I've kept prompting as simple as possible.

Various expressions (new model) - The new model on the other hand seems to reproduce these expressions a lot better even with simple prompts. I'll have to play around with the lighting however as while it is very dynamic and cinematic and all, it's not too great for these kind of tests...

Another set of common expressions (default style) - While the pouting/angry expression is quite similar I think they are pretty accurate. For the embarrassed expression I almost always add in some blush and half-closed eyes as well as a few other expressions like shy or worried, to get the desired results. For the gesugao I always used a resource to do it before, as it doesn't recognize it as a prompt in itself. I usually use a mixture of "uneven eyes, evil smile, parted lips" to do the expression, and that's what I did here as well, but I guess I need a resource to make uneven eyes. And adding all these resources will slightly change the artstyle as well, so that's not good.

Another set of common expressions (new model) - Similarly to the default style pouting/angry expressions are quite similar (I suppose pouting is a type of angry expression after all...), but it actually managed to pull off a pretty good gesugao expression on its own. While it's not that big of a deal that the default style can't reproduce it as I rarely use it, it's good to know that the new one can actually do it.

And that's all for this update. As always, feel free to tell me your opinions and feedback! Have a nice day, and see you all in the next post! o7