Page 1 of 1

Must see on Deviant Art

Posted: Thu Oct 08, 2026 6:33 pm
by qsemergency
I supect that this person must be a member of quicksand fans ....
but you should really check out his amazing AI Videos: https://www.deviantart.com/parselmompracem [url]https://www.deviantart.com/parselmompracem[/url]
His 4 minute video about a woman named Marietta is my favorite

Re: Must see on Deviant Art

Posted: Fri Oct 09, 2026 2:11 am
by SecretDude77
They have cohesive narratives like Viridian but I still think we’re far from being able to combine this level of dialogue and storytelling with the realistic settings and motion that Grok provides. It’s frustrating that even with all the options available, no LLM manages both realistic mud physics and believable sinking dialogue. Lately people are asking for more struggling in my clips or to make them longer at least. I’m capped at 30 seconds max and any extension after the first 6/10 slightly reduces video quality. It’s just not feasible. Grok isn’t free either but damned if I don’t get my moneys worth.

You unfortunately have to pick one or the other for now.

Re: Must see on Deviant Art

Posted: Fri Oct 09, 2026 3:45 am
by redjak6t4
There's another member of DA in parcelmompracem's Favourites who's into QS as well.

https://www.deviantart.com/gotchadeviant/gallery

Hadn't heard of them before today either.

So, good spot there qsemergency.

Re: Must see on Deviant Art

Posted: Fri Oct 09, 2026 4:11 am
by SecretDude77
redjak6t4 wrote:There's another member of DA in parcelmompracem's Favourites who's into QS as well.

https://www.deviantart.com/gotchadeviant/gallery

Hadn't heard of them before today either.

So, good spot there qsemergency.



I can tell their older stuff was Sora. I miss that app.

Re: Must see on Deviant Art

Posted: Fri Oct 09, 2026 2:53 pm
by Sekani
SecretDude77 wrote:They have cohesive narratives like Viridian but I still think we’re far from being able to combine this level of dialogue and storytelling with the realistic settings and motion that Grok provides. It’s frustrating that even with all the options available, no LLM manages both realistic mud physics and believable sinking dialogue. Lately people are asking for more struggling in my clips or to make them longer at least. I’m capped at 30 seconds max and any extension after the first 6/10 slightly reduces video quality. It’s just not feasible. Grok isn’t free either but damned if I don’t get my moneys worth.

You unfortunately have to pick one or the other for now.


Challenge accepted

Re: Must see on Deviant Art

Posted: Fri Oct 09, 2026 5:15 pm
by Viridian
I'm inclined to agree, though the upcoming release of Kling 4.0 might change that. Generally though, adding dialogue does fry the AI "brain". It almost always prioritises the dialogue over adhering to the prompt. However, it can be done.

How "realistic" the scene is will be subjective though. Here's my take on Kling 3.0.

Re: Must see on Deviant Art

Posted: Sat Oct 10, 2026 9:57 am
by MadMax359
looks like a useful tool (at least in your hands) :twisted:

Re: Must see on Deviant Art

Posted: Sun Oct 11, 2026 12:44 am
by Itself_Surprised
Sekani wrote:
SecretDude77 wrote:They have cohesive narratives like Viridian but I still think we’re far from being able to combine this level of dialogue and storytelling with the realistic settings and motion that Grok provides. It’s frustrating that even with all the options available, no LLM manages both realistic mud physics and believable sinking dialogue. Lately people are asking for more struggling in my clips or to make them longer at least. I’m capped at 30 seconds max and any extension after the first 6/10 slightly reduces video quality. It’s just not feasible. Grok isn’t free either but damned if I don’t get my moneys worth.

You unfortunately have to pick one or the other for now.


Challenge accepted


After roaming your page, we can also say that you've come far closer to it than you think. They're amazing.

Re: Must see on Deviant Art

Posted: Wed Oct 14, 2026 11:07 pm
by qsgroupie
Viridian wrote:I'm inclined to agree, though the upcoming release of Kling 4.0 might change that. Generally though, adding dialogue does fry the AI "brain". It almost always prioritises the dialogue over adhering to the prompt. However, it can be done.

How "realistic" the scene is will be subjective though. Here's my take on Kling 3.0.


I have found using Kling 3.0 Omni gives some really great results. You can place up to seven images or create your own elements for characters with unique voices and, as long as the prompt is detailed enough, can produce a high-quality video that matches up to the vision you are wanting to achieve.

It can be hit or miss, though, but a majority of the time, it comes out looking great.