Lately I've recognized this direction

Lately I've recognized this direction on the rise and been inspired to create a collection of 18 images blending the Classical Art - Renaissance oil painting mastery with digital pixel/ASCII aesthetics, Classcii. It is a mix of centuries, a modernized and futuristic direction of the classic.

Each image is crafted around a specific tension, the classical figure or scene rendered in full oil mastery, while the digital/ASCII layer intrudes at a specific point.

Using the @Contra Labs ecosystem to choose the most adequate Ai models from the Creative Arena Leaderboard, I wanted to briefly share my personal taste and evaluation on the outputs these models generated, based on the benhchmark methodology.

For the imagery references I used Seedream 5 Lite and Google Nano Banana 2, whereas for the Image to Video I used Kling 3.0 and Veo 3.1 Fast.

In terms of Visual quality & aesthetics overall I would rank them fairly simillar with Seedream having a slight edge on the composition, specfically achieving more symmetrical outcomes featuring center alignment and a better use of whitespace, opposed to Nano Banana 2 where the use of rule of thirds is more prominent, making it visually in comparison less balanced.

Considering the Prompt adherence & accuracy, I find both models very much aligned, having Nano Banana 2 more intensified use of asscii / pixels across all outcomes, which according to my minimal taste I would rank this a weakness rather than a strong side.

Finally, in terms of Motion Realism I would unanimously rank Kling 3.0 as the better option here, considering the fluidness and physics consistency among majority of the elements in the reference. Additional advantage will be the ability to perform better at no prompt insertion compared to Veo 3.1 fast. You can download the collection here.