CPU3DAI tools for 3D, video and audio

Imagination-Studio-by-keys-Wasserman 4.4.1 brings back the 4.3.1 screens

Version 4.4.1 of the Imagination-Studio-by-keys-Wasserman tool has been released. According to the release description, at Wasserman's request the 4.3.1 screens have been restored, while everything underneath them remains from 4.4.0: native Windows, auto-update, Access, Finish and Premiere, keys-HDR, dual-GPU support, Sparks fixes, SPARK_PROGRESS_FILE and task preview. Removed from the screens are the color themes, the studio's simple mode, the Quick Edit tab, the "no sound" option, the new Library layout with pagination, and the 4.4.0 additions to the Home composer and the Windows pages. Separately, the Video Turbo adapter has been fixed: installation via oneshot-setup.sh --turbo failed with a 404 error because the copy of the Phosphene release had disappeared; now its own LightX2V file is downloaded with a pinned sha256.

What it means

The Imagination-Studio-by-keys-Wasserman tool page lists v4.3.1 as the latest release, while the repository is already at v4.4.1. According to our data, the tool is an open source video generator under the Apache-2.0 license, and the drama-3 model from Fish Audio handles text-to-speech. The model weights are distributed under the Fish Audio Community License: research, non-commercial use and commercial use with annual revenue under 500,000 USD are permitted; hosted speech generation services require a separate license. The largest model weights file is 3.5 GB, all files together are 38.6 GB. As described by the author, it needs 8.3 GB of VRAM and 128 GB of system RAM, and the platform is Metal on Apple-silicon Mac and NVIDIA DGX Spark. Release 4.4.1 concerns primarily the interface: the 4.3.1 screens are back, and some of the 4.4.0 innovations are removed from the visible part, although the mechanics underneath them remain. For a reader choosing a tool, this means the interface has returned to the 4.3.1 state, while the Video Turbo fix eliminates the installation failure on systems that use this adapter.
Imagination-Studio-by-keys-Wasserman 4.4.1 Screens 4.3.1 Restored on request of Wasserman Interface internals back to 4.3.1 Mechanics 4.4.0 Native Windows Auto-update, Access Finish, Premiere, keys-HDR Remained under screens Removed from UI Themes Simple studio mode Quick Edit, "mute" Library pagination, Home composer Video Turbo fix oneshot-setup.sh --turbo crashed with error 404 Now downloads LightX2V with pinned sha256 Requirements and license VRAM: 8.3 GB System RAM: 128 GB Metal on Apple-silicon Mac NVIDIA DGX Spark drama-3 model by Fish Audio — text-to-speech · Weights: 38.6 GB · License Apache-2.0 Fish Audio Community License: non-commercial and commercial up to 500,000 USD per year
How the method works. The diagram was drawn for this news note.

See also