Repository navigation
JamtabaZap v3.0.0-beta.4
Pre-release
Pre-release
Beta. See the README for how to install it and its known limits.
Changes in 3.0.0-beta.4
- HD video no longer falls behind on Windows and Linux. The H.264 encoder in these builds lacked its processor optimizations, so on a laptop each HD frame took longer than 24 fps allows, frames were skipped and video lagged. These builds now include them, and the encoder uses every CPU core. On a two-core laptop an HD frame now takes about 11 ms instead of 55 ms.
- GPU video decoding works on Windows for streams it could not take before. NinjamZap and other real-time senders use the H.264 Baseline profile, which the GPU decoders rejected.
- Diagnostics: the log now records, once per interval, where each video frame's time goes (camera, conversion, encoding). It also records why a GPU decoder falls back to the CPU. Please attach log.txt when you report a video problem.
Changes in 3.0.0-beta.3
- Video no longer takes your second audio channel. Video is sent on its own channel after your audio channels. You can add and remove audio channels while the camera is on.
- Received video is decoded on the GPU when one is available (Windows: D3D11VA/DXVA2). If the GPU cannot decode a stream, that stream is decoded on the CPU. The macOS x86_64 build still decodes on the CPU.
- Video preferences are easier to find. All preferences tabs are visible at once, and the new Preferences > Video ... entry opens the Video tab directly, in the VST plugin too.
- Translations now cover the video and audio quality settings, video rooms and the SD notice in every language.
- The update notice no longer gets lost when the version check answers before the main window is ready.