Help

Troubleshooting

Common issues and their solutions.

anelo requires a CUDA-compatible NVIDIA GPU or an AMD GPU with ROCm support. If the GPU is not detected:

Update your GPU drivers to the latest version
Verify CUDA installation by running nvidia-smi in a terminal (NVIDIA) or rocm-smi (AMD)
On laptops with hybrid graphics, ensure the discrete GPU is selected in your system settings
Check that your GPU has at least 6GB of VRAM

If processing fails with CUDA out-of-memory errors:

Close other GPU-intensive applications (games, other AI tools, browser hardware acceleration)
Reduce the processing resolution — try 1080p instead of 4K
Enable UHD mode for frame interpolation (tiles the frame to reduce memory usage)
If upscaling, try 2x instead of 4x

Expected processing times for 1 hour of 4K content:

If processing is significantly slower than these estimates:

Check that the GPU (not CPU) is being used — GPU utilization should be near 100% during AI stages
Ensure power management is set to maximum performance
Check that thermal throttling is not occurring (GPU temperature above 90°C will throttle)
Disable real-time antivirus scanning on the output directory

If the stereo output shows depth errors, shimmer, or ghosting:

Enable preflight scoring and increase the minimum fitness threshold to filter problematic frames
Try a different depth model — some content works better with MiDaS or ZoeDepth
Increase temporal smoothing to reduce inter-frame flicker
If the source is noisy or compressed, enable preprocessing denoise and deartifact stages
Reduce the stereo baseline if depth looks exaggerated

anelo uses FFmpeg for container and codec handling. Most common formats are supported. If a file is not recognized: