Formats & Codecs

Choosing the right source format and optimization codec is one of the most consequential decisions in show design. The codec determines GPU vs. CPU decode load, file size, scrub responsiveness, maximum resolution, and alpha channel support. This article covers every format WATCHOUT supports, how the optimizer converts between them, and when to use each option.

How the Optimizer Chooses an Output Codec

When a source file is added to the show, the optimizer does not simply re-encode it to HAP. Instead, it consults the codec mapping table — a configurable set of rules that maps each source codec to a specific output codec. The default mapping targets GPU-friendly formats:

You can override any row in this table through the Asset Manager Settings dialog. Non-default mappings are highlighted with a colored arrow.

Why HAP is Recommended

HAP is the default optimization target because it decodes entirely on the GPU using DXT texture compression:

• Minimal CPU usage — leaves the CPU free for effects, compositing, and control tasks.
• Instant scrubbing — every frame can be accessed independently (no inter-frame dependencies), so timeline navigation is smooth even with 8K content.
• High frame rates — easily handles 60 fps and beyond without CPU bottlenecks.
• Large resolutions — supports 4K, 8K, and higher. The GPU handles decompression in parallel with rendering.

The trade-off is larger file sizes and higher storage bandwidth requirements compared to CPU-decoded codecs. A 1-minute 4K ProRes video at ~2 GB may become ~8 GB as HAP. Plan storage and network capacity accordingly.

When to Use Other Codecs

• NotchLC — when you need GPU decoding with smaller files than HAP. NotchLC produces higher visual quality at comparable or smaller file sizes. Ideal when your source is ProRes.
• HEVC — when file size is the primary concern and you have GPUs with hardware HEVC decode. Seek performance is slightly worse than HAP due to GOP structure, but the size reduction is significant.
• Raw — when you need zero decode latency and have sufficient disk bandwidth. Raw formats are used for uncompressed texture streaming, typically for smaller assets or specialized pipelines.
• Pass-through — when your source is already in HAP, NotchLC, or Raw format. The optimizer copies without re-encoding, saving time and avoiding generational quality loss.

Alpha Channel Handling

Alpha (transparency) support depends on both the source and output codec:

When the source has an alpha channel and the output codec supports it, transparency is preserved through the entire pipeline. If you map a source with alpha to a codec that does not support alpha (e.g., standard HAP), the alpha channel is discarded during optimization.

The Optimization Pipeline

When a source file is added, it passes through these stages:

1. Upload — the source file is copied to the Asset Manager's internal storage.
2. Codec mapping lookup — the optimizer reads the source codec and consults the mapping table to determine the output codec.
3. Transcode — the source is decoded and re-encoded to the output codec at the configured quality level. Pass-through codecs skip this step.
4. Chunk storage — the optimized output is written as content-addressed chunks (Blake3 hashed). Identical data across assets is deduplicated.
5. Distribution — the optimized file is transferred to display servers when the show goes online.

Track Management

For source files that contain both video and audio tracks, the optimizer offers four track management modes (configured in Asset Manager Settings):

Bandwidth Limit

The Asset Manager Settings dialog includes a Bandwidth Limit setting (in Mbit/s) that caps the data rate used when transferring assets to display servers. Set to 0 for unlimited bandwidth. Use this when WATCHOUT shares a network with other traffic. See Asset Transfer for transfer optimization guidance.

Best Practices

• Deliver in the target codec. If possible, have content creators export directly in HAP or NotchLC. This eliminates transcoding time and avoids generational quality loss.
• Match frame rates. Source frame rate should match the show's frame rate (or be an exact multiple/divisor). Mismatched frame rates cause dropped or duplicated frames during playback.
• Use HAP Q for hero content. For content that will be displayed prominently (e.g., main stage, close-up screens), HAP Q provides visibly better quality than standard HAP at the cost of larger files.
• Reserve HEVC for constrained storage. HEVC produces much smaller files but requires hardware decode support on the GPU. Verify your display servers have compatible GPUs before committing to HEVC as the output codec.
• Test alpha pipelines end-to-end. Verify that the source → optimize → playback chain preserves alpha correctly by placing a transparent asset over a colored background and checking the result on a display server.

Troubleshooting