In Shotcut 25.12 and later, choose a processing mode after checking the actual filter chain. Native 10-bit CPU is a faster candidate when the required filters are available in the 10-bit path; it is not the most flexible filter option. The official documentation describes Native 8-bit CPU as more flexible because it includes more video filters, but using an 8-bit filter in a 10-bit project reduces the bit depth of the affected clip or track. Choose Linear 10-bit CPU when linear pixel mixing matters enough to justify slower processing; consider Linear 10-bit GPU/CPU only if you accept an option labeled experimental. Processing mode does not determine the exported file’s bit depth.
Decide from the actual filter chain
Shotcut’s Processing Mode controls image processing after a video or image is read and before it is displayed or exported. The official Processing Mode documentation distinguishes Native processing from Linear processing and describes the performance and compatibility tradeoffs. It also notes that many, but not all, CPU video filters have been converted or marked safe for 10-bit projects.
In Shotcut 25.12 and later, use this sequence before committing to a mode. The Processing Mode menu was added in 25.12 and replaced Settings > GPU Effects. Earlier versions have different menus and processing behavior, so do not assume these options or comparisons apply unchanged.
- In the Shotcut version you are using, open the filter menu and identify the filters, transitions and track blending used in the project. Check the filter’s type label, such as
#10bit, rather than judging support by a node or icon color. Colors are only clues; the label is the relevant indicator. - Mark each operation that interpolates or blends pixels: examples include scaling, blur, sub-pixel positioning and track compositing. These are the cases for which Linear processing may be important.
- Check whether each selected filter is suitable for the intended 10-bit or GPU pipeline. If an essential filter is available only through an 8-bit path, record that as a quality risk rather than assuming the project remains fully 10-bit throughout processing.
- Choose Native 10-bit CPU when compatibility and speed are the priority and the expected pixel-mixing result is acceptable. Choose Linear 10-bit CPU when linear mixing matters enough to justify slower processing. Treat the GPU/CPU Linear option as experimental, not as a guaranteed performance improvement.
- Compare representative timeline sections in the candidate mode before settling on the project setting. Include the most demanding blend, scale or blur, and any filter with uncertain support.
The official explanation of linear color processing says gamma-encoded pixel mixing can cause color distortion and is technically incorrect. Linear processing addresses that mixing behavior, but the documentation also warns that it is slower because it involves more conversions.
Hypothetical filter worksheet and acceptance check
The following is a decision exercise for an imagined short sequence, not a test result or a prediction of Shotcut’s rendered appearance. The observations are proposed checks for the editor to perform.
| Hypothetical operation | Mixing or support concern | Candidate decision | Observation to record during review |
|---|---|---|---|
10-bit interview clip with a #10bit color filter | Filter is labeled for 10-bit; no scaling in this section | Native 10-bit CPU as a faster candidate, provided the required filter is available in that path | Whether the intended tone adjustment remains acceptable in the displayed comparison |
| Title over footage using track compositing | Composite blends pixels | Compare Native and Linear 10-bit CPU | Whether edge appearance and color around the title are acceptable in the chosen mode |
| 10-bit clip scaled and blurred | Scaling and blur interpolate pixels | Include Linear 10-bit CPU in the comparison | Whether the result is preferable for this shot, and whether processing delay is tolerable |
| Required effect has an 8-bit filter path | Using it in a 10-bit project reduces bit depth for that clip or track | Keep as an explicit exception, or find a suitable alternative | Whether the resulting image shows unacceptable banding or other visible degradation at delivery viewing size |
Accept a mode only after reviewing the relevant sections at consistent playback conditions and checking that essential filters are available in the chosen path. Record exceptions and unresolved visual concerns; the table’s checks do not guarantee a particular result.
Limits and what to do if it fails
A 10-bit processing mode does not, by itself, guarantee a 10-bit export. Processing and export encoding are separate decisions: select and verify the delivery format and bit depth in the export workflow as well. The official documentation describes processing before display or export; it does not establish that every filter or export format preserves 10-bit precision end to end.
If Linear processing is too slow, try Native 10-bit CPU and review the sections with the most pixel mixing. If an 8-bit filter is unavoidable, test its output on the affected clip or track and decide whether the visible compromise is acceptable; the documentation warns of reduced bit depth but does not prescribe a universal severity threshold. If the experimental GPU/CPU path behaves unexpectedly, compare with the CPU path and keep the experimental result only if it meets the project’s requirements. When support remains unclear, consult the filter’s actual type label and compare a short representative section rather than inferring support from appearance alone.
Common questions
Is Native 10-bit the same as exporting a 10-bit file?
No. Native 10-bit names a processing mode. Export bit depth depends on the selected export format and settings and must be checked separately.
Does every 10-bit project need Linear processing?
No. Consider Linear when pixel-mixing operations are significant and their results justify slower processing. For other chains, Native 10-bit may be the more practical choice.
Can I use an 8-bit filter in a 10-bit project?
The documentation says you can, but the affected clip or track loses bit depth through that filter path. Mark the exception and inspect the result; do not treat the project’s 10-bit mode as proof that all filters are 10-bit.
Sources and next step
This guide interprets public documentation and uses explicitly hypothetical acceptance examples, not application tests. Controls and availability can change; verify the cited manual against your installed version.
For work created in Imagild, carry the input requirements and acceptance checklist into the film workspace. The external software controls discussed above belong to their respective products.