Preserve the failing command and configuration
When mpv changes behavior after customization, keep the exact failing input and installed build as your starting evidence. Record the operating system and package source as well as the version. A graphical application embedding mpv can add its own controls and defaults, so identify whether you are testing the mpv executable itself or another application.
Keep a private copy of the current configuration before editing it. Do not discard a carefully maintained setup simply to establish a baseline. Choose one authorized local file that reliably shows the problem and note the shortest action that triggers it, such as a seek, a subtitle selection or a change of output.
Run the developer’s configuration comparison
The mpv bug-report guide asks users to try --no-config before reporting a problem. Run the same executable against the same file with that option, keeping the environment and selected media constant. Quote your real file path according to the shell in use; a command copied from a website does not create its example file.
Observe whether the exact symptom remains. Record a positive or negative result without adding unrelated quality presets. If the issue disappears, the comparison points toward the custom option set. If it remains, retain that baseline too; it is useful evidence that the ordinary configuration was not sufficient to explain the behavior.
Reintroduce options one at a time
When the baseline works, mpv’s guide recommends adding options back individually until the smallest triggering set is found. Grouping dozens of settings into one test recreates the original uncertainty. Keep a short local table of the added option, the command and whether the symptom appeared.
Use the same media and trigger throughout. A different file or display output can change the result independently of the option under test. If two settings are jointly required, record that interaction as the smallest known set. Avoid declaring the last option generally broken when the failure depends on a particular combination or platform.
Prepare a reproducible and private report
A useful report includes the minimal command, relevant platform information and the exact action that fails. For display issues, mpv requests appropriate GPU and driver details. Remove private media URLs and tokens before sharing a command or log; a token-bearing input can be exposed through a terminal transcript.
If a sample is needed, use material you can lawfully share and verify that the sample still reproduces the issue. Keep the original configuration available for recovery. The expected outcome is a small, explainable reproduction, not a claim that the software was independently benchmarked or that every custom configuration is safe for every system.
Questions answered
What does it establish if --no-config removes the symptom?
It shows the custom option set is relevant to that reproduction. Add options back individually to identify the smallest triggering set.
Should I send a command containing a private stream token?
No. Review commands and logs before sharing and redact the private input. Use a lawful reproducible sample where possible.
Sources & further reading
Official references for the subjects discussed. Service terms and device software can change.