EDITORIAL POLICY

How we test

Every article on videobycode follows the rules on this page. They exist so a reader can repeat a test and check whether the result still holds.

1. Who writes here

Articles are written and published by the videobycode Editorial Team. They are not signed with individual names, and no article is presented as a personal story. The subject of an article is the test, not the person who ran it. Where an opinion is unavoidable, it is stated as an opinion and kept separate from the recorded results.

2. What we record on every test

Any article that reports a number states these four things, in the article itself, next to the result they belong to.

  • Tool and version. The exact release installed, including the runtime it ran on (for example the Node.js, Python, or FFmpeg build), so the same setup can be reconstructed.
  • Machine and operating system. CPU, memory, GPU where it matters, and the OS version. A render time without the machine behind it is not a measurement.
  • Render time. Wall-clock time from the start of the command to the finished file, measured on the output settings stated in the article, not the figure the tool prints.
  • Number of attempts. How many runs were needed to get a usable output, and what failed on the way.

3. Failures are published

  • When a tool does not complete a task, the article says so and shows how far it got. Tests are not quietly dropped because the outcome was unflattering.
  • Errors, workarounds, and version-specific breakage are written down with the message the tool produced, because that is usually the part a reader is searching for.
  • If a failure was caused by our own setup rather than the tool, the article says that too.

4. Comparisons

  • The criteria and the scoring are stated before the first result, and the same criteria are applied to every tool in the article.
  • Rankings follow the scores. A tool is not moved up or down for reasons outside the stated criteria.
  • Tools that have no affiliate or partner program are tested and scored on exactly the same basis as tools that do. See the affiliate disclosure.

5. Article types

Each article is labelled with one of these types, shown at the top of the page.

Measured
Numbers produced by running the tool ourselves, with the setup written down.
Tutorial
A task implemented step by step, with the code and the run time included.
Comparison
Tools scored on one fixed set of criteria, stated before the results.

6. Sources and figures

  • Numbers we did not produce ourselves carry a source next to them. Figures without a source are not published.
  • Numbers we did produce are labelled as our own measurement, and the conditions listed in section 2 are given with them.
  • Prices and limits taken from a vendor are dated, because they change without notice.

7. Updates and corrections

  • Articles that report measurements are revisited when the tool releases a version that affects the result. The update date is shown at the top of the article.
  • When something is wrong, the passage is corrected and a note describing the correction and its date is left at the end of the article. Articles are not deleted to resolve an error.

8. Use of generative AI

Generative AI may be used for drafting and for organising structure. It is never the source of a measurement. Before publication, the videobycode Editorial Team checks the full text against the recorded run output, and anything that cannot be traced to an actual run is removed.

9. Contact

Corrections, disputed results, and test requests: info@videobycode.com.