search

EdgeTX / edgetx-sdcard-sounds

Sound packs for EdgeTX
GNU General Public License v2.0
82 stars 48 forks source link

Developer docs: a few questions to get started contributing to this repository (engineering, with first-hand localization experience) #88

Closed gonzalo-bulnes closed 6 months ago

gonzalo-bulnes commented 6 months ago

Hi @pfeerick (and everybody else who can answers these questions!)

I'm interested in contributing some thinking and tooling to this repository. I see opportunities to reduce duplication, automate further and make the job easier for people interested in providing or correcting sounds in their own language (including by contributing translation docs). But I'd like to make sure I understand how this repo relates to the other EdgeTX projects before making the wrong assumptions.

Can I check my understanding of the goals with you?

  1. This repository releases include:

    • one archive per language-locale-voice, with the contents of the /SOUNDS/<language code> directory of the EdgeTX SD Card.
    • one sounds.json file
    • some other archives that don't follow a strict naming pattern, but also expand to the /SOUNDS/<some language code> directory of the EdgeTX SD Card.

    1. Who consumes sounds.json? I don't see it used in the SD Card, is it retrieved from the releases page by Companion or Buddy?

    2. Regarding locales and voices: my understanding is that EdgeTX is not currently aware of the locale and voice, only the language. By that I mean: I can place the contents of the en_GB-Libby archive in my SD Card's /SOUNDS/en directory and the contents of es_CL-Catalina in /SOUNDS/es. That will allow me to switch the voice language from English to Spanish from the radio settings, but I can't have en_US-Ryan at the same time, unless I hack my way around (e.g. my placing the contents of the archive in /SOUNDS/de and selecting German. Is my understanding correct?

    3. The files in /SOUNDS/SYSTEM are directly referenced by EdgeTX code, and as such are required for a language pack to be complete. But the files directly at the root of /SOUNDS are not, and can be considered optional. Is that true?

    4. Similarly, the files in /SOUNDS/INAV are required for the INAV language pack to be complete, but the script does not rely on any file outside of that directory. Is that also true? (Or SHOULD it be? I understand that depends in part on what the maintainers of the INAV and other scripts do.)

    5. Companion and Buddy rely on the GitHub Releases page directly to retrieve language packs and information about them, so the content of the Assets section of each release can be thought of as the interface of this repository with Companion and Buddy. Is that correct? (I'm half-guessing here.)

  2. There MUST be an archive named edgetx-sdcard-sounds-<language code>-<edgetx version>.zip for each of the supported languages for Companion and Buddy to function as expected. But the archives that follow a different naming pattern (e.g. edgetx-sdcard-sounds-es_cl-catalina-2.9.0.zip) are only meant to be downloaded by individuals who visit the Releases page. Is that correct? (Again, I'm guessing here.)

  3. If you've got pointers to the EdgeTX code that loads or consumes the content of the /SOUNDS directory, I'd welcome them! In particular, I'm interested in understanding if / how the first column (ID) of the CSV files is currently used (or not used - see also https://github.com/EdgeTX/edgetx-sdcard-sounds/issues/48#issuecomment-1873611492).

pfeerick commented 6 months ago
  1. Yes, Companion uses that to identify the packs and where they go. I believe Buddy does also.
  2. Correct, the firmware currently only recognises as single two letter folder, and specific ones that that. One of the long term goals is to remove that requirement.
  3. /SOUNDS/SYSTEM is used by the firmware itself, /SOUNDS are the user presented list of sounds, available for use by special functions, etc.
  4. This was added in order to allow the end user to replace voices included with iNav, BetaFlight and/or YAAPU telemetry scripts, for consistency with the rest of the radio phases and sounds. It is entirely up to the developers of those scripts if they use them natively or if end user has to replace the files as needed.
  5. Yes, the releases page is the interface for Companion/Buddy
  6. No. The releases page is the sole place the archives are exposed, so Companion must use a combination of the sounds.json and the list of files present in a given release.
  7. For the firmware, have a look at the tts_##.cpp files in https://github.com/EdgeTX/edgetx/tree/main/radio/src/translations For Companion, have a look at https://github.com/EdgeTX/edgetx/blob/main/companion/src/updates/updatesounds.cpp

Feel free to make any suggestions/changes to the edgetx-sdcard-sounds structuring that help with moving towards something like crowdin (for machine translation and ease of non-developer translator access) and doesn't completely tie us into a specific TTS service. Anything that touches on the firmware/buddy/companion will take longer, as the current system, while not optimal, works, so isn't as high a priority. But any assistance from someone who actually knows what they are doing would be greatly appreciated :grin:

gonzalo-bulnes commented 6 months ago

I'll archive/close this question thread BTW!