[Galene] Camera test for Galene

[Galene] Camera test for Galene

[Juliusz Chroboczek]

Hi,

Valent wrote:

> I also like your idea of a test your microphone and camera button on the
> login page.

I've writtent a proof-of-concept for a camera and microphone test.
There's a live demo here:

  https://galene.org:8443/camera-test.html

and the sources live in the branch "camera-test" of the repository.  As
usual, this is 100% organic code written by a human being, an LLM was only
used for proofreading.

I'm not quite sure how to integrate it.  One option would be tu put the
"Request camera permissions" button on the login page, and have the UI
unroll when the user presses it.  Another option would be to open a new
tab when the user presses a "test camera" button, but I'm not sure how
well that would work on mobile.

Note that the code is integrated with Galene's settings, so if you choose
a given camera in the test page, this will be reflected in Galene's main
interface.

Things to do, if any of you has free time:

  - work on the CSS to make it pretty;
  - work on the CSS to make it better when the camera is switched off
    (microphone-only);
  - test under Firefox and Safari.

I've made a serious effort to get the error handling right (the state
should be properly reset if something goes wrong), but of course it's
always worth checking.

-- Juliusz

[Galene] Re: Camera test for Galene

[Gabriel Pruvost-Kerneis]

Hi,

On Fri, 24 Jul 2026, at 13:29, Juliusz Chroboczek wrote:
>
> Things to do, if any of you has free time:
>   - test under Firefox and Safari.

Basic functionality works in Firefox 153 on Android 13.

Things to improve: 
- the "enable" button is unclear. "Enable camera" or "enable camera+microphone"?
- Firefox displays a "play" button instead of the video before you click "enable", but hitting it does nothing. It's confusing. 
- CSS: the video is on the right of the screen (unclear if it's too large and doesn't fit or just not centered).
- the permissions are requested twice: once when clicking "request permissions" (for mic+camera) and another when clicking "enable" (for camera only). Probably because I didn't check "remember for this site" but still annoying. 

Best, 
Gabriel

[Galene] Re: Camera test for Galene

[Juliusz Chroboczek]

Thanks, Gabriel.

> - the "enable" button is unclear. "Enable camera" or "enable
>   camera+microphone"?

I think it should go away -- we should enable straight away once the user
has picked "Enable permissions".

> - Firefox displays a "play" button instead of the video before you click
>   "enable", but hitting it does nothing. It's confusing.

I agree.

> - CSS: the video is on the right of the screen (unclear if it's too large and doesn't fit or just not centered).

It needs some CSS love.  Actually, all of Galene needs some CSS love.

(The current UI was written in a rush, with the aim of having something
acceptable for the start of teaching season during Covid.  I think the
CSS pretty much needs to be rewritten from scratch.)

> - the permissions are requested twice: once when clicking "request permissions" (for mic+camera) and another when clicking "enable" (for camera only). Probably because I didn't check "remember for this site" but still annoying. 

Yeah, I'm not sure how to work around this particular Firefox quirk.

-- Juliusz

[Galene] Re: Camera test for Galene

[Juliusz Chroboczek]

I've played with it some more, and here's the current plan for the UI.

Below the login dialog, there's what Mac people call a "disclosure
indicator":

    ▶ Test camera and microphone

When the user clicks on the triangle, the browser produces its permissions
dialog.  If the user rejects the permission dialog, we remain in the
folded state.  If the user accepts, the indicator unfolds (▼), and the UI
for the video test is shown.  If the user clicks on the triangle again,
the camera is released, and we return to the initial state.

Do people feel that's discoverable enough for a naive user?

Related question, do people feel that the Fourier transform (spectrograph)
display is too geeky, and that I should replace it with a simple volume
indicator (0% — 100%)?

I also think I need to display a big red message that says "no audio, your
microphone might be muted", but I need to do some experimenting in order
to find out how to determine that reliably from the Fourier transform.
(I'm currently requesting the browser to compute the Fourier transform,
not the RMS, and I'd rather not be computing both.)

-- Juliusz