WebAudioEvaluationTool icon indicating copy to clipboard operation
WebAudioEvaluationTool copied to clipboard

Is it possible to create a listening test of a wav selected upon user indirect choice?

Open asapsmc opened this issue 7 years ago • 5 comments
trafficstars

Hi! I'd like to know if there's an "easy" way to create a listening test where the user makes a group of options, then (in background) the corresponding wav file is selected and played for the listening evaluation? Thanks in advance

asapsmc avatar Jun 20 '18 17:06 asapsmc

Hi there! Any comment on this? Thanks.

asapsmc avatar Jun 28 '18 06:06 asapsmc

Could you explain what you mean by 'make a group of options'? What you describe seems like the basic functionality of the tool, so I'm sure I'm missing something. Thanks for your interest in our tool!

BrechtDeMan avatar Jul 05 '18 13:07 BrechtDeMan

I'm pretty sure your tool is ready for my use-case, so I'd guess I'm the one missing something about your tool. Let me try a better explanation about my listening experiment:

Full Sequence- For each audio file (let's call it source) in a pool of files I need the following:

  1. The user listens to it and selects (e.g 2 comboboxes) overall characteristics of what he heard (e.g is tempo slow or fast);
  2. based on that characterization and on some additional logic, the "target" audio file is selected;
  3. the user listens to it and performs a subjective evaluation (4 or 5 likert-type predetermined question/answers)
  4. the results (to save) of my full sequence will be: a) the choices made in 1); b) the "target" audio file in 2; c)the subjective evaluation of 3.

This was what I intended to ask in the first place. To seize this opportunity, to fully describe the experiment:

  • I would like to run "Full Sequence" (at least) twice for each subject (listener), in a randomized order (in relation to the order of source audio selection).

Thanks in advance.

asapsmc avatar Jul 05 '18 15:07 asapsmc

Hi. It can work, but not out of the box. Here's how:

  1. Normal listening test, each user selects 2 options on characteristics
  2. Upon submission, server side script can analyse and produce a new test XML with the desired target audio
  3. Listener performs new test and submits results

So you'd need to do step 2. We have a PHP server stack for handling saves, which you could modify to perform the automatic steps. Some modification of the core.js would also be required to automatically continue to the next step as well.

Another way is to get the user to click a link after completion which, through PHP Session cookies, could track the original submission file and create the new file (this way is probably easier to do, but if someone has cookies disabled then it may cause problems unless you alter core.js for other things).

nickjillings avatar Jul 09 '18 15:07 nickjillings

Thank you so much for the answer. Well, at first glance it seems complicated, as this would be for 20 randomised files/pages, plus another randomised full run... Best.

asapsmc avatar Jul 09 '18 16:07 asapsmc