provide scripted test-scenarios so we can repeat tests and get comparable results
We should have more testing and testing should be reproducible and comparable etc.. Here is a basic test-sequence:
- Hans (only german) und Chad (only english) have a conversation
- two interpreters (DE->EN and EN->DE)
- settings:
| Person | floor-mic | Available Language | interpeter-Mic |
|---|---|---|---|
| Hans | on | DE | off |
| Chad | on | EN | off |
| interpreter DE->EN | off | floor | EN |
| interpreter EN->DE | off | floor | DE |
config of the client and specific version should be documented here
- sequence:
| # | Hans says | Chad says | Chad hears | Hans hears |
|---|---|---|---|---|
| 1 | Hallo! Wie geht es dir? | Hello! How are you? | ||
| 2 | Wie geht es dir? | |||
| 3 | Hi! I am fine thank you. | Hallo! Danke mir geht es gut. | ||
| 4 | I can even understand you! | Ich kann dich sogar verstehen! | ||
| 5 | Das war ja einfach. | That was easy. | ||
| 6 | Ich kann ja doch Deutsch. | Ich kann ja doch Deutsch. |
- at step 6 Hans hears the original audio, otherwise both hear the interpretation
- instead of having two interpreters, we could have one who quickly switches back and forth between the languages
I would like to document these cases in the wiki and i would like some feedback before i do. What am i missing? Would you contribute your test cases?
Activity
mentioned in issue #38 (closed)
marked this issue as related to #44 (closed)
I do not think, that this is a realistic test-case, just speaking a few words. Testing translation means, that somebody has to speak for some time in the floor - above 1 min. continuously like in a real sitation - and another person translates continuously. This would be for the translation of a conference presentation.
For translation of a discussion, there should be a real discussion, and the interpreters try to translate this according to the language spoken.
Test case: Language DE und EN
- Presenter: Speaks in DE for 2 min. - with slides
- Moderator: Speaks in EN language, and interrupts the presenter after 30 sek to ask a question
- Translator: Translates to both languages - and changes the output, when the moderator speaks in another language
- Listener: Listens to channel 1
- Listener 2: Listens to channel 2
- Listener 3: Jumps between the channels
That would be more realistic from my view
Edited by Zange Christian
Scenario with DE and EN tested successfully.
Lesson learned: microphone of translator was randomly selected. This caused the audio problems #55 (closed) #59 (closed) of some translators. Preselection of input device for additional SIP tracks to be fixed.
mentioned in issue #60 (closed)
