IRC log of audio on 2012-06-13

Timestamps are in UTC.

18:48:40 [RRSAgent]
RRSAgent has joined #audio
18:48:40 [RRSAgent]
logging to
18:48:42 [trackbot]
RRSAgent, make logs world
18:48:42 [Zakim]
Zakim has joined #audio
18:48:44 [trackbot]
Zakim, this will be 28346
18:48:44 [Zakim]
ok, trackbot; I see RWC_Audio()3:00PM scheduled to start in 12 minutes
18:48:45 [trackbot]
Meeting: Audio Working Group Teleconference
18:48:45 [trackbot]
Date: 13 June 2012
18:49:12 [ot]
18:49:19 [ot]
Agenda+ Testing
18:49:36 [ot]
Agenda+ Loop start/stop points (Bugzilla 17390)
18:49:43 [ot]
Agenda+ rechartering update
18:50:11 [ot]
Agenda+ MIDI issues (Bugzilla 17416 and 17417)
18:53:09 [ot]
ot has joined #audio
18:58:06 [Zakim]
RWC_Audio()3:00PM has now started
18:58:14 [Zakim]
+ +1.862.201.aaaa
18:59:51 [cwilso]
cwilso has joined #audio
19:00:16 [joe]
joe has joined #audio
19:00:25 [ot]
zakim, code?
19:00:25 [Zakim]
the conference code is 28346 (tel:+1.617.761.6200, ot
19:00:41 [Zakim]
19:00:49 [ot]
zakim, ??P9 is me
19:00:49 [Zakim]
+ot; got it
19:00:58 [cwilso]
zakim, who is here?
19:00:58 [Zakim]
On the phone I see +1.862.201.aaaa, ot
19:00:59 [Zakim]
On IRC I see joe, cwilso, ot, Zakim, RRSAgent, MikeSmith, automata, shepazu, kennyluck, trackbot, philipj, paul___irish
19:01:08 [cwilso]
zakim, +1.862 is me
19:01:18 [Zakim]
+cwilso; got it
19:01:26 [Zakim]
+ +1.978.314.aabb
19:01:41 [Zakim]
+ +1.510.334.aacc
19:02:40 [ot]
zakim, aabc is joe
19:02:40 [Zakim]
sorry, ot, I do not recognize a party named 'aabc'
19:02:52 [ot]
zakim, aabb is joe
19:02:52 [Zakim]
+joe; got it
19:03:08 [Zakim]
19:03:14 [ot]
zakim, aacc is crogers
19:03:14 [Zakim]
+crogers; got it
19:04:16 [Joebob]
Joebob has joined #audio
19:04:19 [jussi]
jussi has joined #audio
19:04:57 [chris_]
chris_ has joined #audio
19:05:21 [Zakim]
19:05:51 [jussi]
Zakim, ??P14 is me
19:05:51 [Zakim]
+jussi; got it
19:05:59 [jussi]
Zakim, mute me
19:05:59 [Zakim]
jussi should now be muted
19:08:02 [ot]
zakim, take up agendum 1
19:08:02 [Zakim]
agendum 1. "Testing" taken up [from ot]
19:08:26 [Zakim]
19:08:50 [Zakim]
19:08:52 [tmichel]
tmichel has joined #audio
19:10:23 [ot]
Olivier: several conversations about testing, now seems like a good time to start discussing our approach, agree on a strategy and get a test lead (for the group or for each deliverable)
19:10:37 [chris_]
19:10:45 [ot]
Doug: testing is very helpful to help where specs are unclear
19:10:53 [ot]
… also important when you start having several implementations
19:11:17 [ot]
… it has been useful in other groups to have a "test lead" who helps coordinate the test effort
19:11:30 [Zakim]
19:11:34 [ot]
… may be the person who makes the tests, or extracts test requirements from the spec
19:11:44 [ot]
… or encourages other people to add them
19:11:56 [ot]
… can think of the test coordinator as the editor for the test
19:12:02 [ot]
… will need one for each spec
19:12:21 [ot]
… for the midi API it is likely that one of the two editors can manage the tests
19:12:47 [ot]
for the web audio API, Chris has a lot of work with the spec, it may be best to have another person in charge of the testing
19:13:18 [ot]
doug: it also engages more people, and adds checks and balance (Chris is not in charge of everything and wielding too much power)
19:13:48 [ot]
doug: do we have any volunteer to be test controller for each of the spec
19:14:03 [ot]
crogers: we already have over 60 tests written ->
19:14:10 [ot]
… any of those can be shared
19:14:34 [ot]
doug: that's good. one bit of value in the tests is that the implementors can use them to test their own build
19:14:57 [ot]
… it's importance for the spec interoperability, but in practice it's useful to test the implementations
19:15:07 [ot]
… so it's helping when each implementor contributes tests
19:15:18 [ot]
… are these unit tests?
19:15:36 [ot]
crogers: fairly basic tests, testing the gain node, etc
19:15:40 [ot]
… pretty targeted
19:15:54 [ot]
doug: will need hundreds, maybe thousands
19:17:30 [ot]
Olivier: I've seen simple specs with thousands of tests, testing combination and parameters… those tend to be automated
19:18:13 [ot]
doug: w3c is developing a test framework, a test harness
19:18:33 [jussi]
19:18:59 [ot]
19:19:08 [Zakim]
19:19:11 [ot]
19:19:20 [Zakim]
19:20:23 [rtoyg]
rtoyg has joined #audio
19:21:28 [ot]
Olivier: wondering if the webkit tests look like the method mentioned by Philip?
19:21:44 [ot]
crogers: very similar to pixel tests
19:23:23 [ot]
Olivier: how do you cater for the fact that you don't get a predetermined signal as output
19:23:55 [ot]
crogers: a human being has to generate the baseline for each rendering engine
19:24:55 [ot]
… the tests will be shared, but the baseline may not
19:24:58 [Zakim]
19:25:56 [tmichel]
zakim, who is here ?
19:25:56 [Zakim]
On the phone I see cwilso, ot, joe, Doug_Schepers, jussi (muted), ??P17
19:25:58 [Zakim]
On IRC I see rtoyg, tmichel, chris_, jussi, joe, cwilso, ot, Zakim, RRSAgent, MikeSmith, automata, shepazu, kennyluck, trackbot, philipj, paul___irish
19:25:58 [Zakim]
+ +1.650.214.aadd
19:26:07 [ot]
zakim, aadd is crogers
19:26:07 [Zakim]
+crogers; got it
19:26:41 [ot]
crogers: a majority of our tests don't compare against a ref file
19:27:04 [ot]
… what they do is render audio internally and use js to walk the rendered data and verify that it is correct
19:27:09 [ot]
… with a very small tolerance
19:27:24 [ot]
… many of those tests can be shared
19:29:50 [ot]
olivier: did you have pixel tests with svg, doug?
19:30:03 [ot]
doug: probably similar to what is called ref test in css
19:30:36 [ot]
… a reference view, sometimes reproduced with a combination of simpler features
19:31:03 [ot]
… in svg, we actually had reference images, and we did the testing manually
19:31:21 [ot]
… hundreds of tests and we had f2f meetings where we ran tests together in a room
19:31:32 [ot]
… also important for competitive reasons
19:31:43 [ot]
… for svg2 we plan on doing web tests
19:31:59 [ot]
crogers: a majority of these tests can be automated on baseline
19:32:03 [shepazu]
19:32:54 [joe]
19:33:14 [ot]
olivier: similar to the graph that Philip sent today
19:33:36 [ot]
crogers: not for oscillators, but for e.g biquad, yes
19:33:46 [ot]
… similar to rendering lines
19:33:52 [ot]
… lines will not be rendered exactly
19:33:59 [ot]
… the same
19:34:11 [ot]
… all are an approximation of a mathematical ideal
19:34:34 [ot]
… most of the audio stuff is not that complicated, it's actually precise
19:35:00 [shepazu]
19:35:00 [ot]
ack she
19:35:15 [ot]
ack joe
19:35:26 [ot]
joe: wanted to share a useful testing technique
19:36:30 [ot]
… we alternate between the baseline and the test
19:36:39 [ot]
crogers: you mean a diff?
19:37:17 [ot]
… might not work so well for web audio
19:38:20 [ot]
doug: I like the idea of having a baseline
19:39:30 [ot]
crogers: 2 kinds of tests
19:39:42 [ot]
… one fully automated with js walking the data and comparing
19:39:56 [ot]
… the other compares the generated audio with a baseline
19:40:12 [ot]
… bit-exact compariso
19:40:17 [ot]
19:40:35 [ot]
doug: is there some functionality in the API enabling this?
19:40:53 [ot]
crogers: we had to build an extra tool in our test harness
19:41:37 [ot]
doug: would that be useful for some audio analysis use cases
19:42:19 [ot]
crogers: probably not because in a normal environment you can generate the wav file and upload it to a server via xhr
19:42:37 [ot]
… we just have a specific hook in our test harness
19:42:47 [ot]
… details would be different in another test tool
19:44:26 [ot]
ot has joined #audio
19:45:33 [Zakim]
19:46:00 [jussi]
zakim, unmute me
19:46:00 [Zakim]
jussi should no longer be muted
19:46:04 [ot]
jussi, you've posted a link to some tests
19:46:27 [ot]
jussi: testing that the interfaces are there
19:46:53 [jussi]
19:47:46 [ot]
jussi: not sure about more complex tests
19:47:58 [ot]
… don't think you can do automated test
19:48:05 [ot]
… you'd need a tester with MIDI device
19:48:59 [ot]
ot: could that be "faked"?
19:49:06 [ot]
jussi: a virtual midi device could work
19:49:24 [ot]
cwilson: would be hard to make cross-platform, obviously
19:49:46 [ot]
jussi: I think OSX lets you do that
19:49:47 [ot]
crogers: yes I think you can do that
19:50:06 [ot]
… harder to write that test than to write the MIDI implementation itself
19:50:27 [ot]
jussi: I had a virtual device in my original proposal, decided to keep it out of v1
19:50:38 [ot]
… some platforms don't support it at all.
19:51:57 [ot]
Olivier: I will send a call for volunteers on the list, to see who wants to be in charge of testing for each of the specs
19:52:15 [ot]
jussi: if noone else volunteers I can be in charge for the MIDI API, for now
19:52:45 [ot]
zakim, take up agendum 3
19:52:45 [Zakim]
agendum 3. "rechartering update" taken up [from ot]
19:53:11 [ot]
19:55:05 [Zakim]
19:55:06 [Zakim]
19:55:06 [Zakim]
19:55:09 [Zakim]
19:55:11 [Zakim]
19:55:15 [Zakim]
19:55:16 [Zakim]
RWC_Audio()3:00PM has ended
19:55:16 [Zakim]
Attendees were +1.862.201.aaaa, ot, cwilso, +1.978.314.aabb, +1.510.334.aacc, joe, Doug_Schepers, crogers, jussi, +1.650.214.aadd
19:55:42 [ot]
rrsagent, draft minutes
19:55:42 [RRSAgent]
I have made the request to generate ot
19:55:49 [ot]
zakim, bye
19:55:49 [Zakim]
Zakim has left #audio
19:55:53 [ot]
rrsagent, bye
19:55:53 [RRSAgent]
I see no action items