| [00:03:31] |
* zcorpan has joined #respimg |
| [00:06:51] |
* eeeps has quit (eeeps) |
| [00:11:11] |
* bruceontheloose has joined #respimg |
| [00:44:41] |
* marcosc has joined #respimg |
| [01:17:10] |
* yoav has joined #respimg |
| [01:17:18] |
* newtron has joined #respimg |
| [01:21:29] |
* joemcgill has joined #respimg |
| [02:32:17] |
* Wilto has joined #respimg |
| [03:18:02] |
* marcosc has joined #respimg |
| [04:33:50] |
* newtron has joined #respimg |
| [04:54:25] |
* yoav has joined #respimg |
| [05:14:35] |
* yoav has joined #respimg |
| [05:33:04] |
* marcosc has joined #respimg |
| [05:58:18] |
* newtron has joined #respimg |
| [05:59:59] |
<yoav> |
zcorpan: Thoughts on ? |
| [06:43:27] |
* joemcgill has joined #respimg |
| [06:59:22] |
* newtron has joined #respimg |
| [07:06:58] |
* zcorpan has joined #respimg |
| [07:18:36] |
* nemzes has joined #respimg |
| [07:21:40] |
* MarcDrummond has joined #respimg |
| [08:00:37] |
* newtron has joined #respimg |
| [08:23:07] |
* marcosc has joined #respimg |
| [08:25:05] |
* darobin has joined #respimg |
| [09:01:12] |
* newtron has joined #respimg |
| [10:01:58] |
* newtron has joined #respimg |
| [10:24:08] |
* newb has joined #respimg |
| [11:02:42] |
* newtron has joined #respimg |
| [11:19:51] |
* zcorpan has joined #respimg |
| [11:32:09] |
* Kolombiken has joined #respimg |
| [11:42:05] |
* zcorpan has joined #respimg |
| [11:52:30] |
* paul___irish has quit (Ping timeout: 180 seconds) |
| [11:58:40] |
* PicBot is now known as PicBo867 |
| [11:58:42] |
* PicBo867 is now known as PicBo851 |
| [11:58:44] |
* PicBo851 has joined #respimg |
| [11:58:46] |
* yoav has joined #respimg |
| [11:58:54] |
* Rossen_away has joined #respimg |
| [11:58:57] |
* adrianba has joined #respimg |
| [11:59:00] |
* igrigorik has joined #respimg |
| [11:59:27] |
* MikeSmith has joined #respimg |
| [11:59:29] |
* zcorpan has joined #respimg |
| [11:59:48] |
* jz has joined #respimg |
| [11:59:58] |
* paul___irish has joined #respimg |
| [12:00:20] |
* tn has joined #respimg |
| [12:00:55] |
* marcosc has joined #respimg |
| [12:01:22] |
* darobin has joined #respimg |
| [12:02:16] |
* Wilto has joined #respimg |
| [12:03:28] |
* newtron has joined #respimg |
| [12:04:00] |
* joemcgill has joined #respimg |
| [12:04:30] |
* zcorpan has joined #respimg |
| [12:14:08] |
* nemzes has joined #respimg |
| [12:14:15] |
* nemzes has joined #respimg |
| [12:15:52] |
* newtron has joined #respimg |
| [12:41:18] |
* Wilto has joined #respimg |
| [13:01:38] |
* marcosc has joined #respimg |
| [13:03:38] |
* newtron has joined #respimg |
| [13:32:53] |
* zcorpan has joined #respimg |
| [13:42:18] |
<yoav> |
landed |
| [13:43:03] |
<yoav> |
so images won't upscale for up to (and including) 1x DPR |
| [13:43:19] |
<yoav> |
we still need some research tho |
| [13:45:38] |
* zcorpan has joined #respimg |
| [13:51:36] |
* joemcgill has joined #respimg |
| [13:54:28] |
<newtron> |
yoav: a guy at my work offered to help get some research started |
| [13:54:44] |
<newtron> |
he'll work off the proposal i posted yesterday, unless anybody has feedback |
| [14:01:20] |
<yoav> |
newtron: Awesome! :) |
| [14:14:51] |
* MarcDrummond has joined #respimg |
| [14:26:45] |
* MarcDrummond has joined #respimg |
| [14:45:24] |
<newtron> |
we can get a user's DPR with JS, right? |
| [15:00:00] |
* joe_mcgill has joined #respimg |
| [15:02:44] |
* joemcgil_ has joined #respimg |
| [15:10:36] |
* marcosc has joined #respimg |
| [15:12:48] |
<yoav> |
newtron: window.devicePixelRatio IIRC |
| [15:13:06] |
<newtron> |
yoav: thx |
| [15:50:35] |
<yoav> |
newtron: I would add ssim scores to each phase in your survey |
| [15:51:13] |
<yoav> |
(as part of building the survey, rather than asking people) |
| [15:51:52] |
<yoav> |
(although, we could render them on canvas and calculate ssim on the user's machine</crazy>) |
| [15:53:50] |
<newtron> |
yoav: ya, i dssim as something to do a subgroup analysis on |
| [15:54:17] |
<newtron> |
i think the test cases can stay the same, but we can log the ssim/dssim for each pair and see if there's an effect |
| [15:54:59] |
<newtron> |
once i have the test images generated i'll make sure there's a decent enough spread of dssim scores or if i need to generate more test cases |
| [15:55:21] |
<newtron> |
(i'm doing this part because i'm a sucker; nik is going to build the interface for the survey and data collection) |
| [15:56:04] |
<yoav> |
very cool |
| [15:56:35] |
* zcorpan has joined #respimg |
| [16:35:39] |
* joemcgill has joined #respimg |
| [16:36:14] |
* joemcgill has joined #respimg |
| [16:37:00] |
* joemcgill has joined #respimg |
| [16:41:02] |
<zcorpan> |
gregwhitworth shared spartan's test results for - 6 fails, of which 2 are expected (since they only support x) |
| [16:42:05] |
<zcorpan> |
Fail "data:,a 1w 1h" |
| [16:42:05] |
<zcorpan> |
Fail "data:,a 1h 1w" |
| [16:42:05] |
<zcorpan> |
Fail "data:,a 0x" |
| [16:42:07] |
<zcorpan> |
Fail "data:,a -0x" |
| [16:42:09] |
<zcorpan> |
Fail "data:,a 1.x" |
| [16:42:11] |
<zcorpan> |
Fail "data:,a +1x" |
| [16:43:05] |
<zcorpan> |
yoav: r? (spartan correctly fails this btw) |
| [16:45:52] |
* eeeps has joined #respimg |
| [16:48:54] |
<yoav> |
looking |
| [16:49:07] |
<yoav> |
zcorpan: Did you see my question on ? |
| [16:49:15] |
<eeeps> |
newtron: would it be possible to vary not only on software (browsers) but also on hardware (devices)? |
| [16:49:55] |
<zcorpan> |
yoav: nopew |
| [16:49:55] |
<newtron> |
eeeps: you mean the devices on which the test images are generated, or the devices users are using to view the test images? |
| [16:50:04] |
<eeeps> |
for viewing |
| [16:50:23] |
* eeeps might be totally not understanding the test architecture |
| [16:50:27] |
<zcorpan> |
yoav: ... HTML4? |
| [16:51:17] |
<newtron> |
eeeps: ya, i think that makes sense |
| [16:51:21] |
<newtron> |
especially in terms of DPR |
| [16:51:35] |
<newtron> |
that's why i asked above about collecting DPR with JS |
| [16:51:53] |
<newtron> |
do you think it makes sense to ask the users what kind of device they're on? |
| [16:52:01] |
<yoav> |
zcorpan: Can you comment on the bug with the HTML5 link that makes this irrelevant so I can delete the irrelevant test? |
| [16:52:06] |
<zcorpan> |
yoav: yep |
| [16:52:28] |
<eeeps> |
newtron yep! then you can kind of work out the angle each pixel subtends on their retina based on dpr and a guestamate of the viewing distance for that device |
| [16:52:39] |
<eeeps> |
~guesstimate |
| [16:52:44] |
<newtron> |
eeeps: good idea |
| [16:53:14] |
<newtron> |
so just like phone/tablet/laptop/desktop/other? |
| [16:53:30] |
<newtron> |
(+tv, +other handheld?) |
| [16:54:07] |
<eeeps> |
yeah maybe. more detail = better, but seems like you have to strike a balance between getting more data and making the test cumbersome, and I have no idea where that lies |
| [16:55:05] |
<newtron> |
ya that's what i'd be worried about. i'd rather they spend time rating images instead of getting bored by the prelim questions |
| [16:55:09] |
<zcorpan> |
yoav: done |
| [16:55:31] |
<newtron> |
a broad device question + user agent string should be able to give us a pretty good idea in most cases |
| [16:55:55] |
<eeeps> |
newtron: *thumbs up* |
| [16:56:17] |
<newtron> |
+ maybe screen/viewport dimensions from JS too |
| [16:56:48] |
<eeeps> |
yep collect all of the “free” data that you can… |
| [16:58:27] |
<newtron> |
anybody have a windows machine + spartan and want to do some screenshots for me? |
| [16:58:50] |
<newtron> |
(there would be a lot…) |
| [17:01:50] |
* nemzes has joined #respimg |
| [17:06:06] |
* darobin has joined #respimg |
| [17:10:38] |
* bruceontheloose has joined #respimg |
| [17:14:43] |
* zcorpan has joined #respimg |
| [17:30:37] |
* darobin has joined #respimg |
| [17:33:01] |
* Wilto has joined #respimg |
| [17:35:25] |
* joemcgill has joined #respimg |
| [17:41:35] |
* Wilto has joined #respimg |
| [17:41:35] |
* Wilto has quit ("Lost terminal") |
| [17:41:47] |
* Wilto has joined #respimg |
| [18:24:29] |
* nemzes has joined #respimg |
| [18:25:07] |
* zcorpan has joined #respimg |
| [18:30:33] |
* joe_mcgill has joined #respimg |
| [18:42:50] |
* tn_ has joined #respimg |
| [18:44:04] |
* Wilto has joined #respimg |
| [19:04:30] |
* joemcgill has joined #respimg |
| [19:09:15] |
* darobin has joined #respimg |
| [19:16:07] |
* Wilto_ has joined #respimg |
| [19:23:29] |
* Wilto has joined #respimg |
| [19:30:29] |
* zcorpan has joined #respimg |
| [19:42:14] |
<yoav> |
and async image loading has landed: |
| [19:46:11] |
<Wilto> |
Niiiice. |
| [19:48:22] |
* bruceontheloose has joined #respimg |
| [20:00:09] |
* gregwhitworth has joined #respimg |
| [20:21:06] |
<newtron> |
yoav: like images are just loaded async by default with that change? |
| [20:21:54] |
<yoav> |
async in the sense that the load is not triggered synchronously, but using a microtask |
| [20:22:02] |
<yoav> |
just like Blink does it |
| [20:45:20] |
* joemcgill has joined #respimg |
| [21:17:36] |
* joemcgill has joined #respimg |
| [21:18:59] |
* nemzes has joined #respimg |
| [21:19:13] |
* joemcgill has joined #respimg |
| [21:19:45] |
* joemcgill has joined #respimg |
| [21:31:26] |
<newtron> |
cool |
| [21:35:54] |
* zcorpan has joined #respimg |
| [22:04:56] |
* joe_mcgill has joined #respimg |
| [22:23:16] |
* Wilto has joined #respimg |
| [23:03:08] |
* yoav has joined #respimg |
| [23:07:12] |
* bruceontheloose has joined #respimg |
