Two-photometer test for LCD spatial timing

[GoogleCodeExporter]

LCD displays have become standard.  But their spatial and temporal 
characteristics are less obvious than those of CRT displays.  We would like to 
have a spatial test for characterizing LCD displays.

A simple test would fill the screen with alternating, distinct colors.  It 
would read voltage traces from two photometers, placed at different locations 
on the display.  Each photometer should have a fast response which varies 
monotonically with display brightness.  The test should compare the signals and 
signal phases from the photometers.

We would be interested in whether signal and signal phase vary with photometer 
position.

Original issue reported on code.google.com by Benjamin.Heasly on 27 Oct 2011 at 6:12

[GoogleCodeExporter]

It would be hard to write the whole test in a generic way, since different 
photometers would have different interfaces and data units and formats.

I think it's worth including a generic test of this sort, and leaving it up to 
users to hide photometer details in a couple of wrapper functions.

I'll include this in version 1.0, since it's not that big.

Original comment by Benjamin.Heasly on 8 Jan 2012 at 10:17

• Added labels: Priority-High
• Removed labels: Priority-Low

[GoogleCodeExporter]

Hi, one issue is that LCD timing depends on grey-level transitions, i.e. it 
depends on the first and second luminance to give a specific response time. A 
more accurate test would not use only two distinct colours but a defined set of 
luminance transitions. I think the luminance transition  inhomogeneity is far 
more significant and problematic for Neuroscience, and the luminance overshoot 
problems are a direct consequence of trying to compensate for these...

See Fig. 2 in http://www.mis.mpg.de/preprints/2010/preprint2010_33.pdf for an 
example.

Original comment by ian...@gmail.com on 9 Jan 2012 at 12:04

[GoogleCodeExporter]

That looks like an interesting paper.  Now I'm wondering if the topic is beyond 
the scope of Snow Dots.

Maybe the best thing Snow Dots could provide (and best payoff from limited 
development time) would be a utility for taking light readings at all.  It 
would be a working example that *does* make assumptions about the hardware 
being used.  Users could use it or modify it depending on the question they 
want to answer or the hardware they have.

Original comment by Benjamin.Heasly on 22 Mar 2012 at 6:32

[GoogleCodeExporter]

I looked at the AInScan1208FS class, with is in 
utilities/measurementComputing/.  This can read analog input waveforms and can 
be configured for one or more inputs channels and various scanning frequencies.

This seems like a good basis for taking light readings in Snow Dots.  The 
obvious missing piece is one or more photodiodes or other light meters.

I'm going to push down the priority of this issue until it's clear what else 
Snow Dots should take on.  Is there one single, canned test for LCDs that would 
be generally useful?

Anybody want to write one?

Original comment by Benjamin.Heasly on 26 Mar 2012 at 5:22

• Added labels: Priority-Low
• Removed labels: Priority-High