[argyllcms] Re: Can Eye-one Pro failure the calibration?
- From: Sam Berry <samkberry@xxxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 5 Oct 2010 13:29:17 +0100
On 5 October 2010 13:09, Graeme Gill <graeme@xxxxxxxxxxxxx> wrote:
> Sam Berry wrote:
>
>> have a storage scope or similar device to give you proper numbers. I do
>> know
>> that PVA displays can take can take a long time (100ms+) to fall to black.
>> Given the 60ms delay through the frame buffers this is 160ms. CRTs also
>> show
>> significant persistance after reverting from high brightness. The bulk is
>> dropped of after only a few ms, but a trail remains for several hundred
>> ms.
>>
>
> Hmm. That's certainly a lot slower (an order of magnitude!) than the
> headline response times for LCD displays.
The headline figures that come to mind to most are those of TN displays,
which although fast are not really suitable for critical photographic use.
Photographic IPS and VA panels are noticeable slower. They also usually
these days quote a figure known as the grey-to-grey response, which is (a
bit like the CRI measurement) an average of the times taken to move over the
centre part of the s-curve brightness-time response. The figure does not
include any initial time to start the colour change; a display which takes 2
days to respond but then changes instantly gets a good figure :) The figure
also only needs to the display to get most of the way to the target colour.
The margin is nowhere near close enough for colourimetric work. The figure
is massively improved by overdriving the display, which only works for
greys, and is useless for white or black (or saturated colours). TN panels
are low to get to white, and PVA and IPS panels are slow to get to black.
Suffice to say the figures are for comparison purposes only :)
My display has an overdrive switch on the menu system, although the effect
is very obvious to the eye, it made only a small difference to the readings,
suggesting delay is the prime factor rather than the response time. I tried
it on my second display, and old NEC, and it showed some of the effect too,
although reduced.
>
>
> Given the small increase in total measuring time, I would suggest 200ms as
>> a
>> more conservative margin.
>>
>
> Yes, maybe that's needed to be safe. It does start to impact overall
> measurement
> time though (10 - 20% of the time will be settling time).
>
That is true. Perhaps a /fast switch for those happy to risk it? I repeated
the test with my DTP-94 and could not reproduce the effect. Perhaps the
longer integration times reduce the effect or perhaps it simply takes longer
to start measuring?
>
> Graeme Gill.
>
>
Sam Berry
www.satsumatree.co.uk
Other related posts:
