Re: Humans (and cameras and scanners) do not have a color gamut (?)
Re: Humans (and cameras and scanners) do not have a color gamut (?)
- Subject: Re: Humans (and cameras and scanners) do not have a color gamut (?)
- From: edmund ronald via colorsync-users <email@hidden>
- Date: Wed, 8 Jan 2020 08:13:02 +0100
BTW I think this type of discussion is quite helpful because it exposes all
sorts of computer geeks and photographers who lurk here to the idea that
what they are doing is twiddling the buttons on some very cumbersome tech
that is quite distant from what humans perceive. This can be demoed easily
when one takes a multispectral capture, and suddenly realises that although
a lot of information has been stored it is hard to transform this back
into valid dense-impressions.
Edmund
On Wed, Jan 8, 2020 at 7:54 AM edmund ronald <email@hidden> wrote:
> I don’t see what a camera really has to do with color; the word
> colorimetric doesn’t really come into it except rather late because
> colorimetry starts out as a way of making more precise the psychophysics of
> color perception. As a camera can does not create a color impression per
> se, we would have to create a whole capture and rendering environment
> before we can talk about colorimetry related to that camera, even if we
> assume a “standard observer” - such observer being themselves a fiction
> created by the priests of ... colorimetry.
>
> Disclaimer, my knowledge of the topic only comes from talking Dr. Hunt’s
> introductory course, and of course I am just an amateur scientist, and not
> as smart as Iliah or Andrew.
>
> Edmund
>
>
> On Tue, Jan 7, 2020 at 11:35 PM Iliah Borg via colorsync-users <
> email@hidden> wrote:
>
>>
>> On Jan 7, 2020, at 5:12 PM, Henry Davis via colorsync-users wrote:
>>
>> > This part is still confusing:
>> >> The closer the transform from camera observer to human observer is to
>> one-to-one, the easier it is to speak of “colour accuracy”.
>> > It just seems that there has to be limitations within the response of
>> the camera.
>>
>> Yes, measurement range and measurement accuracy impose limitations.
>> Convolution of spectrum to 3-channel data numbers impose limitations.
>> Measurement accuracy and ambiguity of convolution varies with the source
>> colour (the colour presented to a camera).
>>
>> > So, don’t call it a gamut - but if there is some limitation what would
>> it be called.
>>
>> Accuracy.
>>
>> > Is the camera able to respond to the entirety of the visual spectrum
>>
>> Like an exposure meter, or, better, light meter, camera responds to any
>> light visible to a human (and wider, some UV, some IR that still passes hot
>> mirror and is shorter than 1200)
>>
>> > accurately, one-to-one?
>>
>> No. Metameric error is always present, convolution results in error, plus
>> there is the question of measurement accuracy.
>>
>> > I get how there is not a one-to-one transform from camera to the human
>> observer.
>> >
>> > The transform that takes place from world to camera is made with
>> numbers. If those numbers are exactly the same numbers that are used to
>> define/describe the visible spectrum then I think I can better understand
>> the discussion.
>>
>> A camera is a colormeter, and colormeter readings are just 3 numbers.
>> Suppose they are XYZ. Restoring spectrum from XYZ data alone is impossible,
>> because XYZ represent a convolution of spectrum (based on an observer).
>> During this convolution a lot of original data is lost.
>>
>>
>> --
>> Best regards,
>> Iliah Borg
>> LibRaw, LLC
>> www.libraw.org
>> www.rawdigger.com
>> www.fastrawviewer.com
>>
>>
>>
>> _______________________________________________
>> Do not post admin requests to the list. They will be ignored.
>> colorsync-users mailing list (email@hidden)
>> Help/Unsubscribe/Update your Subscription:
>>
>>
>> This email sent to email@hidden
>>
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden