Re: Resizing Input Targets
Re: Resizing Input Targets
- Subject: Re: Resizing Input Targets
- From: Steve Upton <email@hidden>
- Date: Thu, 22 Mar 2007 17:05:34 -0700
At 8:09 AM +1100 3/23/07, Graeme Gill wrote:
>Steve Upton wrote:
>
>>At 1:20 PM -0500 3/21/07, Anthony Sanna wrote:
>>
>>>I vaguely recall from somewhere that if a scanned target image, such as
>>>a HCT, is down-sized before profiling, bicubic is not the best choice
>>>for the job. Have I got that right? I think nearest neighbor was the
>>>one to use.
>>
>>yup, you have it right. Nearest Neighbor doesn't create all the intermediary
> > colors between patches that just throw things off.
>
>But you'll lose the values of pixels that aren't sampled. Technically
>it's better to use bilinear since then you are sure to use all the
>pixels in the bulk of the patches, although yes, you loose the
>ability to get the ones at the edges, but they are likely to
>be discarded in any case.
good point. I guess you also receive some smoothing of the points as well, which is another part that worried me. Probably not as destructive as full loss though... It's not the smoothing per se but the method and whether it may be the best.
I would normally try to scan at a reasonable file size and then not have to worry about the down-sampling method.
Regards,
Steve
________________________________________________________________________
o Steve Upton CHROMiX www.chromix.com
o (hueman) 866.CHROMiX
o email@hidden 206.985.6837
o ColorGear ColorThink ColorValet ColorSmarts ProfileCentral
________________________________________________________________________
--
_______________________________________________
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