Re: Media Testing for maclife.de
Re: Media Testing for maclife.de
- Subject: Re: Media Testing for maclife.de
- From: "edmund ronald" <email@hidden>
- Date: Fri, 5 Sep 2008 16:56:16 +0200
Iliah,
Decoding the Raw is one thing, adding a "look" is another.
Image Engineering has a device which they sell for about $10K which
can determine the spectral response of a camera. I believe this is
well within the budget of most software firms. Here in Europe, loaners
of SLR cameras can be easily obtained for profiling, when I was doing
it professionally I had loaners.
As for profiles and noise, yes there is an issue at that stage of the
pipeline - but that is exactly why you should come to the meetings of
the ICC to debate the solutions. For myself, I don't think that
profiles should be used as a hammer at the raw2xyz stage, only used
for "look" later.
Anyway, even if you have the spectral response, the consensus is you
will still need to add some "look" profile for decent decoding
appearance, because of camera metamerism ...
Edmund
On Fri, Sep 5, 2008 at 3:12 PM, Iliah Borg <email@hidden> wrote:
> Dear Edmund,
>
>> profiles can be employed in a meaningful way in the
>> image rendering pipeline
>
> Yes, they can be applied in a meaningful way; but the problem is - currently
> one can argue he can do better without those profiles.
>
>> the Raw
>> converter is a place where existing profile technology can be applied
>> several times,
>
> My biggest problem with applying profiles (even if in one step and in one
> place) is that they add noise to the image. (On a side note, interestingly
> enough Argyll was the lowest noisy solution - comparing Adobe CMM,
> ColorSync, lcms and Argyll). If a converter wants to provide good shadows,
> both in terms of noise and saturation - current ICC specs are not really an
> option. We can extract data from a profile, manipulate it in a certain way,
> and that allows a custom profile to be applied. However, it is suboptimal
> compared to spectral data we use for our own colour transforms in RPP.
>
> On still another note, things could be much easier if camera makers and / or
> sensor makers were providing spectral characteristics of CFAs and IR filters
> they use in front of the sensor. That would for sure allow much better
> results then profiles obtained from shooting colour targets and edited to
> suite the taste of the folks in raw converter development teams; or to
> provide results similar to the looks of renditions through the converters
> designed by the cameramaker.
>
> For a raw converter developer (small volume, enthusiast, OpenSource,
> freeware type) it is hard to justify the price of the equipment (including
> each camera he is going to support) and the labour of a setup that would
> allow to determine the spectral responses with a good degree of accuracy. As
> a result, the development of converters does not have enough competition.
> One can develop a beautiful demosaicing and wonderful UI but fails to
> provide acceptable colour - such a converter is mostly dead at birth.
>
> An initiative to specifically address camera profiling (that way or another)
> in EXIF, DCF, and ICC standards is long overdue.
>
> --
> Best regards,
> Iliah Borg
> 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