Re: Rips
Re: Rips
- Subject: Re: Rips
- From: Graeme Gill <email@hidden>
- Date: Thu, 19 Sep 2002 14:04:22 +1000
DuWayne wrote:
>
>Hmmm. Linearisation puts the printer in a balanced repeatable fashion, a
>
>known state if you like.
>
>
It assumes that the printer is not in a repeatable fashion. This isn't true
>
because Heidelberg tested Epson & Roland's over months when developing
>
proofing systems and found that they vary by .40 & .35 delta E. Even in a
>
very controlled environment too may other factors came into play like the
>
spectrophotometer precision is +-.25 Delta E.
How can you make such a broad claim "Epsons and Rolands only vary .4 delta E",
and expect it to be believed ? Its just plain incorrect. Dye inks on
the epson change by well over .4 delta E over 24 hours, and there are
dozens of different Epson models, and at least 3 different inksets.
I've certainly seen prints from an Epson 10000, printed a few days apart,
that have visually noticeable differences. Such differences go away using
calibration.
>
If you assume that the printer will drift due to other factors -
>
environment, media coating etc... to a state that you don't like. You must
>
also acknowledge that you have the possibility of hue & chrome shifts which
>
linearisation will not correct. So you might as well build a new profile.
Yes, there is a possibility of chroma shifts, but because the basic color
of the inks themselves doesn't vary so radically, shifts in ink interaction
are of a much lower impact than changes in basic channel linearity. Linearizing
the device channels gets you 95% of the way there, and the remaining
device to profile mismatch is down in the noise by comparison.
>
>>Gamut by proper linearisation is increased not decreased. When too much
>
ink is laid down at a certain >>point the Chroma values reverse. Trick is to
>
maximize the gamut by linearisation, something that the >>Epson drivers
>
didn't have.
>
>
I understand what you mean but try it in the real world.
It depends what sort of linearization is done, and what sort of
compromise has been chosen between repeatability and maximum
gamut size. You can linearize to a fixed maximum density, or
you can linearize to the maximum density of the particular
device in question.
The former does compromise the maximum density achievable on
a particular device to a very small degree, but gives repeatable
and uniform responses, day to day, and device to device.
And you need to keep in mind that the small amount of gamut
you loose here, is only lost because it isn't available on
all instances of that device in the first place.
If you linearize to the device maximum (or don't linearize
at all), then you will get the maximum gamut for that particular
device, but it won't be stable, or repeatable from device to device.
This is fine for pretty pictures, but not for proofing.
Making high quality profiles as a means of calibration is
not a practical proposition at this point in time. Our high
quality profiles use about 3000+ test patches, and hence cost
to print, and cost time and effort to read and process. By
comparison, printing and reading a few step wedges is cheap
and fast. A calibration system that is too difficult to use
is nearly as useless as no calibration system as all, because
users simply won't use it.
>
Download Profilemaker demo and use their gamut viewer in profile editor.
>
Look at all the canned profile that you can download for all the rips in the
>
market that do linearisation and compare them to ImagePrint's canned
>
profiles. The gamma differences are huge.
I've taken a look at one of ImagePrints profiles (Epson 10KCF on
Epson Gloss Paper Photo Weight), and it looks, well, not credible.
It claims for instance that yellow has an L*a*b* value of 92, -8, 125,
and there is no possible way that the pigment ink on the 10000CF has
a b* value of 125 !!! (in fact, I defy you to find any yellow printing
ink with a b* value of 125).
Typical of a printer being run through an RGB mode, the gamut in the
dark and dark saturated colors is severely (and unnecessarily) limited.
You aren't getting anything like the full potential of this printer
using this profile !
>
Its going to be 3 new drivers for a whole new type of head. Epson has the
>
10600 with 6 color ultrachrome. The rip developers just don't have enough
>
qualified bodies to write the code from scratch - so the rig it. Well you
>
can only rig it some much till you get trapped- that where a lot of rips are
>
like onyx, colorburst etc...
Speak for yourself. For us, the 10600 was dead easy, and the Ultrachrome ink
works very well. Haven't had to change a thing, apart from creating some
new color profiles. The 7600 was more fun (in generating the light black
data channel), but pretty straightforward, and the profiling was pretty simple.
In general, keeping up with driving inkjet printers directly is relatively easy,
and the current crop of manufacturers print drivers can't achieve anywhere
close to the same results.
Graeme Gill.
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.
| References: | |
| | >Rips (From: "DuWayne" <email@hidden>) |