Re: filmscanners: Printdpi

[Arthur Entlich <artistic@ampsc.com>]

Tony Sleep wrote:


> 
> 
> Yes, I agree. The Epson driver is broken, and they haven't fixed it, which 
> kind-of suggests to me that they cannot fix it without revealing some deeper 
> flaw. I would summarise the problems I see as:-
> 1. Broken colour management - the driver cannot do a profile transformation 
> adequately

They seem afraid of greens, have you noticed?

> 2. Poor profiling. The canned profile is way off the mark, as can be proved 
>by 
> comparing 1. with doing a profile-profile within PS and then bypassing the 
> printer driver CM.
> 3. Variability/inconsistency of output. As you mention, different carts or 
> paper batches affect consistency, though IME this is a minor problem

One thing they don't like to talk about is that after the carts are 
about 1/2 empty the amount of ink flow changes, making delicate color 
management impossible.  Constant ink systems or systems like the 3000, 
which don't use the above head cartridge system might be better.

> 4. Metamerism. I see massive amounts in vanilla 1200 prints on Epson Photo 
> Paper. It's a joke, you need a calibrated daylight reference to be able to 
>view 
>  prints! A $1 daylight tungsten artists bulb works fine for conventional 
> darkroom prints.
> 5. Dry-down colour shifts. I see excess red which calms down within a couple 
>of 
> hours. You can't rebalance a print before that, except by guesstimation.

Now, come on... are you telling me wet color prints come out looking 
color balanced?  Some papers look very blue until they dry, Ciba looks 
magenta brown.  At least they did when I used to do color darkroom work.

> 6. The gamut is wide, but has some sharp discontinuities and weaknesses. This 
> plays havoc with trying to get a precise match in some colours/tones - 
> grass/foliage can be especially impossible, and pale European skin tones will 
> drive you stark staring mad. 

Tell your models to get a tan or eat more carrots! See my response to #1 
regarding greens.

I think this characteristic alone limits the 
> opportunity for custom profiling, at least any attempt to DIY using a 
>flatbed. 
> A precision spectrophotometer might stand a better chance, but I suspect that 
> the printers are so twitchy that 3, inconsistency, would scupper even that - 
> unless you were able to reprofile for every image.
> 7. Archival longevity I am not even going to mention.
> 

I think you just did ;-)

> Overall - and I am being picky - these are adequate tools for casual use, but 
>I 
> doubt the aftermarket industry borne of discontent with the OE product, can 
> really conquer many of the issues. But I have no experience of either CIS 
> inkset or Cone's new colour kit. Maybe they have fixed things, the trouble is 
> it is lots of $$ to find out.

Approaching perfection is always costly.  What does a really good 
enlarger and lens cost?  What does temperature and time controlled 
darkroom equipment set you back?


> I should also say that I don't think $100,000 Iris printers do any better, 
>but 
> for different reasons. As for HP etc, forget it. Epson's are the best, so 
>far, 
> but not there yet. The new Canon S800 looked very promising indeed, largely 
> because the test print I saw was subtle and looked 'right' - Epson demo 
>prints 
> are always high contrast, high saturation, and gloss over the weaknesses.
 leben.com to see 
> that people are really struggling with these printers. They are good, 
> tantalisingly so, since it is easy to produce a print which is *almost* 
>right. 
> But I despair of producing one which *is* right, that sooner or later I don't 
> go back to and think 'urgh, that sucks'.
> 
> 

You're either too hard on yourself, or the printer, or both.  I'm not 
trying to be unkind here, but if you cannot afford the tools necessary 
to make the result you are after, either go back to an earlier method 
that works (chemical) or accept the flaws and recognize that 99% of the 
people looking at it will not see it the way you do.  Only you have the 
internal vision of what your expectations are, and although I'm not 
suggesting you toss those goals away, you need to recognize the 
limitations of the tools you have to work with now, and that you will 
improve them as they become more affordable to you.


>> I am getting to the point that I will be taking my disk to my friendly 
> 
> Noritsu/Fuji
> 
>> printer lab.  Hopefully we we can come up with the adjustments needed to get 
> 
> the
> 
>> print to match the disk image as I see it on my screen.

I sold many people on PCD when film scanners and CD-R burners were 
either vastly expensive or totally unreliable.  Now, I recommend they do 
it themselves if possible.  The market changed.  If the Noritsu/Fuji 
silver paper machines can get you where you want to be, it make be 
cheaper than a $7500 5000 with RIP, and it is certainly cheaper than 5 
years of therapy ;-).


> Hm, well. My test Noritsu print was closer to the screen/scan than I'd have 
>got 
> from the Epson after much struggling, and with no tweaking at all - the 
>Noritsu 
> seems to do CM properly! But there was an obvious posterisation in light, 
> graduated tones, which looks to me like an aliasing issue. I tackled the lab 
> about this and they said 'ah, yes - we've spotted that too'. They aren't 
> pleased, having just bought the machine.
> 
> I can see a need for hardass, objective reviews of printers too...
> 

None will fulfill your needs in the price range you are looking at right 
now, at least with current technologies.

Art