Digital Prints on Fuji Frontier:
How to get prints that honor the files

Auteur: pehache (p e h a c h e _ p r i m @ l a p o s t e . n e t)
Revision: 2.7 (19 Sep 2005)


What you should know about the Fuji Frontier before printing from your digital files, in order to obtain prints that honor the original colors and brightness. This requires a color managed solution, with ICC profiles, and the insurance that no automaticcorrection is applied to the images. 

I got the informations on this pages from various sources, along my trial-and-error process to obtain good prints:

1 - Introduction

The Fuji Frontier is a digital lab. Or rather a collection of labs, with several models that differ mainly by their productivity. A typical configuration is:

Some other digital labs (Agfa, Noritsu..., ) are as good -even better according to some people- as the Fuji Frontier, but it turns that the latter is the most widespread around: you therefore have significant chances to have to deal with it one day or another. Hence this page...

The digital labs can reach an excellent quality. Their arrival a few years ago has boosted the average average quality of the consumer prints market. A skilled operator is still needed to operate the lab and optimise the prints, but the various automatic correction and optimisation algorithms can perform some miracles.

That said, we are not interested here in these automatic corrections. This is all the opposite, in fact: we want to obtain prints as identical as possible (colors, brightness,...) to the files as they are displayed on the monitor. For that, however, we need to know a few things about how the lab works.

Obviously, you need to visualise your images on a calibrated monitor, which displays the "correct" colors. And more generally, you need a fully calibrated workflow, based on a Color Management System (CMS) and on ICC profiles. If all of this is like chinese for you, I advice that you start reading some literature/sites on this topic (google is your friend ;-)).

2 - Calibration of the Frontier

This the operator's task, and the user (you) doesn't need to know about the details (and I don't know anything about them). A good store is supposed to perform at least one calibration every day.

Beware: do not make the confusion between calibration (playing with some hardware and/or software parameters to bring a peripherical onto a reference behavior) et profiling (analysing of the actual behavior, and editing an ICC profile). In a CMS workflow, the profiling is essential, but not the calibration: one can profile a non-calibrated peripherical (e.g. personal inkjet printers). However, the advantage of the calibration is to reduce (ideally suppress) the variations from one peripherical to another (of the same model, of course): hence, the profiling is less dependant on the particular copy of the peripherical, and the resulting ICC profile can be considered as 'generic' (up to a certain point).

A poorly calibrated lab will induce color shifts on the prints. As long as they are weak, these shifts are difficult to bring to the fore on color prints. A unstopable test consists in printing a digital B&W image: any color shift is immediately visible.

It sounds clear that a well calibrated lab is essential to obtain good prints. Otherwise, all the advices belows are useless. Do not hesitate to try several stores to find a good one (which offers constant results).

One performed, the calibrations are active whichever the printing mode of the Frontier.

3 - The printing modes of the Frontier - Color spaces

The first thing to know is that the Frontier doesn't support the ICC profiles. The profiles possibly embeded in your file simply won't be considered. This doesn't mean that the Frontier does not manage at all the colors. It has a proprietary solution, with a simplified management compared to full ICC standards.

The Frontier has two main printing modes: "sRGB" and "Print Direct/no convert" ("PD"). The PD mode is pretty uncommon, most of the consumer/prosumer stores -if not all- use their Frontiers in the "sRGB" mode. The operators often know this latter one only. To get the PD mode, it's better to go to a small local store with its own machine, or to a pro store.

3.1 - The "Print Direct" (PD) mode

In this mode, the file is sent "as is" to the printer unit. The native gamut is therefore used. The comparison of this gamut to the sRGB's one shows that none of the two can entirely contain the other one: some sRGB colors fall outside of the native gamut (and hence can not be printed), while some colors in the native gamut fall outside of sRGB.

The file is actually supposed to be already in the Frontier native colorspace: it is therefore required to have the ICC profile of this colorspace,  in order to convert the files before sending them.

A generic ICC profile for the PD mode is available on request on the Fuji Europe site(see http://www.fujifilm.de/engl/serv/farbm_faq.html). If you contact me, I can also send it to you.
Edit 2012: alternate link to the PD mode profile (right click/save as); warning : this profile is from 2004, and there's no garantee that its the current Frontier machines...

3.2 - The "sRGB" mode

In this mode, the files are supposed to be in the sRGB colorspace, which is the defacto standard for the consumer digital photography (even if  Adobe98 is often prefered by the prosumers). They are converted to the Frontier native space inside the Frontier itself.

The results are correct in terms of color fidelity, as long as the file are effectively prepared in sRGB. If they are not (if they are in Adobe98 for instance), they should be converted to sRGB before being sent.

The results are correct, yes, but not fully satifactory, with significant color shifts, mainly in the yellow and blue tones. Of course, since the native gamut does not contain the whole sRGB space, it is impossible to have prints perfectly identical to the files (this problem is very common on a vast majority on printing systems). However, I doubdt it can explain all of the observed shifts. In my opinion, a part of the problem comes from the internal conversion engine of the Frontier: we have no control over it (it is difficult to obtain such informations from Fuji), and it is a possibility that the algorithms do not fit all kinds of images.

Fortunately, it is possible to enhance the results when using the sRGB mode. This mode can indeed be considered like a "black box", which receives a file and outputs a print. What happens inside the box does not matter. If we have an ICC profile of this black box, it is valid to convert the files to this profile before sending them to the lab, trying various rendering intents and possibly various conversion engines, and retaining the combination that gives the "best" result. For this, the soft-proofing on the monitor gives a fairly good idea of the final (printed) result.

Indeed, the tests show that this method significantly enhances the fidelity of the prints. The main advantage is of course that we have the control over the conversion, instead of being tied to the Frontier internal conversion. It would in fact better stand to reason to work directly in the PD mode, but as stated previously, it is often difficult to find stores that use this mode in practice. 

A generic ICC profile for the "sRGB" mode can be downloaded from the Fuji Europe site, at http://www.fujifilm.de/eng/serv_1120.html. With this profile, we can check that, strangely, the "sRGB" mode doesn't use the whole Frontier native gamut, but rather the common part of this gamut and the sRGB's gamut.
Edit 2012: alternate link to the sRGB mode profile (right click/save as); warning : this profile is from 2004, and there's no garantee that its the current Frontier machines...

Exemple: using the generic ICC profile of the Frontier "sRGB" mode.
left: original image (file in sRGB), displayed without soft-proofing on the monitor.
middle: obtained print when sending the file "as is"
right: obtained print when converting (perceptual intent) the file to the "sRGB" mode ICC profile before sending it.

The middle and right images are of course not the prints, but rather the soft-proofing displays. They pretty well simulate the actual prints.

Obviously, the prints have less saturated colors than the original image. The print obtained after conversion is much closer to the original image than the one without conversion: see the yellow flowers, and the yellow-green transition on the colorscale. 

3.3 - So what ? "sRGB" or PD ?

If you edit your pictures in the sRGB space and if you wish not convert your files before sending them to the lab, then the "sRGB" mode is the one to choose.

If you are willing to convert your files, then the PD mode is the most pertinent one: you have access to the entire Frontier native gamut. The only problem with this mode is its scarcity: only a local store with its own lab will eventually propose prints in this mode. The "sRGB" mode is hence often the only available solution, but it is a very acceptable compromise.

3.4 - Additionnal comments on the profiles and colorspaces

4 - Picture corrections et optimisations

As with any other lab, the operator can manually correct the pictures before printing them (brighness and color balance mainly). But in addition, the Frontier can apply automatic automatic corrections and enhancements (namely DSC corrections) to the contrasts, saturation, sharpness,...

It is easy to request NO manual correction, but for the DSC corrections it is not so obvious. For, they are by default always enabled on the Frontier (in contrast to the similar corrections on some other labs), and the operator has to explicitly disable them for each picture (which is pretty restricting).

Fortunately, there are some means to ensure that the DSC corrections won't be applied, whatever the operator does. For, they are in fact enabled only when all of the following conditions are checked:

If you use the srGB mode, a simple way to ensure prints without correction is to remove the EXIF headers from the files. If your softwares do not offer this, here is one possible method::

I've been also given a link to a utility that is able to directly remove the EXIFs from the files, but I didn't test it: http://www.sentex.net/~mwandel/jhead.

Note also that some stores explicitely propose an option to select or not the automatic corrections.

5 - The papers

The Frontier labs are often used with the Fuji Crystal Archive paper. Or rather without one of the Crystal Archive papers:

I don't know to which degree these different papers have the same color rendering and if the above generic ICC profiles fit all of them or not. I tend to think "yes, they fit all of them", thanks to the calibration step, which is suppose to compensate for the possible differences, but I didn't really tested this point.



Well, you know nw all what you need to know (at least as much as I know) about the Frontier. Just do it, now :-). Do not hesitate to contact me if needed. Voilà, vous savez tout, ou presque (en tous cas autant que moi), sur la Frontier et les couleurs. Y'a plus qu'à :-). N'hésitez pas à me contacter si besoin est.

I didn't browse here the issues related to the resolution, the picture size, and their consequences (resampling, margins, cropping,...). It was beyond the original scope of this paper, and I didn't made any tests. The problem is similar to the one for the colors, in the sense that it is somehow difficult to know what the lab is able to do, what it does in practice, and what are the choices (if any) of the operator (who knows if he can for instance force a resampling, even for a file prepared at the exact size...). You will find a few tests about that and a table of the exact print sizes on these pages: http://fred.just.free.fr/Photo/frontier.html et http://fred.just.free.fr/Photo/fuji/result02/index.html (apparently broken links - under investigations).