Review of Color from Intrepid Enlarger by the Naked Photographer

[albada]

Here's the link to the Youtube review:
Intrepid Enlarger Print Results

Here's a summary of what he says:

• For black and white, this enlarger is fine. It reaches grades 00 and 5. Their grade-settings don't quite match Ilford filters, but folks can make that minor adjustment.
• For color, its LEDs don't emit the correct wavelengths to get accurate colors, and will thus frustrate both beginner and advanced users. At 15:00, he discusses technical details of the problem, which is the use of combo LEDs (RGB in one LED), and suggests that Intrepid switch to separate LEDs for red, green, and blue, so they can produce wavelengths better-suited for color paper.

Mark

 

[koraks]

At 15:00, he discusses technical details of the problem, which is the use of combo LEDs (RGB in one LED), and suggests that Intrepid switch to separate LEDs for red, green, and blue

Hm, that sounds eerily familiar: https://tinker.koraks.nl/photography/why-rgb-leds-suck-for-a-color-ra4-enlarger/

In fact, I think I've voiced similar concerns (without having actually used the Intrepid enlarger myself!) back when this device appeared on the market, also based on example prints that were shown in videos etc.

 

[albada]

Hm, that sounds eerily familiar: https://tinker.koraks.nl/photography/why-rgb-leds-suck-for-a-color-ra4-enlarger/

In fact, I think I've voiced similar concerns (without having actually used the Intrepid enlarger myself!) back when this device appeared on the market, also based on example prints that were shown in videos etc.

@koraks, I was hoping you'd reply to this thread. I re-read your article, and noticed this in one of the comments:

Noritsu laser unit uses 690nm laser for red, 532 nm for green, 473nm for blue​

690 for red is superb, but I expected around 550 for green and <=450 for blue. But I remember your dare to try blue at 480 because that's where the blue and green paper-curves have their largest difference. Your comments?

 

[koraks]

Long story short: there are several ways to skin a cat, and I've come to believe that there's no single 'right' answer.

A few notes, based on interactions with Fuji engineers:
* The choice of red shouldn't be too critical as long as the red LEDs don't have a secondary emission in the green spectrum. If you use 620nm red (default in RGB) you just need a lot more power than if you'd use 690nm LEDs. Otherwise, it should work. Again, provided that the color purity of the LEDs is good. This is what my main concern is/was with the common 620nm LEDs.
* 550nm for green seems to be optimal. 525-532 is on the low side. This may affect magenta purity; I've done some testing recently that suggested this is indeed the case. Sadly, 550nm power LEDs are still scarce - but they're getting there. This is one test I'd still like to do, but I've yet to order some LEDs for this.
* For the blue, there are a few lines of thought. One person (Fuji) says that you should get better yellow purity if you pick a low wavelength, i.e. below 450nm, since the green emulsion has virtually no sensitivity there. However, I find this argument not really convincing also based on early testing I did which showed that wavelengths around 430nm really didn't work very well at all (poor yellow purity, in fact). 480nm seems optimal based on theoretical exercises, but I yet have to test this, too. It seems to make good sense.

This chart remains highly relevant:

Note that if you look at commercial (digital) exposure systems, the wavelengths they pick are pretty much all over the place. This suggests that manufacturers readily accept a compromise and then try to make the best of it through LUT's and/or ICC profiling, and in the end probably accepting a little gamut shrink. Taking the Noritsu as an example, the blue seems very logical given the peak of the blue layer on the paper. Red also matches very well. But green seems to be a less fortunate choice, although they probably get away with it because it just manages to skirt the blue sensitivity curve. I suspect that there's a technical reason for this compromise (why didn't they hit the green layer peak like they did with blue & red?), and that compromise is undoubtedly a lack of availability of a usable 550nm emitter.

When it comes to enlarging negatives, you also have the dye absorption curves to contend with, of course. This is ignored in all digital exposure systems for obvious reasons.

Another potential problem with the Intrepid system is that they likely use 8-bit PWM resolution, which is insufficient for color filtering; the steps especially for the blue and green channels are too coarse to fine-tune color, leaving you with a significant cast in every print. Edit: I was just scrolling through the Naked Photog video and he confirms my suspicion around the 12:30 mark that the filter settings are too coarse especially on the blue channel. This confirms my earlier expectations/reservations about this system. It also seems that the Naked Photog doesn't recognize (yet) the relevance of this aspect and he seems to bunch it together with his argument on wavelengths - which is also relevant, but a different issue altogether.