LED choice for video using microscope

[Andy Davies]

I'm using a Nikon D810 that has Bayer array. I've emailed Nikon technical support as to when the WB correction is made. I'm an NPS member but doubt whether they will give an info on this as they are very secretive when it comes to technical details.

Does the brightness (power) difference between the two emitters mean that the spectral response of the 6000K LED is shifted upwards compared to the lower power 3000K emitter? If so, would this mean that the CRI difference would effectively not be as great?

[Asha]

I am wondering if the article should be read with caution--the camera he used has a Bayer filter (2 green, 1 red, 1 blue in a block pattern). Cameras with different kinds of sensor designs (eg a stacked sensor such as Foveon) may not behave the same way.

This article should not be taken as universal, for sure (I wouldn't call it "with caution" - it is just normal to understand limits of applicability of any scientific theory). But how much non-Bayer sensors and sensors with significantly different per-channel quantum effectiveness are on market?

I'm not sure--that is why I was asking the question. I know all the latest Fuji cameras from the past few years to present use their own proprietary design. And Sony has new stacked sensors in their smaller cameras--probably TBD whether those end up in the high-end bodies (and Nikon's, since they use Sony chips).

[Asha]

Does the brightness (power) difference between the two emitters mean that the spectral response of the 6000K LED is shifted upwards compared to the lower power 3000K emitter? If so, would this mean that the CRI difference would effectively not be as great?

By brightness (power) do you mean luminous flux?

There probably is a difference in the total radiance for the LEDs that are shifted toward the blue end based on luminous flux numbers, and based on knowledge of Planck's curves. However, the spectral curves they show are normalized to 100%, and each color temperature range is normalized to its own maximum. Because of this, it isn't really possible to determine absolutes between the different color temperature ranges from the spectral output they show in their data sheet.

[Andy Davies]

Yes, I was talking about luminous flux.

[Bushman.K]

In case of LEDs with their uneven spectrum, usage of term "color temperature" is a bit "forced". Also, luminous flux is a parameter, that is normalized to human eye. Therefore, these parameters are partially irrelevant to digital photography. When speaking about human vision, we need weighted parameters such as luminous flux, because we can't take RAWDigger and tap into an optical nerve. But with digital camera we obviously can examine RAW files and experiment with different light sources, giving us better picture right out of camera sensor.

As I said before, for photography, it is more important to get more even exposure in each channel and better spectral range span. So, personally, I would prefer one of those warm white LEDs with stronger emission in deep reds (and relatively smaller "green pit").

[Bushman.K]

To be more precise, in case of digital photography, only amount of photons converted into electrons by sensor channels does matter. Therefore, any weighted parameters, depending on human eye (luminous flux) or wavelength, are distorting the picture of situation.

[Andy Davies]

After a few emails and a telephone call with Nikon NPS it turns out that the video signal collected from the sensor is only 8 bit and so any WB correction is applied to this signal as I first thought.

In light of this, which would be the recommended LED? The brighter and little colour tint 6000K one or the less bright yellow one?

Many thanks

Andy

[benjamind2014]

The thing that concerns me, and I can't remember anyone replying to my concerns, is that LEDs with DIMMERS can sometimes produce these funny horizontal lines or like a kind of flickering.

The camera picks this up. Well, mine does anyway. Not sure what would cause this. I know it isn't noise because it's caused by the rolling shutter of the camera and the LED driver is a 21kHZ speed so it's very fast. I am still stumped as to how the camera can pick this up.

Maybe it's the camera itself. If so I'll just use my Canon 600D with a 2x projection lens adapter and give the other camera to my nephew who would be using it with an Olympus GB scope. I plan on gifting the GB to him at some point in time when he is mature enough to know how to use a scope.

[rjlittlefield]

After a few emails and a telephone call with Nikon NPS it turns out that the video signal collected from the sensor is only 8 bit and so any WB correction is applied to this signal as I first thought.

Interesting. I wonder then how they apply the gamma curve without getting serious gaps in the histogram. Some analog circuit before the A/D converter? (I'm actually skeptical that the answer you've gotten is exactly what you want to know, but it really doesn't matter.)

In light of this, which would be the recommended LED? The brighter and little colour tint 6000K one or the less bright yellow one?

I don't know the answer to your question, but I can suggest how to answer it yourself.

Shoot some raw images of a gray card with your camera and process them through dcraw using its -D option to output the sensor data as "original unscaled pixel values." Do this with a variety of light sources, and see which sort of spectrum gives the least prominent checkerboard pattern. That should be the one that will require the least adjustment of any RGB channel.

--Rik