Monochrom tonal peculiarity

[mdarnton]

In order to better understand what I was seeing, I loaded the Chrome browser with a histogram viewer that displays histograms of photos on the web with a right click. Yesterday I was going through one of the Monochrom threads I noticed something that was pervasive in photos from many different photographers, and I'm wondering about it.

Many of the histograms display regularly spaced missing values on the right half, regularly spaced spines of peaking values on the left half, and a strong zero value followed by the first four or five values being completely empty, then a small spike, then another, smaller, complete gap, and then a normal histogram. This pattern does not ever display with photos from other sources, so I'm counting it as a Monochrom peculiarity, not something about the browser plugin.

What I'm wondering is the cause. Is it some common conversion software that many people are using for dngs, or is it something with the camera's internal brain? The pattern is remarkably similar across users, but not inevitable, so I'm suspecting some common post-processing glitch.

Any ideas?

 

[daveleo]

(I have no Leica Monochrome experience. This is based on Fuji camera output.)

I think that any postprocessing adjustment of the "value" (intensity? luminosity?) histogram produces discrete gaps similar to your description.
The fact that the final histograms look somewhat similar is possibly because people tweak to the same approximate tonal results (?)
This example comes from a simple adjustment of the histogram of a color JPG:

 

[Calzone]

In order to better understand what I was seeing, I loaded the Chrome browser with a histogram viewer that displays histograms of photos on the web with a right click. Yesterday I was going through one of the Monochrom threads I noticed something that was pervasive in photos from many different photographers, and I'm wondering about it.

Many of the histograms display regularly spaced missing values on the right half, regularly spaced spines of peaking values on the left half, and a strong zero value followed by the first four or five values being completely empty, then a small spike, then another, smaller, complete gap, and then a normal histogram. This pattern does not ever display with photos from other sources, so I'm counting it as a Monochrom peculiarity, not something about the browser plugin.

What I'm wondering is the cause. Is it some common conversion software that many people are using for dngs, or is it something with the camera's internal brain? The pattern is remarkably similar across users, but not inevitable, so I'm suspecting some common post-processing glitch.

Any ideas?

I use Heliopan 2X yellow filters marked "Digital." I discovered that the filters marked "Digital" have built in IR and UV filtering so that basically noise that does not add to IQ is removed from the histogram. This also increases the signal to noise ratio and the result is that one can shoot more to the right (increase exposure) and create cleaner files with a broader distribution utilizing all ten zones.

I found out from Michael Reichman's review of the Monochrom that Leica engineered the Monochrom to be uses with a light yellow filter for panochromatic response. The yellow filter I find hits the sweet spot of the sensor, and I find that when using a yellow Heliopan filter marked "Digital" that I'm able to capture the most data and produce the broadest histograms that display proper contrast without the need of much post processing. Basically I minimize post as much as possible.

Much of my technic is to think like a large format shooter who is trying to maximize information for contact printing, so the idea is to do everything right at the time of image capture. I do not think most are thinking or doing what I am doing, and most are conservative with their exposures to not blow highlights and are relying on post processing to bring out detail.

I'm thinking the histograms you are seeing are where the exposure/signal is compressed, rather than broad and open like mine.

BTW getting rid of both the IR and UV minimizes or eliminates clipping so some of those bumps to the far right might be some of that signal that is really just noise.

Know that with the Monochrom if you have the detail in the shot and perfect technic you can approach large format tonality and IQ. Not hard to do with the ten zone histogram display and the clipping indicators set at 1% an 99%.

Perfect exposure is easy because basically the histogram provided by Leica does not lie, and most shooters IMHO are not trying to maximize data capture at exposure and rely on post to rescue data. If you print really big, you want to minimize digital artifact and noise as much as possible. I see a broad midrange, mucho shadow detail and a lot less contrast in my images. On my prints I get asked if my Piezography prints are silver wet prints all the time.

"You can't print what's not there," my friend Steve once said.

Cal

 

[Calzone]

(I have no Leica Monochrome experience. This is based on Fuji camera output.)

I think that any postprocessing adjustment of the "value" (intensity? luminosity?) histogram produces discrete gaps similar to your description.
The fact that the final histograms look somewhat similar is possibly because people tweak to the same approximate tonal results (?)
This example comes from a simple adjustment of the histogram of a color JPG:

My histograms are mostly midrange. The taller slope is to the left with a ski slope like drop into the highlights. At the ends are my black and paper white that are a bit spikey. Basically the histogram on LR5 after post strongly resembles the histogram on my Monochrom.

In many ways my histogram on my MM is more like one would see as the distribution of tone from a printer that has 9 shades of black and paper white for ten zones.

There are valleys and rises but no gaps. More continue-us I say.

Cal

 

[Particular]

I think the gaps in the histogram come from adjusting/stretching an 8bit file. Possibly a jpeg out of the camera. 16bit files can be stretched a lot more before the histogram shows gaps.

 

[daveleo]

I think the gaps in the histogram come from adjusting/stretching an 8bit file. Possibly a jpeg out of the camera. 16bit files can be stretched a lot more before the histogram shows gaps.

Very possibly.
My example was an 8-bit JPG !

 

[willie_901]

I doubt this has anything to do with unique aspects of the Leica Monochrome data stream.

This sort of histogram data is extremely common with many camera brands and models. There are several causes and most of them are of no consequence. One common source of histogram gaps is using digital multiplication to increase global luminance in-camera after the data is digitized. Digital multiplication simulates very high ISO settings (so the light meter reading makes sense) once further analog amplification becomes pointless.

Here is just one article that discusses histogram gaps.

Here is are a series of more detailed blog entries that discuss with this subject.

Of course one can create gaps during raw rendering and, or applying lossy data compression algorithms.