March 27, 2023

Arsentsumebite from Tombstone

The specimen was collected on 3-25-2023 in a small stope above the 300 level of the Empire Mine in Tombstone.

I am shooing with the R5 camera and the MPE-65 set at 5x and f/2.8. I let EOS utility capture the images and put them into P:/lightroom/2023/2023_tether I set the camera to capture cRAW and use a 1:1 crop in the camera. Being a complete dummy, I neglect to connect the shutter cord to the camera.

With this in mind, my focus rail software needs an abort button! I have a putty session logged in as root and I can just type Control-C there. I wrote down my bottom and top and am using a 15 step increment. It would also be nice if it told me what step it is on and shows "DONE" somewhere when it is done.

Away we go again, this time with the cord connected (and I hear the shutter). I am using EFCS, so I do have a mechanical shutter closing that makes some noise.

EOS utility makes a directory "2023_03_27" to put the files into. The files are 22.7 megabytes, even with the 1x1 crop set.

The R5 is communicating with Wifi, and the downloads take 22 seconds, so I have a 30 second delay after moving the stage. The current stack will have 54 images, so it will take 27 minutes to take them all! This should improve significantly when I get a new wireless access point that does 802.11ac.

The current stack is from 99800 to 100600 at an increment of 15 steps. My stage gives me 10 microns for 6 steps, so this is a stacking interval of 25 microns.

Since we are now using a full frame sensor, at 5x, the field size top to bottom will be 4.8 mm. With a 45 megapixel sensor, it takes quite a while just to export TIFF files.

Exposures were 1/40 second at ISO 500. The cRAW images are 22M but when converted to TIFF they are 180M. Zerene handles them OK, but it takes some time and when I use "top" on my linux system I see Java running at 400 to 500 percent. I'll admit that I am appalled to learn that Zerene is written in Java. There is a strong argument to move to Helicon if it has been shown to run faster.

I crop it. The final crop is 2458 pixels square. The R5 sensor is 8192 by 5464 pixels, so the crop is 0.45 of the full image. The full image is 4.8, so the cropped image is a 2.16 mm field.

Analysis

This image should be a baseline to try to improve on.

If I was to shoot this again, I would use ISO 100 and I would introduce the light diffusion that I keep talking about (a vertical tube of white cotton). I should also turn off image stabilization - this seems somewhat controversial but I have seen it said that the R5 had some issues with IBIS on tripod that were partially addressed by firmware changes. I am surprised at the amount of what looks like sensor noise at ISO 500.


Feedback? Questions? Drop me a line!

Tom's Computer Info / tom@mmto.org