Improving microscope images

[porina]

 

I just got the above chips to try and photograph at high magnification as I love die shots, and wanted to try myself with an x86 CPU some time. Obtaining or extracting an x86 die from packaging is proving beyond my means at the moment, so I bought these 4-bit microprocessors to play with for practice. This shot was a hand held macro using a DLSR. Each die is about 5mm on short edge. For high resolution, I need more than a camera lens. Enter the microscope.

 

I have a compound microscope which has a single 10x eyepiece and also a camera port. With an adapter (which says 2x on it) I can connect it to a DSLR, in my case an APS-C sensor model.

 

I have 4 objectives:

1. 4/0.1 160/-
2. 10/0.25 160/-
3. 40/0.65 160/0.17
4. 60/0.85 160/0.17

 

My understanding of the numbers is they are: magnification/numeric aperture, some kind of distance/coverslip thickness

 

With the DSLR fitted, my first test shots are as shown:

 

With 4x objective

 

With 10x objective

 

I couldn't use the higher magnification pieces, as they had to be so close to the subject there wasn't any room for me to get light in there. The 'scope has light under the sample so is designed more for shining through things, than reflecting off things.

 

Open question, how do I improve on the images? The focus doesn't feel quite right, and as I was fine adjusting it, it felt like the detail on the subject had enough height variation I can't get it all in focus at once anyway. Focus stack might be needed. I also wonder if using monochromatic light might help against chromatic aberrations. At least, separate narrowband RGB lighting.

 

Presumably a coverslip is not required for the two objectives I've used so far, but are required for the other two should I find a way around lighting from the same side. On that note, I have one plan that is crazy enough it could work. In a test, I can shine light into the eyepiece, and I can see a lit area where the subject would be. Since I'm imaging using the camera port, I hope the two can work together, but I haven't tested it yet.

 

[Radium_Angel]

8 minutes ago, porina said:

Focus stack might be needed

There is software to help with that, but it's still a PITA to use right

[porina]

1 minute ago, Radium_Angel said:

There is software to help with that, but it's still a PITA to use right

I tried it a long time ago for regular macro photography. Was kinda hoping things might have moved on from then!

[Radium_Angel]

2 minutes ago, porina said:

I tried it a long time ago for regular macro photography. Was kinda hoping things might have moved on from then!

While there are more versions out there, I've not had great luck with the main ones, and just gave up, never caring for the results.

 

[porina]

Just now, Radium_Angel said:

While there are more versions out there, I've not had great luck with the main ones, and just gave up, never caring for the results.

I don't know what the software I last tried was, but I do recall I never really took enough slices for the stack to work well. Making small enough adjustments was difficult.

 

I love macro photography, and see microscopy to be an extension of that. I just wish the laws of physics allowed us more depth of field without diffraction limitations. One workaround, if I don't mind an monochromatic image, is simply to use shorter wavelength lighting. But then I have the bayer pattern getting in the way... it is never easy. I wish anyone other than Leica would offer unfiltered (panchromic) sensors in a consumer level body/pricing. Same problem applies to my other former hobby of astrophotograhpy.

[Radium_Angel]

8 minutes ago, porina said:

unfiltered (panchromic) sensors in a consumer level body/pricing.

https://www.lifepixel.com/

 

I can personally vouch for their quality. I believe they offer this service.

[porina]

1 minute ago, Radium_Angel said:

https://www.lifepixel.com/

 

I can personally vouch for their quality. I believe they offer this service.

If I'm reading it correctly, they only do IR conversions? I've done that myself already to some cameras. What I'm interest in is not having the Bayer colour filter. I have seen people DIY this, which basically is mechanically removing it from the sensor front itself. The risk of damage is too high for me to try.

 

Alternative is getting a camera with a sensor that never had it fitted in the first place. There are many models used in astrophotography but at any half decent resolution they're really expensive. I believe Leica has a model, but that is really expensive. I think because it is so niche, the prices are rather high. Seems backwards that I want less on the camera, but to do so means I would get charged a lot more. I do understand economies of scale.

[Radium_Angel]

17 minutes ago, porina said:

they only do IR conversions?

They are most well known for that, but they do other things as well, including (IIRC) the removal of that filter you are talking about

[Radium_Angel]

https://www.lifepixel.com/product-category/our-services

 

[porina]

3 minutes ago, Radium_Angel said:

They are most well known for that, but they do other things as well, including (IIRC) the removal of that filter you are talking about

If they do, I can't see it. I see they offer what they call full spectrum mod, but that's still in essence only removing the IR filter. It is full spectrum in the sense nothing is significantly filtered before the sensor, but the colour filter remains on sensor.

[Radium_Angel]

3 minutes ago, porina said:

If they do, I can't see it. I see they offer what they call full spectrum mod, but that's still in essence only removing the IR filter. It is full spectrum in the sense nothing is significantly filtered before the sensor, but the colour filter remains on sensor.

Ah, I was thinking this

https://www.lifepixel.com/product-category/our-services/anti-aliasing-filter-removal

 

They used to do monochrome conversion, looks like they stopped doing that.

My bad...

[Radium_Angel]

https://www.monochromeimaging.com/

 

Some firms exist that do conversions to monochrome, but I can't speak for their quality

[porina]

17 minutes ago, Radium_Angel said:

https://www.monochromeimaging.com/

 

Some firms exist that do conversions to monochrome, but I can't speak for their quality

Judging by their pricing, it is still no simple job to perform this. 

[Radium_Angel]

35 minutes ago, porina said:

Judging by their pricing, it is still no simple job to perform this. 

And it's a narrow market, therefore...high price.

The alternative, is track down an old (errr...ancient) Kodak dSLR based on the D1x series (or F5, IIRC) that were monochrome right from the factory

https://www.nikonweb.com/dcs760m/

 

[porina]

11 minutes ago, Radium_Angel said:

And it's a narrow market, therefore...high price.

The alternative, is track down an old (errr...ancient) Kodak dSLR based on the D1x series (or F5, IIRC) that were monochrome right from the factory

I'd argue it would be as big or bigger potential market as IR, which is in essence a subset use case of it. The difference is the difficulty in conversion where the sensor does not exist in native form. IR mods essentially come down to removing the IR block filter external to the sensor, and optionally putting in a new filter of the user's preference. The mono conversion is literally scraping away the colour filter on the front of the sensor. The page previously linked also mentions that phase AF is broken by doing so, presumably as microlenses on the sensor front are also lost along the way. It is a much more difficult job to do.

 

An ancient digital camera is no real solution either. The main point of having a modern "monochrome" sensor is you get to keep the resolution and advances in sensitivity. Early gen mono digital cameras would probably perform worse than a colour one of recent years. For practical purposes, I consider colour sensors to have 1/4 the effective area resolution than stated. 1/4 of a modern sensor is still better than a full old sensor.

[Radium_Angel]

6 minutes ago, porina said:

old sensor.

I get that, it's just the only option I knew of that wasn't a steep price to pay.

And that's ultimately the problem isn't it? Cost vs Market size.

 

Sadly, I have no other answers or options, but your microphotography is very cool.

[porina]

1 minute ago, Radium_Angel said:

I get that, it's just the only option I knew of that wasn't a steep price to pay.

And that's ultimately the problem isn't it? Cost vs Market size.

Thanks. It was the same conclusion I came to years ago, and thing haven't really changed. Maybe I should check out what current astrophotography sensors are like, but they weren't cheap then either.

 

1 minute ago, Radium_Angel said:

Sadly, I have no other answers or options, but your microphotography is very cool.

Thanks. We seem to have gone on a bit of a tangent. I'd still welcome any advice on improving imaging using a microscope. Like normal photography, I think I need to work more on lighting also.

[Radium_Angel]

4 minutes ago, porina said:

lighting

Photography is painting with light, so yeah, of course with a microscope this will take some experimentation but that's half the fun of learning.

[GDRRiley]

10 hours ago, porina said:

Thanks. It was the same conclusion I came to years ago, and thing haven't really changed. Maybe I should check out what current astrophotography sensors are like, but

EOS RA is 2500

maybe a used 60Da, thats an 18mp sensor

[porina]

4 hours ago, GDRRiley said:

EOS RA is 2500

maybe a used 60Da, thats an 18mp sensor

Those have a modified IR block filter, so for my purposes they suffer the same problems from the colour filter as most other devices and thus for this application are no better.

 

Following link is the kind of device I'd like, it if wasn't for the price tag. https://www.ikarusimaging.com/qsi-cameras/qsi-6120-12mp-monochrome-ccd-camera