Rayfact 105mm f4.5 vs Coastal Opt 105mm f4.5 - monochrome camera

[JMC]

Having recently got a Coastal Optics 105mm f4.5 lens I wanted to test it against the Rayfact 105mm f4.5 Some days, testing goes really smoothly. Others, it just doesn't. Today was one of the 'others'. Changing sunlight, rapidly moving clouds, , not checking focus, and chasing down light leaks I've not experienced before all lead to a fun set of experiments.

 

Anyway, back to the interesting bit. This was all done with my monochrome EOS 5DSR and using a Baader U filter. I had a 40mm'ish' lens hood on both lenses. All done in sunlight, with the camera mounted on a tripod, and using a 2s timer. I did not however use mirror lock up. ISO640 for all tests. f value and exposure mentioned by the shots.

 

1. Sunflower photos.

These were captured as JPEGs in the camera and used a custom white balance setting from a PTFE tile. f11 and 1/4s used for them. Approximately 50cm distance to subject. Camera did not move between lenses. Firstly, the full size images (resolution redone to fit here, but left as original field of view);

 

Rayfact - full field of view

 

Coastal Optics - full field of view

 

First thing is the field of view is different. So are they both actually 105mm lenses?

 

EDIT - at infinity they have the same field of view, so yes they are both 105mm (or at least they are the same).

 

Now, going into the same area on both images for a really tight crop;

 

Rayfact - cropped

 

Coastal Optics - cropped

 

The Rayfact looks to have higher contrast, and also be a bit sharper to me. It also seems to be a slightly brighter image, so does it have better light transmission in the UV?

 

2. Light transmission

I used the Spectralon diffuse reflectance targets for this. Took a picture directly down on them. Same lighting for both, and images captured as RAW files this time. ISO640, f8 and 1/60s. Here's what the image looked like (as a jpeg);

 

I took the RAW files, opened them in RawDigger as Raw composite files, and measured the R, G, B, and G2 channel response for the 8 channels. Averaged these together, and plotted the average response for each lens as a function of reflectance from the 2 different lenses, and got this;

 

As with the sunflower shots, the Rayfact is letting more light through for a given reflectance standard than the Coastal Optics lens. Looks to be between 1/2 and 1 stop difference. I shall run this lens through my more controlled camera spectral sensitivity measurement system when I get chance as well to see if there is variation with wavelength.

 

3. The light leaks

This surprised me, as I have rarely seen light leaks before, but I was getting them with both lenses today in different places. The EOS-Nikon adapter I use was a little lose on the lens (I have one adapter per lens). The Rayfact had a tiny amount of play in its adapter, and the Coastal Optics more so in its one. Thankfully the adapters can be tweaked slightly to make them a better fit, and once I had done this the light leaks disappeared, but it does highlight the need to make sure the adapter is a snug fit on your lenses. I have not had this issue with the M42 lenses, but they are screw fit, so that's not a surprise.

 

The Coastal Optics bayonet fit had about 0.1-0.2mm more gap between the mounting face and the inside of the bayonet than the Rayfact did. This explains the extra wobble. It doesn't look to be damaged, and the only other person I know with one of these has told me he had the same issue too. Not sure how you can get machining a bayonet that wrong, but it does look to be an issue with these specific lenses, although by all accounts the 60mm fits much better.

 

Not sure what the fit of the Coastal Optics lens straight onto a Nikon body would be, and whether this wobble is still a problem. Shall have to try that at some point.

 

4. Overall conclusion (based on data in the whole thread, and updated Aug 25th 2018)

The Coastal Optics has slightly less light transmission than the Rayfact when used at the same settings, and was a worse fit mechanically.

 

Using mirror lock up and doing everything possible to reduce the effects of wind on subject movement the lenses gave similar contrast and sharpness (at f5.6), just with reduced brightness as backed up by sensor sensitivity measurements.

 

If you can ever find a Coastal Optics 105mm second hand, they go for less than the UV Nikkor, and are definitely less than a new Rayfact.

[nfoto]

There is movement blur on both images, thus making any difference in UV sharpness and -performance difficult to establish.

 

The location of the entrance pupil and the distance to the subject will dictate the actual perspective. In the close-up range, even small variations in the position of the pupil can result in different field of view even though focal lengths might be identical.

[JMC]

Yes, it's windy here today, and combined with not using the mirror lock up presumably resulted in the slight movement.

 

I was wondering about whether the internal construction would vary the field of view, especially in macro. Will have to check them both out at infinity too.

[Andrea B.]

Birna's got the explanation there for you about the difference you are seeing in field of view at close range. The entrance pupil determines your field of view but it is not always at the same location as the physical aperture.

 

I like this explanation (link) of how to find the entrance pupil. It is well illustrated. You can find the EP closely enough using an ordinary ball head, a level of some kind and Live View. Nikons have an internal level that can be brought up in Live View which makes this easy. Do Canons have that?

http://www.vrphotogr...ptalign-tn.html

 

Both those lenses can resolve the pollen stuff on the florets or conical cells on the surface of a petal. But all sharpening techniques need to be used to do that: tripod, mirror-up, remote shutter trigger and...... no breezes or fast shutter speed. In UV we seem to get an extra stop before diffraction sets in, but do experiment between f/8 and f/11 for the particular system in use to see which has the most edge sharpness.

 

I know the feeling very well about having everything go wrong in a UV photo experiment !! I think part of why we enjoy this so much is that is often is not easy at all.

[JMC]

Just gone and checked, and at infinity they have the same field of view. Have updated the initial post accordingly.

 

Further uv testing will have wait until the sun comes back and the wind dies down.

[Andrea B.]

Has Birna been sending in those gale-force winds from Norway again?

 

We have so been lacking in sunlight recently here also. However from all the rains our August is very green and lush instead of the typical summer brown-out.

[JMC]

Westerlies today Andrea, so I guess it's from you guys. Definitely much breezier today than it has been. Just started raining too....

[JMC]

Little bit more data to share. I ran the CO lens through my system for measuring spectral response for the camera sensor - thankfully I can do this when it's raining or dark. Used the monochrome 5DSR and compared it with the Rayfact 105mm f4.5. Here's the response for the 2 lenses from 280nm to 480nm;

 

I get the same shape line for the Coastal Optics lens as for the Rayfact, but a lower response. This ties in with the original posting where I mentioned less light coming through the CO lens for the same settings as the Rayfact, and is a more quantitative test.

 

On the current Coastal Optics advert for this lens they show a lens transmission plot, and it has an odd sinusoidal wave to it. The same plot for the Rayfact, supplied by Togichi Nikon, does not show this type of wave. I see no evidence for a 'wave' like variation in my Coastal Optics lens transmission as a function of wavelength. This could be down to the resolution of my measurement setup, or perhaps as mine is an older version the coatings are different to the current ones? Not sure on that though.

 

Anyway, as a final model of the data, I superimposed data for sunlight intensity and Baader U transmission as a function of wavelength, on to the curves and got this;

 

This is a modelled curve, but shows the response of the camera sensor for each lens, but combined with the change in intensity of sunlight with wavelength, and Baader U transmission. Very similar shaped curves for both lenses, and as above, no sign of a 'wave' in the transmission of the CO lens. Area under the curve would show about 0.5 stop difference between the expected exposures for the 2 lenses at the same settings.

[JMC]

Given the comment about blur in yesterday pictures, I tried a quick experiment this morning with a few changes. I used mirror lock up, had the flower nearer the ground, and the tripod as a result also lower. It was less windy here this morning too. I reduced the aperture to f5.6 to try and split them apart as well.

 

First the full sizes images (reduced in resolution for posting here).

 

Rayfact

 

Coastal Optics

 

Now, cropped to a part which was in best focus (unfortunately not the same part for each lens).

 

Rayfact

 

Coastal Optics

 

Mirror lock and less wind have helped here and there is much less blur evident.

 

The Rayfact image is brighter, as has been seen with the other shots, and demonstrated in the sensor response test.

 

As for sharpness, I think it is a much closer competition now. Perhaps the Rayfact still has the edge, but I'm not certain on that any more. A lens test chart might be the way forward here.

 

I'll update the conclusions in post #1 accordingly.

[JMC]

One more test for now - focal shift going from visible to UV. Both lenses now at f4.5. Focussed in the visible on a part of the sunflower head. Add on the Baader U and lens hood. Mirror lock up used again. Same exposure for both.

 

Firstly the full sized, uncropped images (rescaled from original resolution for here).

 

Rayfact

 

Coastal Optics

 

This was the area focussed on (the white rectangle).

 

There was a small feature in the centre of the area, which is shown in the following shots.

 

Cropped images, this time cropped to have the same field of view as each other, and left at original resolution.

 

Rayfact

 

Coastal Optics

 

The feature focussed on in the visible light was the triangular one at the centre of the cropped images. Both lenses show no significant change in focus point when the Baader U filter was used compared to visible.

[nfoto]

There still are some traces of movement blur. Mirror lockup needs to have a lead time before the exposure is made, in particular if the tripod support is not perfect.

 

You should try focus stacking with each lens and run them at say f/5.6 - f/8 to move into the zone of maximum performance. Do this indoors with flash illumination.

[nfoto]

The first pair of curves in post #8 probably refers to the camera response more than that of each lens. The UV-Nikkor 105 and its current incarnation (Rayfact UV-105) go almost flat down towards 220 nm.

[JMC]

The first curves in #8 aren't meant to be transmission curves for the lenses Birna. As described these are the sensor response of the camera and are measured with the specified lens attached, so it's way of checking the response of the camera/lens combination.

[dabateman]

Well it looks like you still have a winner of a lens. How does the minimum focus distance compare between the lens?

In your post 10 above it seems as though the CO has slightly closer minimum focus distance, leading to shallower depth of field. This can also have slight loss of light as you're tighter in on a shot.

 

1/2 stop less doesn't seem huge, I think that would account for the cost difference, he says trying to be funny.

 

[JMC]

David, the comparison shots were done from the same position, so the change in field of view is as Birna says, down to internal differences in the lens. The Rayfact allows me to focus closer by a couple of cm. However at those close ranges the CO lens shows a little more magnification at a given distance.

 

Thanks, it seems to be a good lens. It has some quirks though, like no distances of focus scales on the lens. I find the focus lock on the Rayfact useful when putting on and taking off filters especially in macro, again the Coastal Optics doesn't have that. Overall though I'm happy with it :)

[Andrea B.]

Good testing, Jonathan. Thank you for these efforts.

 

Have you linked this test to a photo of the CO 105/4.5 in the Lens Gallery?

 

A bit of musing ...... There never is a "perfect" lens for UV is there? Seems like each UV-capable lens has its little quirks, yes? We always have to try to learn how to get the best from each lens whatever its quirks. :D

 

An interesting thing for you to do (perhaps?) is to learn how to also test for some of the aberrations and distortions. Have you ever done the crumpled foil test for chromatic aberrations? Or shot any sharpness diagrams? As long as you have learned this much why not continue and learn a bit more about optics testing?? It is rather fascinating. And time-consuming, lol !!!

 

Also it would be of interest to most all of us to see you these lenses for their behaviour in infrared. Do they focus shift between Vis and IR? How long to record images using Schott RG 1000? That kind of thing. I'm also curious to see your sensor modeling for IR as you have done above for UV. How do you feel about tackling the IR side of things?

 

I'm wanting to see a distinctive title prominently displayed for those graphs which model sensor + lens + filter + sunlight. I'm always worried (that is, as an editor/admin) that someone will look at your chart and not quite understand what it is about. Should the y-axis be called Total Response or System Response rather than Sensor Response?

 

And a final question. Do you think that the system you are testing really cannot go below 320 nm? If I were much more confident about those narrowband hard-hardihard-coated Edmund filters, I would order the 320 and the 330 just to see what happens in the 15 or so nanometers around those peaks. But there is so much flare from those darned filters, that I'm quite hesitant to spend the money. :unsure:

[ulf]

Good testing, Jonathan. Thank you for these efforts.

 

Also it would be of interest to most all of us to see you these lenses for their behaviour in infrared. Do they focus shift between Vis and IR? How long to record images using Schott RG 1000? That kind of thing. I'm also curious to see your sensor modeling for IR as you have done above for UV. How do you feel about tackling the IR side of things?

 

 

I agree fully with Andrea's comments.

 

For testing behaviour in infrared, (NIR) I think it would be good to divide the wavelength testing range into two or three smaller parts.

For UV we are at best used to use a range of 60-75nm below 400nm.

In the NIR-end we are talking about a usable wavelength range of around 300nm. That equals the full VIS-range!

 

Ideally the testing could be done with suitable band pass filters, but I think that difference composites of two images with different OG-type filters might work too.

It might be a good idea to ponder the suitable ranges and a good test setup for the experiment.

[JMC]

Thanks Andrea, comments added in below;

 

Good testing, Jonathan. Thank you for these efforts.

 

Have you linked this test to a photo of the CO 105/4.5 in the Lens Gallery? No, not yet.

 

A bit of musing ...... There never is a "perfect" lens for UV is there? Seems like each UV-capable lens has its little quirks, yes? We always have to try to learn how to get the best from each lens whatever its quirks. :D Oh, they are quirky, that is definitely the right thing to call them.

 

An interesting thing for you to do (perhaps?) is to learn how to also test for some of the aberrations and distortions. Have you ever done the crumpled foil test for chromatic aberrations? Or shot any sharpness diagrams? As long as you have learned this much why not continue and learn a bit more about optics testing?? It is rather fascinating. And time-consuming, lol !!! I'll add it to the list. While I would love to dig deeper into optical testing at the moment, I must remember to do some paid work occasionally, otherwise the mortgage doesn't get paid, and more importantly I have no money for lenses.....

 

Also it would be of interest to most all of us to see you these lenses for their behaviour in infrared. Do they focus shift between Vis and IR? How long to record images using Schott RG 1000? That kind of thing. I'm also curious to see your sensor modeling for IR as you have done above for UV. How do you feel about tackling the IR side of things? My monochromator only goes to 800nm, and my spectrometer to 850nm, so I'd need some new equipment to do modelling in the IR. Also, I'm not sure how far the light source goes into the IR. Ideally I'd want to get out to 1400nm to make sure I am at the end of where the cameras are sensitive. As a final thing, new fibers optics collimators, as the ones I have are optimised for Uv and Vis. If the opportunity comes along and someone wants their cameras characterising up there, then I'll definitely look into it, but not at the moment. As Ulf says band pass filters may also be an option.

 

I'm wanting to see a distinctive title prominently displayed for those graphs which model sensor + lens + filter + sunlight. I'm always worried (that is, as an editor/admin) that someone will look at your chart and not quite understand what it is about. Should the y-axis be called Total Response or System Response rather than Sensor Response? Yeah, not sure how best to do that. Been thinking about it myself too.

 

And a final question. Do you think that the system you are testing really cannot go below 320 nm? If I were much more confident about those narrowband hard-hardihard-coated Edmund filters, I would order the 320 and the 330 just to see what happens in the 15 or so nanometers around those peaks. But there is so much flare from those darned filters, that I'm quite hesitant to spend the money. :unsure: The monochrome camera definitely has some usable response down at 320nm, the colour multispectral one much less so. The issue with being down that low is OD blocking in the rest of the range. If the response in the area of interest is so low anyway, then the effective blocking in the rest of the range needs to be much higher than the OD3-OD4 range, otherwise by turning up the exposure you'll just be getting more out of band contributions. My advice, if one comes up cheap go for it and try it, but don't go spending loads on those filters.

[Andy Perrin]

Ideally I'd want to get out to 1400nm to make sure I am at the end of where the cameras are sensitive.

 

JMC, the cameras are not sensitive past 1100nm, 1150nm at the outside, at least with the Bayer/microlenses on. Silicon becomes transparent at that point, so you just can't get anything, at, say 1200nm.

[nfoto]

Film hardly got past 900 nm thus digital is a huge step in increasing the possible range for recording electromagnetic spectrum by direct photographic means.

[JMC]

That's good to know Andy. Still way past where I can measure at the moment though.

[dabateman]

Jonathan thank you for all your open and interesting studies.

On the point of IR, I think it would be interesting. However, much harder to test. The uv problem seems relatively easy. As we are only looking at 300nm to 400nm. A great linear ccd spectrometer can see at 280nm, whereas our cameras as you have shown have limits at 340nm. Also spectrophotometers can see much higher ensergies.

 

IR gets harder as you would need all new equipment. Also I think our cameras are as good as what would be easy to test with. So the limits would run into the instruments used. So ultimately, what we could look fir may not be sensitivity.

 

That all being said, I recently purchased a 1100nm long pass filter with hard cut off at 1080nm. So as Andy pointed out I am interested to see if I can see anything. I probably will only see 1080 to 1150nm, but that is still larger range compared to the uv I can see. I am hoping my Halogen lamps outputs enough light also to see up there, as that too will be a limit.

I will let you know how it goes in a separate post in the future.

[Andy Perrin]

I have a Premium Edgepass filter from Thorlabs that blocks from 200-1100nm to OD5+. I can’t get anything but noise with that one on there. (Its purpose is to allow 1550nm SWIR through, so I can focus it on a fluorescent screen, which then emits 900nm light visible to my sensor.) I only mentioned 1150nm because I wonder if the Bayer+microlenses is stopping a little light from getting through, even though it’s pretty transparent.

[NA_joey]

This is great info thanks i been months since ive been on here.. I bought a used Costal Optics 105mm 4.5 from eBay last year been just finally testing it out more with stacking photos...