Much needed OLPF coming for the UM4k and UM4.6k

After encountering a very high level of Morie' on a resent shoot with my UM4.6, I contacted MOSAIC Optoelectronics to see if they had or were working on a OLPF filter for the UM cameras. The following is the very thorough reply I received from David Cubanski of Mosaic:

YES - we are definitely proceeding to bring out a filter for the Ursa Mini (both versions). We've made a couple of prototype filters already; performance is similar to that of our BMCC / BMPCC / BMMCC filter. For the Ursa Mini, we're furthermore proceeding to design a revised filter, and right at this point we're trying to get this new filter tested even this week. Here's the basic idea: Our standard BMCC / BMPCC / BMMCC filter, gets rid of "most" aliasing and moire - meaning it works very well most of the time, though it sometimes leaves small residues of aliasing artifacts even in the filtered video, though these are attenuated to a low level that's generally not very visually distracting. For the Ursa Mini, we think we can do better - actually to remove "all" aliasing and moire, without exception - so that filtered video will basically be completely clean, similar to the experience with a camera like an Alexa. We think right now that we can do this without significantly compromising the native sharpness and resolution of the Ursa Mini. This will be a more complex filter, though especially in the case of the 4.6K Ursa Mini, we feel that the additional complexity will be justified by the level of performance that it enables.

Sorry that we haven't posted information about this effort on our web site; I'll keep your email here and let you know as things proceed!

Thanks again and best regards--

--dave cubanski

David Cubanski
Mosaic Optoelectronics, Inc.
210 W. Hamilton Ave. #290
State College, PA, 16801, USA
(+01) 814-359-6290

This is exciting!

Sent from my Pixel XL using Tapatalk

Well Benton, looks promising. Thank you for taking the initiative on this. If they can eliminate virtually all moiré and aliasing, they should sell the proverbial boatload of them!


Sent from my iPad using Tapatalk

Very interesting news. If you are in touch with them Benton, can you ask them if it would be possible to create a version with IR cut added as well. I've got a Hoya IR Cut living on my Ursa Mini pretty much all the time. Would be great to have it built into the camera and not even have to think about it.

Yes an IR cut would make this even better.

Very interesting news. If you are in touch with them Benton, can you ask them if it would be possible to create a version with IR cut added as well. I've got a Hoya IR Cut living on my Ursa Mini pretty much all the time. Would be great to have it built into the camera and not even have to think about it.

Great suggestion Jamie! Having IR cut built in would make that new filter the ultimate Bandaid! I will definitely pass your suggestion on.

No it wouldn't! You would lose the creative control over where you cut IR, how severely you cut, and so on. BMD is doing us a favour by having a very mild (if any) reduction on IR. You're in danger of losing some of those beautiful skin tones if you cut too soon. And the IR Cut filters perform differently affecting different wavelengths as you vary the lens focal length. For example a 680nm cut may be good for a normal lens, but a wide angle may fair better with a 715 or 750nm IR Cut. Why else do you think Schneider gives you these three options? Please think before you leap with that suggestion. Thank you.


Sent from my iPad using Tapatalk

I agree with you in this Rick, same goes for the OPLF filter, sometimes you are better off without one. A nice solution would be the old film camera trick, of a behind the lens filter slot in the camera one could insert the desired filter, like a OPLF and/or IR of choice.
Cheers

Please think before you leap with that suggestion.

No need to be so harsh Rick...Jamie was just trying to help and to be part of the discussion,
and I think most of us agree with him too.

Furthermore, most of the times a well tuned IR filter (it can be tuned perfectly for the sensor in this case),
not only produce better results than a trio of generic ones, but it's also much better than dealing with filters (expense, time involved, cumbersomeness, reflections, etc etc) and optically superior not being in front of the light path.

Exactly, good point. That is why I would like to see a return of the behind themlens filter slot of old! It would be a nice compromise of both approaches.
Cheers

No it wouldn't! You would lose the creative control over where you cut IR, how severely you cut, and so on. BMD is doing us a favour by having a very mild (if any) reduction on IR. You're in danger of losing some of those beautiful skin tones if you cut too soon. And the IR Cut filters perform differently affecting different wavelengths as you vary the lens focal length. For example a 680nm cut may be good for a normal lens, but a wide angle may fair better with a 715 or 750nm IR Cut. Why else do you think Schneider gives you these three options? Please think before you leap with that suggestion. Thank you. Sent from my iPad using Tapatalk

While I agree that a 680nm cut would be too steep to have that always baked in, I think something closer to 700nm (which is where the Hoya cuts) would be a safe bet. I have compared the Schneider True Cut 680 and the Hoya IR cut and the Hoya lets in a full saturation of deep visible red while the 680 mutes deep red colors too much. Yet even with the Hoya, traces of IR pollution can still be detected in blacks, but it's so minimal that it's totally acceptable to me. I would welcome to at least have the option of having a OLPF that has IR cut built in if it is fully tuned for the best performance of the sensor. Also, I was not aware that IR pollution changed with focal length. I always thought it was strictly a response of the sensor. I learn something every day!

FWIW I talked to them last May and they said the same thing. I hope its legitimately moving forward.

Tommaso, sorry if I sounded 'harsh' as I just intended to put forward a different opinion. Could have been more gentle of course! Having the IR Cut component being optional would satisfy those who want it included and those who did not, so I like that compromise.

I'm planning to do another test illustrating the effect of a wide angle lens (25mm) versus a long normal (50mm) and short telephoto (85mm) with my Schneider 715 IR Cut (715nm) and my B+W 486M IR Cut (680nm) and my Image Enhancer Pro (IR reduction) filters once I get my Misfit Atom matte box in a week or so and my APO lenses returned later this month. Hopefully that test will be instructive. And as a bonus I'll even try to have my model in focus this time!

Wish I had an even wider lens to illustrate the point. Or perhaps Anatoly will kick in and explain the reasoning that filters on wider lenses cut differently than longer lenses. In other words, if you have a 680 nm cut filter that looks great on a longer lens, you'll find on your widest lens that it is cutting too far into the red wavelengths.

So my solution is trying a range of strengths. My guess is the 715nm will work well on my 25mm APO, but when I get an 18mm APO for example, I'll pick up a Schneider 750nm IR Cut for that lens. Apologies I can't illustrate all this yet! When I have all six lenses, I'll try to redo the IR pollution test again.


Sent from my iPad using Tapatalk

Great news, a 18mm APO lens is in the production pipeline, that was the one missing necessary focal length for the UM 4.6
That sill give you a nice basic prime set. Hope your other APOs show back,up,from SLR Magic soon
Cheers

Tommaso, sorry if I sounded 'harsh' as I just intended to put forward a different opinion. Could have been more gentle of course! Having the IR Cut component being optional would satisfy those who want it included and those who did not, so I like that compromise.

I'm planning to do another test illustrating the effect of a wide angle lens (25mm) versus a long normal (50mm) and short telephoto (85mm) with my Schneider 715 IR Cut (715nm) and my B+W 486M IR Cut (680nm) and my Image Enhancer Pro (IR reduction) filters once I get my Misfit Atom matte box in a week or so and my APO lenses returned later this month. Hopefully that test will be instructive. And as a bonus I'll even try to have my model in focus this time!

Wish I had an even wider lens to illustrate the point. Or perhaps Anatoly will kick in and explain the reasoning that filters on wider lenses cut differently than longer lenses. In other words, if you have a 680 mm cut filter that looks great on a longer lens, you'll find on your widest lens that it is cutting too far into the red wavelengths.

So my solution is trying a range of strengths. My guess is the 715nm will work well on my 25mm APO, but when I get an 18mm APO for example, I'll pick up a Schneider 750nm IR Cut for that lens. Apologies I can't illustrate all this yet! When I have all six lenses, I'll try to redo the IR pollution test again.

No worries. It is a wonderful thing when people will tell me I'm dead wrong. I mean that seriously. It's a great way to learn new things

I don't doubt that what you say is true, but every digital sensor has it's unique quirks and I wonder whether anyone proven this to be true in comparison tests with the Ursa Mini 4.6K? In my own work with it, I find I like its images best when the Hoya IR Cut filter is on, including the skin tones no matter the lens. If you know of a test that shows a comparison between having that filter on the lens and off the lens on an Ursa Mini 4.6K where it honestly looks better without the Hoya, I would really love to see it. As I understand it, the Schneider IR filters cut more abruptly than the standard Hoya IR which may lead to some different results between them.

Appreciate your remarks, Jamie. Hopefully the upcoming tests will demonstrate the point even though my range of focal lengths with the primes is currently limited to 25, 50, 85. I could use the Fujinon but it's only a sensor crop and it will work with my 4" Schneider 715 IR Cut in a matte box but I'd have a hard time accurately and safely filter whacking the B+W 486M 86 mm circular filter against the very large front element of the Fujinon. I could give it a try though as long as people would accept the edges of the frame will vignette the circular screw-on filter. Not necessary if I can demonstrate the effect even with the 25mm APO, but if it isn't clear, I may attempt another test with the Fujinon at much wider angles.

To illustrate the effect on black is not a problem, but I need to think about how to illustrate skin tones as well. Ideally I'd like to find a very light skinned model with naturally good colour ranges in her complexion (and a dark skinned model if possible). That might be a challenge for me, but I have a few Christmas parties here this month and I might be able to enlist the help of a guest.

I think what I shall attempt to illustrate has not been done that I am aware of. Unfortunately I don't have the Hoya, but I believe it cuts gradually in a way very similar to the B+W 486M. Schneider does have the wavelength responses of all their filters on a public webpage so you could compare it to the Hoya's chart. The 486M response curve is not conveniently isolated on its own chart, but included with several other filters.


Sent from my iPad using Tapatalk

Rick, no doubt that a IR filter in front of the lens will need different nms for different focal lengths.
But I am quite sure that there are no different nms needed when the IR cut filter sits directly in front of the sensor.

I do believe Robert is right. The different lengths need for in front of the lens has to do with then angle at which they pass through on a wide angle which can cause color shifts in the corners. Behind the lens should be fine as it is realitively always hitting from the same angle. I'm not an optical engineer so don't take my word for granted though.

Furthermore, most of the times a well tuned IR filter (it can be tuned perfectly for the sensor in this case),

Tuning still requires subjective judgement. That's the point. This is the case for both IR cut and the amount of 'softening" to reduce moire. As mentioned in the re-posted email, they've made a judgement about how much to reduce it.

Just like with a diffusion filter, you can choose the degrees of effect...same with IR cut and OLPF for that matter...That's why it takes so long to develop an OLPF. It's a very itterative process.

JB

Rick, no doubt that a IR filter in front of the lens will need different nms for different focal lengths.
But I am quite sure that there are no different nms needed when the IR cut filter sits directly in front of the sensor.

Isn't the whole magenta issue proof positive that the angle does matter, even behind the lens.

Unless you have a telecentric lens like some Schneider primes, the angles from the wider lenses hitting the sensor could be a factor. That's an argument for testing the lenses with IR Cut filters filters you have so you can make your own conclusions.

Having the IR Cut as part of the sensor glass or a third-party OLPF is arguably not the best solution for every situation. Again, if the third-party solution was easy to insert and remove, it may be a choice for many people.

Having the IR Cut at the front of the lens though (screw-on, snap-on or matte box) allows for the greatest flexibility and simplicity. I don't think I'd feel confident about repeatedly placing and removing anything near the camera sensor. Others might, but I get apprehensive just changing PL mount shims although I've done it several times and never dropped a screw on the sensor.

I think it's interesting that only Schneider appears to offer a range of IR Cut filters and their website which includes some commentary on the rationale although it doesn't get into defining when to use each of their options. They just say it's up to the operator. I read up on this when buying their 715 IR Cut as some of you may recall several months back. As far as I'm aware, they are alone in offering the options, but don't want to open up the discussion. Anyone know of other companies supporting multiple design criteria for their IR Cut?


Sent from my iPad using Tapatalk

Tuning still requires subjective judgement. That's the point. This is the case for both IR cut and the amount of 'softening" to reduce moire. As mentioned in the re-posted email, they've made a judgement about how much to reduce it.

Just like with a diffusion filter, you can choose the degrees of effect...same with IR cut and OLPF for that matter...That's why it takes so long to develop an OLPF. It's a very itterative process.

JB

Based on your use of the Ursa Mini 4.6K so far, do you prefer any particular IR cut filter? Do you change your IR cut based on lens choice?

Based on your use of the Ursa Mini 4.6K so far, do you prefer any particular IR cut filter? Do you change your IR cut based on lens choice?

I haven't seen any difference based on lens choice or focal length choice.

I've used Mitomo TRUE NDs for years and I've done work with Firecrest and recently the new Panavision ND
s.

https://www.flickr.com/photos/johnbrawl ... 3844822085

These are the only brands I would use with the UM 4.6K. (and Hoya PRO ND's in screw in)

Oh. And the SLR MAgic for Vari ND's also very good if you have to go variable ND.


JB

Based on your use of the Ursa Mini 4.6K so far, do you prefer any particular IR cut filter? Do you change your IR cut based on lens choice?

I haven't seen any difference based on lens choice or focal length choice.

I've used Mitomo TRUE NDs for years and I've done work with Firecrest and recently the new Panavision ND
s.

https://www.flickr.com/photos/johnbrawl ... 3844822085

These are the only brands I would use with the UM 4.6K. (and Hoya PRO ND's in screw in)

Oh. And the SLR MAgic for Vari ND's also very good if you have to go variable ND.


JB

Thanks for your observations John! For the life of me, I couldn't see how different focal lengths would react differently to the same amount of IR cut filtration. Plus even if it did make a difference, I can't imagine having to change to a different IR cut filter every time a lens was changed. That would drive me crazy!

John is referring to ND filters. I'm talking about IR Cut filters. He's right about ND filters. I'm waiting until I do some testing to see if I can demonstrate a difference with the limited resources I have.

The most interesting finding from these tests will not be in the infrared spectrum, but what's happening in the deep red, near infrared spectrum and how that may affect skin tones.

Sent from my iPad using Tapatalk

John is referring to ND filters. I'm talking about IR Cut filters. He's right about ND filters. I'm waiting until I do some testing to see if I can demonstrate a difference with the limited resources I have.

The most interesting finding from these tests will not be in the infrared spectrum, but what's happening in the deep red, near infrared spectrum and how that may affect skin tones.

Sent from my iPad using Tapatalk

All of the ND filters John is referring to have IR cut built in. We would still be seeing the effect of IR cut across all lenses that these filters are used on.

Not necessarily, as the UV/IR combo,filters are not cutting IR at the same level as the straight IR cut filters that Rick is using. The partial IR cut in the UV/IR filters may not be as noticeable, as it would be with a Schneider 715 IR Cut filter
Cheers

John is referring to ND filters. I'm talking about IR Cut filters.

Indeed,

I don't really bother with the practise of separate IR cut filters.

One thing I should also mention with the above post, was that they are lit with tungsten lights, actual tungsten lights that happen to emit a lot of IR. It's always hard to tell how much actual IR light is floating around. That's why when I test for IR pollution, I try to use tungsten lights, because they happen to emit a lot of IR light and it's a constant. Outdoors, depending on time of day, smog, clouds etc it can change a lot, but not obviously and to your eyes.

In the linked images, we have two skin tones and look at the difference in the darker ladies arms, and we also have synthetic colours (her black top) and the white / cream top shows colour bias. You can even see that in a heavy IR environment, even the Alexa unfiltered has some IR contamination in the blacks with a bit of a green bias in WB.

As already mentioned, there's no absolutely correct IR filter choice, as the amount of IR and the sensor's reaction along with different sensors will all wreak havoc with consistency.

I prefer to sue IR reducing ND's like the true and Mitomo rather than specific wavelength cuts...

jb

Not necessarily, as the UV/IR combo,filters are not cutting IR at the same level as the straight IR cut filters that Rick is using. The partial IR cut in the UV/IR filters may not be as noticeable, as it would be with a Schneider 715 IR Cut filter
Cheers

I look forward to seeing tests that demonstrate this. I have certainly seen dramatic differences in skin tone when different IR cut filters are applied, but never when those same filters are used on different focal lengths.

Sorry it's confusing Benton...

Unless it says Alexa, it's an Ursa Mini 4.6K.

The IR cut was one of the Schneider IR cut filters, but I can't recall which. I was testing the scheniders with the IR cut against the PV filters. They couldn't get me any TRUE's and offered me the PVs, which still hadn't really been released when I tested them. I think they're wonderful filters if that can be said of ND filters...they have a great "look"

Panavision's new NDs, like TRUE, Nisi, Firecrest and Hoya PRO filters, they spray coat them and the IR cut is seemingly built in, though they don't tend to market them specifically in this way.

http://panavision.com/panavision-unveil ... nd-filters

jb

Sorry it's confusing Benton...

Unless it says Alexa, it's an Ursa Mini 4.6K.

The IR cut was one of the Schneider IR cut filters, but I can't recall which. I was testing the scheniders with the IR cut against the PV filters. They couldn't get me any TRUE's and offered me the PVs, which still hadn't really been released when I tested them. I think they're wonderful filters if that can be said of ND filters...they have a great "look"

Panavision's new NDs, like TRUE, Nisi, Firecrest and Hoya PRO filters, they spray coat them and the IR cut is seemingly built in, though they don't tend to market them specifically in this way.

http://panavision.com/panavision-unveil ... nd-filters

jb

Ha. Yes, I deleted my post after I realized I didn't know what I was looking at. Lol. Thanks for the clarification!

I don't have rectangular ND filters yet although NiSi is my leading choice. So if I buy NiSi, before the cows come home, then I'll be able to test those for their infrared and deep red behaviour along with the Schneider 715, B+W 486M, and SLR Magic Image Enhancer Pro. From what I've noticed the manufacturers can be reluctant to market their ND filters as IRND because they don't cut IR like a dedicated filter. They do reduce IR and that may be sufficient as John has found in practice on those quality brands he uses.

Since we know the Image Enhancer also reduces IR, I'm also curious to see if it reduces IR more effectively for a very wide angle versus a longer lens.

I think this particular set of tests I'm planning on doing with gear I own is of interest to me to see how I can best use what I have. My small market town isn't going to loan me items for the sake of testing and I can almost guarantee they don't even have those items in stock for sale. So I can't comment on the Mitomo True ND and so on.

The test I did publish on IR pollution was shot fairly quickly outdoors summer in the golden hour with my model and black material catching the blazing warm sun. No shortage of IR rays that time. Trying to repeat that test in our winter isn't going to have the same intense warmth. I agree that not having controlled lights is less than ideal, but it can still be instructive as those earlier tests demonstrated, for example, showing the 715 was not as effective as the 486M on my 50mm lens. But if I can show the 715 looking better on a 25mm lens than the 50mm or 85mm lenses, my point is proven. But having the 18mm APO (or whatever wider prime SLR Magic releases) would be an easier test of the concept and I don't know when that lens will be available.

Apologies to Benton that my concern about adding IR Cut to the OLPF has derailed the original discussion into a bit of a debate on a different aspect of controlling infrared radiation. I think I have gone on too long here already! Sorry, guys and gals.

When I put up the next test, I'll add it to my existing thread on infrared contamination.

Sent from my iPad using Tapatalk

Rick, I got a couple of NiSi 4x4 filters you are welcome to use in your test. Just let me know.
Cheers

Apologies to Benton that my concern about adding IR Cut to the OLPF has derailed the original discussion into a bit of a debate on a different aspect of controlling infrared radiation.
Sent from my iPad using Tapatalk

No worries at all Rick! I'm very interested to seeing the results of your test. Plus I'd be surprised if Mosaic Engineering would produce an OLPF with only IR cut built in. Chances are, they'll just offer a standard version. But if they can come up with a nice mid range option with IR built in, that could appeal to a lot of people.

Yes, yes it would, including me!
Cheers

any updates on this filter?

yeah, any updates??

The filter isn't available yet, but they're making progress! It will include IR filtration: http://mosaicengineering.com/products/vaf/bm/um.html

David responded to my email and said that the BMMCC, BMCC and BMPCC Mosaic V2 OPLF started shipping last week and they are still trying to fulfill backorders. The URSA Mini OPLF will be for sale in about 4-6 weeks. He said that URSA Mini version will cut all of the moire and aliasing due to the fact that he has more room to play with since you can't put a metabones adapter on it.

So, it is available for the pocket?

Any sample images around?

Any sample images around?

On their website. Looks promising.

Sent from my SM-N910F using Tapatalk

Thanks Ben.

The OLPF from Mosaic engineering for the Ursa Mini 4 and 4.6k is now available and at discount for early adopters: http://store.mosaicengineering.com/VAF- ... _p_46.html It is available with or without built in IR filtration.

Results are very good on the URSA Mini 4K camera with 6.5 micron photosites. I haven't had an answer though if this will be as effective on the URSA Mini 4.6K sensor given its smaller 5.5 micron photosites. And is the bayer arrangement identical on the 4K sensor to the 4.6K sensor? Definitely interested but will wait until that is demonstrated to work as well.


Sent from my iPad using Tapatalk

The mini 4k's sensor is 21.12x11.88mm at 4000x2160. The URSA 4k is listed at 21.12x11.88 at 3840x2160. I'm guessing the mini 4k uses a slightly wider area of the sensor. The 4.6k is 25.34 x 14.25mm at 4608x2592. URSA 4K 21.12/3840=5.5 microns. URSA 4.6K is 25.34/4608=5.5 microns. Seem to be the same size photo-sites.

The question is, IR or not? Seems that with the Pro with the build in ND filters, we may want the non-IR version?

I'm wondering if this OLPF would even work on the Pro given that the Pro has a physical filter wheel which could present a mounting problem for the filter.

Dan, the 4K sensor is 22x11.88mm for 4000x2160 5.5micron photosites. I could have sworn the generation 3 OLPF video mentioned 6.5microns (which is the size for the BMCC/BMPCC sensors). My mistake and thanks for correcting me!


Sent from my iPhone using Tapatalk

The OLPF from Mosaic engineering for the Ursa Mini 4 and 4.6k is now available and at discount for early adopters: http://store.mosaicengineering.com/VAF- ... _p_46.html It is available with or without built in IR filtration.

Fantastic news! Thank you for the heads up on the release Benton