Blackmagic Forum

I've read in a thread (that I can't find anymore, I used the search function), that RAWLITE OLPFs have or might have some "problems" with cropped sensor modes. I am using my Pocket 4K mostly in 2.6K Super16 mode and therefore I am hesitant to buy the OLPF.
But I have three questions:

1) What exactly are these "problems"?
2) Are these problems related to the camera-internal "crop"? In this case I could just record the full sensor and do the cropping in post?
3) You are supposed to be able to do the installation alone, but wouldn't that be stupid if you don't have a clean room?

Maybe John Brawley is reading this, I know he has much experience with the RAWLITE OLPFs

Thanks in advance,
Jannik

There is no difference between full sensor or crop mode as far as an OLPF is concerned.
Every OLPF is reducing the resolution to some degree to counteract moiré, and it's a delicate balance. IMHO, the RAWlite is doing a very good job with this balance. Resolution of a 2.6K image downscaled to HD should look even a bit better than native 4K in UHD.
That said, you may experience the resulting image as a tad soft. But you have to keep in mind, that unfiltered images will exhibit some false detail due to aliasing. This is not precise information as derived from the scene, but it can create some illusion of sharpness.

When cropping on G2 / Pockets etc I tend to always lower my iso by half if possible and to sharpen the picture a bit by closing down.

To reduce noise floire and enhance sharpness.

I didn't saw any difference in sharpness with my Rawlite on G2. But to be honnest I never did a proper test.

I’ve never seen any issue with cropped sensor modes and Rawlite OLPF on multiple different BMD camera models. In theory, an OLPF is tuned to the photosite size which is no different in crop vs full sensor. And even with the full sensor 4K and 8K scaled BRAW sensor modes on the 12K UMP I’ve never heard nor have seen any OLPF related issue on my own 12K UMP + Rawlite OLPF.

Unless the thread describing the issue can be found and it can be replicated, I don’t think there is anything to worry about.

Thanks for all the input so far.

Jamie LeJeune wrote:Unless the thread describing the issue can be found and it can be replicated, I don’t think there is anything to worry about.

Thread has been found: viewtopic.php?f=2&t=154840&p=823291&hilit=+known+to+have+a+VERY+AGGRESSIVE+OLPF+#p823291


"The way the 12K sensor works, you're using different groupings of pixels, so when you shoot at 8K or 4K you have a different pixel spacing, some are white and some are coloured and therefore you can see moire come back in.

So an OLPF tries to limit super fine detail that the sensor can't record anyway. But on a multi resolution sensor, what threshold do you choose ?

With RED, they are well known to have a VERY AGGRESSIVE OLPF. Part of the reason their redcode codec appears to be so efficient is that you're greatly reducing thew amount of super fine detail. Yes it's an 8K file, but the bucket of resolution is that 8K sensor though an aggressively softened filter. The downside of this is that when you change from the full native resolution you also have a lot less inherent resolution. So a 1920 window of their sensor looks like garbage compared to an actual 1920 sensor with an OLPF optimised for 1920."

Especially the last part makes me wonder, if I'd get problems in 2.6K sensor crop mode with my Pocket 4K with an OLPF.

Uli Plank wrote:That said, you may experience the resulting image as a tad soft.

I'm using the crop mode with 60s/70s old Angenieux zooms for 16mm, so I have a very soft image to start with. Do you think the effect would add/multiply or just not be noticable?

Well, the 4K has a conventional Bayer sensor, so the discussion above doesn't apply to you.

And then, the RAWlite is less aggressive than the Red OLPF. But if you use those nice old Angies, why not try with some critical patterns if moiré is really a problem? It's already less pronounced on 4K than on the first BMPCC or the 2.5K BMCC.

Uli Plank wrote:And then, the RAWlite is less aggressive than the Red OLPF. But if you use those nice old Angies, why not try with some critical patterns if moiré is really a problem? It's already less pronounced on 4K than on the first BMPCC or the 2.5K BMCC.

I only once(!) encountered moiré on the P4K 2.6K crop with Angenieuxs, but the main reason for me, why I am considering the OLPF is IR pollution.

Well, there are other solutions to IR pollution.

Uli Plank wrote:Well, the 4K has a conventional Bayer sensor, so the discussion above doesn't apply to you.

As far as I know Uli, the Reds have also 'conventional' bayer sensors. Nonetheless I think on balance an OLPF must be advantageous and since Rawlite now officially promote the Broadcast G2 I will be getting one.

Zweistein wrote: Thread has been found: viewtopic.php?f=2&t=154840&p=823291&hilit=+known+to+have+a+VERY+AGGRESSIVE+OLPF+#p823291

What is written in that thread is entirely speculative rather than a report of an empirical issue with steps to replicate it.
I haven't seen anything like what is written there in actual use of the Rawlite filter installed on any of my own BMD cameras — Original BMCC, Original Pocket, 4K Pocket, 6K Pocket Pro, Ursa Mini 4.6K, Ursa Mini Pro 4.6K G1, Ursa Mini Pro 4.6K G2, and Ursa Mini Pro 12K

Correct. But that speculation is about the 12K sensor, so it doesn't apply here. AFAIK, no serious testing out there yet. I suppose those with a 12K and a RAWlite are happily shooting instead of going scientific about it

...and those with a 12K without any RAWlite are happily shooting.

Probably. I had difficulties provoking moiré on a 12K with anything but test patterns and a very sharp lens.
Even then it was minimal.

Uli Plank wrote:Correct. But that speculation is about the 12K sensor, so it doesn't apply here.

I know. It was the last sentence, that struck me:

"So a 1920 window of their sensor looks like garbage compared to an actual 1920 sensor with an OLPF optimised for 1920."

From that I derived the assumption, that a 2.6K window of my 4K sensor on the Pocket 4K might also have some kind of problems, as the OLPF is optimized for 4K resolution.

Thanks for all your input. What about my third question? Did you install your OLPFs yourselves without a clean room?

As already mentioned above, an OLPF has to be optimised for a specific size of the photocells (aka sensels), not for the absolute number of these.

Uli Plank wrote:As already mentioned above, an OLPF has to be optimised for a specific size of the photocells (aka sensels), not for the absolute number of these.

Okay. Sorry I didn't catch that information.

Last question: Would you think it's a bad idea to install the OLPF by yourself, when you do not have a clean room?

Zweistein wrote:Last question: Would you think it's a bad idea to install the OLPF by yourself, when you do not have a clean room?

You are not removing the sensor, just the cover glass over the sensor. I've installed Rawlite OLPFs myself on two cameras in my living room, no issues. Use a bulb blower (never canned air) to blow any dust out of the inside of the camera before you remove the existing glass and then install the OLPF immediately after removing the existing glass; I also gave the OLPF a quick blast with the bulb blower before inserting it.

The trickiest part on the BMPCC models is removing and replacing the o-ring; not sure how it works on the BMPCC 4K but that was the only difficult procedure on my BMMCCs.

Uli Plank wrote:Probably. I had difficulties provoking moiré on a 12K with anything but test patterns and a very sharp lens.
Even then it was minimal.

Before I installed the Rawlite filter on my UMP 12K, I would too often see moire on finely textured fabrics shot on sharp primes. Less commonly visible than what happens on the UMP 4.6K G2, but still there. A couple weeks ago I encountered the exact same minor moire on a finely textured fabric in on a shot I was grading from a brand new Venice 2 using ARRI Signature primes.

Since installing the Rawlite I haven't seen any moire on images from my UMP 12K. The main reason I installed it though was not for the OLPF, but for the IR cut. Yeah, there are other ways to cut IR, but having it taken care of in the sensor glass is the most convenient.

Thanks for the answers. I ordered it.

So I installed the OLPF on my Pocket 4K today, of course shimming with the included shim.

But I am confused now: When I shim a mount on a camera so that the flange distance is correct again, shouldn't then all lenses that worked before still work correct?

My Sigma 18-35mm + Viltrox EF->MFT Speedbooster works fine, but my Angenieux 12-120 with Cameflex->MFT adapter doesn't: at 12mm I can't focus closer than 4 or 5 meters.

To test what's wrong I put a bit of aluminium foil in the Cameflex–>MFT adapter's "space" for shims and voila it focuses correct again. But how can that be? Shouldn't then the Sigma also be incorrect? Why is the RAWLITE shim correcting some lenses and others not, when they are all correctly adapted for MFT?

Thanks in advance.

MFT and EF mount stills lenses have a wider tolerance than cinema zooms.
I'd bet the tolerance needed for your Angenieux is a lot tighter than your Sigma EF.
Even between PL lenses it might differ. I had to shim my DZO PL zoom lenses separately for the same PL mount where my SLR Magic APO PL primes were spot on. Then, for a rented set of Zeiss super speeds, the mount had to be re-shimmed in order to hit infinity.

Also, it might be worth double checking your camera mount to make sure it was screwed in evenly when re-assembled.

Thank you, Jamie.
I will test all that next week!

Jamie LeJeune wrote:Also, it might be worth double checking your camera mount to make sure it was screwed in evenly when re-assembled.

This may be a stupid question, but how do I check, if it is evenly screwed in?
I mean we are talking fractions of millimeters here, nothing that can be recognized by eye, right?

Yes, I'm sorry I wasn't clear. I meant unscrewing and then re-doing the final attachment of the mount just to make sure it was even.

And make sure that you tighten the screws in diagonal positions in small steps, just as you'd do it on a combustion engines's cylinder head.

Jamie LeJeune wrote:I meant unscrewing and then re-doing the final attachment of the mount just to make sure it was even.

Uli Plank wrote:And make sure that you tighten the screws in diagonal positions in small steps, just as you'd do it on a combustion engines's cylinder head.

I tried that, did not change anything.

I guess I have to go to someone with a collimator, who shims my Cameflex-MFT adapter, right?

Zweistein wrote:I guess I have to go to someone with a collimator, who shims my Cameflex-MFT adapter, right?

You can also do it without. These were the instructions I've followed most recently when shimming lenses.

Thanks, I watched the video. But as the error occurred only after modifying the camera mount, shouldn't the shimming also be done on the camera mount instead of on the lens mount?

Yes, but then all your lenses should be off. Is that so?

Uli Plank wrote:Yes, but then all your lenses should be off. Is that so?

I can't tell without measuring. If like Jamie said EF photo lenses have a great tolerance, they might be off as well, just not noticeable to my eye. I borrowed a test chart and will check again, though.

But all my Cameflex Angenieuxs are off:
12-120mm
12-240mm
9.5-57mm
Type R7 5.9mm

I also have a MFT-PL adapter from MTF Services, I will go to a rental house tomorrow and test a modern PL zoom to see if it is off, too.

It doesn't really make sense. The acceptable back focus tolerance should actually be shallower on the Speedboosted 18-35mm combo which, at the wide end, is an effective 13mm f1.3 Vs. the Angenieux which is a 12mm f2.2. Another thread where, in the lack of images, we're left debating the OP's impressions.

Good Luck

Howard Roll wrote:It doesn't really make sense. The acceptable back focus tolerance should actually be shallower on the Speedboosted 18-35mm combo

Interesting. That indeed make this case even more counter-intuitive for me.

Howard Roll wrote:Another thread where, in the lack of images, we're left debating the OP's impressions.

No need to be so salty, just ask for images, I gladly provide them. As I stated above, I got a test chart now and will do more tests. I can then share the results.

Zweistein wrote:Thanks, I watched the video. But as the error occurred only after modifying the camera mount, shouldn't the shimming also be done on the camera mount instead of on the lens mount?

If only lenses mounted with your MFT to PL adaptor are out, then I would simply shim that adaptor rather than the lenses.

Jamie LeJeune wrote:If only lenses mounted with your MFT to PL adaptor are out, then I would simply shim that adaptor rather than the lenses.

Right now it seem only lenses with my Cameflex to MFT adapter are off (albeit there were fine before OLPF modification).

I only mentioned the PL to MFT adapter, because it was recently professionally measured and shimmed by a lens technician. So if this is off, too, we know it's the camera's mount. If it is fine, it's rather the Cameflex to MFT adapter, who's causing problems.

We'll see today.

My lens technician of trust used a flange distance controller and measured that exactly 0.1mm is missing.
He tested my camera a few weeks ago and it was perfect. So the 0.1mm difference must come from the RAWLITE modification. I guess somehow a wrong shim found its way into my RAWLITE package. I wrote Hans from RAWLITE an email, I hope he will provide me with the correct shim.

I had to shim my UMP 4.6K after the RAWlite mod. Now my zooms are parfocal again.

Hans from RAWLITE will see if he has a 0.1mm shim for me, when he returns to office July 10.
Meanwhile i used a 0.1mm leftover shim from my PL adapter and cut three fitting pieces from it. Works perfect by eye, but I still have to do some proper chart tests and see it on big high-res screen.

I appreciate the detailed updates on it. It’s useful to know that the wrong shim was sent in the kit. I’ll be looking out for that next time.

Jamie LeJeune wrote:I appreciate the detailed updates on it. It’s useful to know that the wrong shim was sent in the kit. I’ll be looking out for that next time.

You're welcome. I appreciate your input.

To clarify: Hans said it can't be a wrong shim, because there is only one shim for the Pocket 4K on his side.
He assumes it is due to a tolerance in the built of the Pocket 4K camera. Which still would be weird, because then my Angenieux lenses with the Cameflex adapter would have been off in the first place, right? Still not a fully solved mystery, but at least a solution for it, as it seems.

I had to add another shim too. Other than PL, mounts for stills cameras have pretty high tolerances.
Not Hans‘ fault.

It’s very normal to be both monitoring and changing your shims on your lenses and mounts over time.

There can be a lot of copy to copy variation as well, also in both cameras and lenses.

This leads to a lot of confusion when something “used” to be right and now isn’t.

I used to prep film cameras and we had very accurate physical depth gauges.

Sometimes the act of undoing the screws and doing them up again would be enough to change the FFD by a critical amount. A bit of dirt on a shim, temperature, uneven tightening. These can all change the FFD.

Putting an extra optical piece between the back of a lens and the sensor also definitely changes the FFD.

Hans is doing a “best guess” of the shim but you may have a camera and lens that together are just the right combination to be at the outer most acceptable tolerance and THEN you’re adding a third thing that can change those numbers.

Again, FFD changing is borderline routine. On a large long running job, I’ve sometimes had to re-shim both lenses and cameras in the middle of the shoot.

It’s one of the first thing a good first AC checks when picking up a rental camera package. You should just assume it’s out till you can confirm it for yourself.

The good thing is that it’s pretty easy to check for yourself and it’s easy (if a bit tedious) to alter. A lot of people don’t understand how to check these kinds of issues methodically and systematically so it’s hard tk then troubleshoot.

JB

Thanks for the input, John!

John Brawley wrote:You should just assume it’s out till you can confirm it for yourself.

Maybe that's a good state of mind regarding this topic.

John Brawley wrote:The good thing is that it’s pretty easy to check for yourself

You mean with a FDC?

So the best thing for checking a body is something like this.

https://www.pure4c.de/accessories/measu ... -fdc-light

That will tell you if the mount is good or if it’s out, how much you need to re-shim.

If you don’t have one the next best thing is to take a “known good lens” and use that.

That of course is hard to know without a lens collimator. But, the best way is to use something wider, as FFD focus is more critical the wider your lens. So you get a (say) 12mm fast prime that you know is good and set it to infinity, go outside and find something that’s at least a couple of hundred feet away and check to see if it’s in focus at infinity.

This only works if you have a “good known lens” and most people probably don’t have that.

To get a good known lens you use one of these

https://www.chrosziel.com/collimators/

And then shim the lens so it is perfect for the FFD you set.

If you’re super nice to your local camera rental company they may let you collimate your own wide lens so you can then check your mounts.

JB

John Brawley wrote:So the best thing for checking a body is something like this.
If you don’t have one the next best thing is to take a “known good lens” and use that.

Yes, my go-to lens technician at my go-to sales/rental house uses a Denz FDC.

Will keep the “known good lens” method in mind.

Thanks to all here for your input.

Follow-up to my original problem from 1.5 years ago, which is now solved!

Last year I purchased the RAWLITE OLPF for my Pocket 4K and after installation (including the shim that comes with the OLPF) my flange distance was off, which I noticed because my parfocal lenses were not parfocal anymore and my lenses in general did not hit close focus and/or infinity anymore.
I wrote Hans from RAWLITE and he sent me a few different shims for the Pocket 4K to try with my OLPF. But at that time I was right before the shoot of a feature and I did not have time and mind to fiddle around more at my mount/sensor, I just wanted a functioning camera, so I de-installed the OLPF and my lenses were parfocal again. After the shoot I had more important things to do and also not really motivation to try the other shims.

Now, 1.5 years later I gave it another try and the combination of the P4K's original shim (0.3mm) and one of the new shims RAWLITE sent me (0.38mm) made the flange correct with the OLPF installed. My lenses are parfocal, hit close focus and (beyond) infinity. Tried it with the MFT>PL Mount and the MFT>EF mount.

So everyone who has flange distance problems after RAWLITE OLPF installation, ask them for different shims. It solved the problem for me.