DSMC 1 OLPF DRAGON 6k

[Mike Krumlauf]

I recently moved from the EPIC MX to EPIC DRAGON and I wanted to get some peoples opinions on this.

My camera came with the Skin Tone Highlight OLPF and I have a Low Light OLPF on the way. I've been told by multiple people that the Low Light should really be the only OLPF I use on the camera as the noise floor is lower and dynamic range increases.

Is there any real reason I should hold onto the Skin Tone OLPF or am I safe to sell it and recoup a little cash back into my pocket.

Where would the skin tone excel over the Low Light OLPF on the OG Dragon DSMC Epic Body?

Thanks

 

[Les Hillis]

The pro's and con's of the Lowlight, Standard and Skintone-Highlight OLPF's are pretty straightforward in terms of how they filter the light hitting the sensor.

Using the same amount of light -

The Standard OLPF distributes the light in a similar way to the Dragon sensor itself, balancing lowlight and highlight exposures.

The Lowlight OLPF lets more light in and so you get more exposure in the 'shadows', but at the same time the highlights are more prone to overexposure.

The Skintone-Highlight OLPF does the opposite and blocks more light, so you get less exposure in the 'shadows', but at the same time also slightly darken the highlights.

Where it gets less straightforward is how those changes in exposure are perceived, objectively and subjectively.

Objectively, you get about a half to a two-thirds of a stop exposure difference either way.

Subjectively, you're dealing with differences in the mid-to-high and mid-to-low ends of the image, where personal opinions and preferences play more of a part in what is considered acceptable in any given image, eg. how noisy is noisy, how clippy is clippy etc.

At the same time, there are also some objective colour differences between the OLPF's, especially with the Skintone-Highlight OLPF.

Nowadays, with more in-depth but easier to use colour-correction available in post, it may be less of an issue making one look more like the other, but the different OLPF's do render slightly different colours on the same sensor.

If you have the Lowlight and Skintone-Highlight OLPF's you'll be able to shoot some tests to see if you have a general preference for one or the other. The tendency is to just pick one and use it all the time, but they are interchangeable, and you could keep them both and change them when and if you have a predominance of dark or brightly lit scenes.

Similarly, you could take a post-processing/grading approach that allows the particular OLPF used to become less significant. Especially with an adequate level of lighting control, eg. lightening the 'shadows' and darkening the highlights, the particular OLPF used could become even less of an issue.

And of course, there's still a point you have to deal with where how you shoot any given shot is beyond the range of influence of any of the available OLPF's.

Not that it matters, but my personal preference has always been for the Skintone-Highlight OLPF. I prefer the colours it gives and dislike dealing with clippy highlights more than noisy shadows.

Lastly, if choosing one OLPF for all occasions, the Standard OLPF probably makes sense.

 

[Les Hillis]

So, curiosity got the better of me and I had a look for myself at the SkinTone-Highlight, Standard and Low-Light-Optimized OLPF's with the Dragon sensor.

This wasn't a scientific test, but it was pretty well-controlled. Same camera and lens, same lights, adjusted the exposures by changing the shutter angle. Made sure to select the correct OLPF in the camera menu and to keep the Temperature and Exposure indicators in the Green, changing the actual blackshaded calibrations as necessary.

Four stops over and under-exposed from the middle exposure measured at ISO800 was enough to see what I wanted to see, didn't need to push it further than that.

Shot at 24fps at 5K on a DSMC1 Scarlet Dragon that can't do 24fps at 6K, but I don't think the lower resolution at 5K changes the basic outcome.

The first video shows the lighter and darker images and the second video shows the over and under-exposures brought back to match the middle exposure.

There is a difference between the OLPF's, but once corrected and even matched to the extent that I was able to do it here (not being an actual colourist), the visible differences are more subtle than I remembered them being, in terms of colour and noise.

The actual maintaining of colour information at the extremes was better than I remembered it, it's just that it gets obscured by the noise (and/or the darkness itself) when underexposing.

Basically, they all look good, within the limits of the Dragon sensor itself.

Any one of the three OLPF's could be kept installed and used all the time imo.

 

[Les Hillis]

Here's a crappy over-compressed picture of what the actual OLPF's look like btw -



And here's a random flower also captured with the Scarlet Dragon, just because -



I'm left thinking the (post-Dragon sensor) introduction of the IPP2 workflow, especially with REDWideGamutRGB/Log3G10 (which replaced DRAGONcolor2/REDgamma4 for me), kind of ironed out some of the inherent differences between all the RED sensors and OLPF's. There are still other differences between the camera's and sensors of course, but the fact that the DSMC3 camera's didn't come with the same range of interchangeable OLPF's as the DSMC1&2 camera's seems to support that idea.

It's often been said that the new improvements in the RED 'color-science' can give new life to old RED R3D footage and I think I can see from these recent tests how that is so. Especially also with the advancements in programs like Resolve, which allow more image control than was readily available, or even around at all, back when the Dragon sensors were new.

Not that the different OLPF's aren't doing anything at all, but I think you'd really have to be trying to stretch your image down to the last percentage to pick one over the other.

 

[Mike Krumlauf]

Thanks for the info. Upon testing them myself, I personally happen to like the image coming from the Skintone over the Low Light honestly. I think I might keep both of them and use lowlight where I see fit. I had no idea they made a Standard OLPF for DSMC1 Dragon bodies, thought it was just Low Light And Skintone.

Seems like the Dynamic Range difference and noise floor differences between OLPFs are subtle.

 

[Les Hillis]

No worries.

It does seem like a more subtle difference between the three OLPF's now.

If I remember correctly, the Standard OLPF got overshadowed a bit by the other OLPF's that had more of a noticeable effect on the previously 'normal' image, even though the new 'Standard' OLPF wasn't exactly the same as the 'standard' non-interchangeable OLPF had been. It's a pretty good balance between the other two imo, tending more towards the LLO than the STH.

RED's coming out with a whole new camera around the same time, in the Gemini, also kind of puts what you can get out of an OLPF in perspective when it comes to increasing low-light performance. There were also some significant improvements in the 'color-science' itself that changed the appearance and performance of the way the sensors moved into under and over exposure too...and on it goes.

The whole OLPF thing is actually a big subject with lots of tangents and inter-related details, just within the RED ecosystem itself.

For what it's worth, after looking at the OLPF/Dragon combinations again myself, the Skintone-Highlight went back in at the end. For better or worse, there is something about what it does to the colours that I prefer and that I can't quite match in post using the other OLPF's, even as there are so many other factors that have more of an impact on the creation of an image than that.

Thanks for raising the subject and giving me an excuse to re-assess my opinions on it.

 

[Alex Lubensky]

I was also digging into the OLPF stuff recently due to another Epic body I've bought (Dragon this time, making it the 4th Epic body I've purchased). It's def. not that fast to change the OLPF, in fact the screws and the idea is the same as on the MX (yes one could swap OLPF on the MX, in fact all early M-bodies were opted to because of the gate flare produced by the materials used in the OLPF surroundings).

I've done some tests myself and to me the LLO is closer to the MX image-wise and color-science wise. It tends to shift neutral grays towards magenta side, has pretty much the same issues that MX olpf had (magenta orbs when closed down and pointed at sun, can be seen in almost any film shot on Dragon in the aerial shots etc), some magenta glare which is basically produced by the same coating. It does seem to have less IR contamination compared to MX though.

Skintone-highlight lets less light in, as already mentioned. Doesn't have the orbs, internal reflections are minimized. It tends to go green tint rather fast in the shadows, so I rate this OLPF at ISO400, which minimizes the amount of green in the shadows. It has more noise in the shadow range of 2 stops under and further, but below that it's quite the same. I do like the skintones and colors more using the skintone-highlight OLPF, but it really thrives for light - I'd say anything shot outside on bright sun is where I'd use the skintone-highlight, otherwise I'm sticking with LLO most of the time (Interiors, interviews, outdoors when the weather is cloudy or the set is long enough to cover vastly different sunlight conditions). Almost any big production was using some kind of LLO also - be it Gone Girl (used pre-production cams with 1st iteration of LLO which was non-removable), Guardians Of The Galaxy Vol.2 (yes orbs in the opening sequence aerial), Martian also utilized LLO as I recall, but might be mistaken.

Standard is a mix of LLO and Skintone-Highlight, a bit closer to LLO maybe. I'd def use Standard over LLO as my main OLPF, but sourcing it for DSMC1 is kind of a hassle. Some people swap the Standard olpf glass from the DSMC2 into the DSMC1 gasket (have to have a spare OLPF to exchange the glass). Overall I find comfortable rating Epic Dragon at @800-1600 with LLO, and 400-640 with Skintone Highlight.

 

[Les Hillis]

Thanks for the info and observations Alex.

I found if I left the camera set up and powered on with the lens attached I could swap out the OLPF's within about a minute or so. But yeah, not recommended, especially in a non-controlled environment and if there is time to do it more properly/carefully.

The one-screw DSMC2 OLPF's were an improvement on the way-too-easy-to-drop-and-lose 4-screw DSMC1 OLPF's, but even so, probably better to just pick one and not have to deal with changing them at all if possible.

I agree with your assessments of the differences between the OLPF's.

Apart from their differences at the extremes of exposure and in the shadows, I found when trying to match them even at a good exposure level, the reds in particular were hard to impossible to match exactly. But even then I don't think it's something anyone would notice outside of A/B examination of sequential images.

With minimal or no post-processing or grading, or control of the shooting environment, the differences could become more of an issue, so as usual, people need to look at what they're trying to achieve and what their priorities are and choose whatever's appropriate for them.

Standard OLPF's for DSMC1 could be harder to find now because most of them were put into the new cameras by default and just left there.

I think it's interesting RED went with non-user-interchangeable OLPF's for DSMC3, although you can still get special-use ones fitted I believe. I guess it could be partly due to the lens mounts not being removable to allow access to the OLPF's, but I wonder if it's also because the image quality advanced beyond needing to eke out the extra little bit the DSMC1&2 OLPF's allowed?

 

[Mike Krumlauf]

Upon further tests, i've decided to put the Low Light back on the camera. I wish I could track down one of the standard OLPFs... seems like thats the sweet spot for the camera.