anamorphic lenses and shake -- problematic?

Hi there,
I'm curious if anamorphic lenses are more prone to shake than sperical lenses due to the stretch applied after recording.

- does the stretch amplify shake?
- is rotational shake a problem?
- is it advisable to use a gimbal with anamorphic lenses?
- does it differ related to the stretch factor (1.33;1,6...)?
- do anamorphic lenses amplify rolling shutter (due to stretch)?

Lots of questions...

Greetings,
Stephan

So, I did some research myself and here are my findings so far:

- rolling shutter gets worse, depending on the squeeze factor as predicted,
- in-camera stabilization can make shake worse, turning it into wobble,
- in-camera stabilization can let flares wobble,

Overall, shake and wobble seems to be a bigger problem compared to aspherical lenses.

Any comments? Experiences?

Anamorphic lenses are unique in the way the present the sensor readout.
If there is any IBIS (in-body image stabilization), it needs to be turned off.
If the camera suffers from poor rolling shutter, then mitigate micro jitters but either adding weight to the camera rig or stabilizing it on supports (i.e. tripod, gimbal, dolly, etc).

With all of the options out there for budget anamorphics, they have tended to be smaller and lighter weight, thus you need to compensate for your rig design. Stuffing a pillow in a backpack and slinging it around your shoulder to lay the camera on can help too.

with a M43 cinema camera without stabilisation, the Pocket 4k, I didn't see any difference between spherical and anamorphic lenses. Probably there is one, but not very noticeable.

The Panasonic G9m2, a camera body with stabilisation, has a special setting for anamorphic stabilisation. Since there are many affordable anamorphic lenses these days, I expect that many modern cameras have such a feature.

with a M43 cinema camera without stabilisation, the Pocket 4k, I didn't see any difference between spherical and anamorphic lenses. Probably there is one, but not very noticeable…

I think whether or not stabilization is affected by anamorphic lenses depends upon the method used in post to desqueeze the image. Let me try to illustrate with a 2x squeeze lens.

Assume you’ve recorded 2x squeeze using 3840x2160 photosites. There are two ways to get that image to look right in post. You can desqueeze the X-axis horizontal dimension to 7680x2160. Or you can squeeze the Y-axis 2x to 3840x1080. The former method doubles whatever instability you had on the X-axis; the latter method halves the instability of the Y-axis.

Well, that’s my arm chair coaching, not a scientific proof, but I always squeeze Y in post to restore the proper pixel aspect to ‘square.’

thank you Rick, squeezing down the y axis makes of course sense. In my case I simply don't know how it is done. DaVinci Resolve recognises anamorphic clips automatically and desqueezes them without my doing. I trust that Resolve does it right. I've only shot with 1.33 squeeze factor. Two times squeeze factor certainly makes a lot more difference with shake and stabilisation.

This is the case in the Panasonic Lumix anamorphic stabilisation thing. One should of course select the correct squeeze factor.

… squeezing down the y axis makes of course sense. In my case I simply don't know how it is done...

On the Media tab of DaVinci Resolve, I leave the pixels as square, ignoring they are anamorphic captures.

On the Edit tab, in the Inspector, the Transform function defaults to linking the X and Y axis. Unlink X and Y.

In the Y axis, use the value 0.5x to squeeze the anamorphic pixels if your anamorphic lens has applied a 2x squeeze when capturing the image. To adjust a 1.33x squeeze, use the value 0.7518797x which is the approximate value of 1 / 1.33.

Resolve will only display the squeeze with three decimal places, but you can enter a more precise value.

If your anamorphic lens’ squeeze factor varies with the distance from the subject, you can enter other values to correct the pixel squeeze. My SLR Magic Anamorphot adapter is 1.33x for Normal distances, but Near distances (less than 5’) have a smaller squeeze, such as 1.22x if your camera is quite close which could use a Y axis squeeze of 0.8197x. At one time I setup tests of known circular objects such as a cooking pan at different Near distances and then applied a squeeze in post to recreate a circular object. You may not need to do that if your lenses is a true anamorphic and you can use the value 0.7518797x for every subject distance. I think my squeeze is variable because I use an anamorphic adapter on a spherical taking lens.

the Sirui anamorphic lenses I use, the so-called "Mars"-set, have a varying squeeze factor depending on the focus. Their design is the cheaper variant of anamorphic glass. The lenses who keep the squeeze factor are more complicated to build, and therefore more expensive.

Thank you for the settings in Resolve. I might try them to see if there is a difference.

the Sirui anamorphic lenses I use, the so-called "Mars"-set, have a varying squeeze factor depending on the focus...

Does Sirui provide guidance on how their squeeze factor varies with distance to the subject?

the Sirui anamorphic lenses I use, the so-called "Mars"-set

I'm closely watching this discussion, since I also like to go that rabbit hole...
Phil, are you happy with the Sirui Mars MFT on BMPCC4k? is it worth the effort?

IMHO, the bmpcc4k camera gyro should also help in stabelizing....

My experience with post stabilisation (gyro or otherwise) with lenses with a lot of 'character' (i.e. vintage lenses and anamorphics) is that post stabilisation becomes incredibly obvious as the vignette, corner abberations, flares etcetera start dancing around the frame.

I don't think, that gyro stabilization and in camera stabilization differ at all regarding warping with anamorphics. It's the stretch factor that emphasizes warp, wobble and rolling shutter.

@Phil999: the varying squeeze of the Sirui Mars, how big is the difference? Do you have to adjust each clip manually, or only on the extreme ends of the focus reach?

I don't adjust clips with different focus. It's 1.33x for me for each clip. And if there is such a case, for example a scene with a circular object, which appears as ellipse, I would simply zoom in a bit and adjust the y axis. So far such a case hasn't occurred yet. I mostly do nature short films, and I rarely do rack focus where the squeeze factor change and breathing can be seen.

The Sirui Mars set is very good in my opinion. I love its colourful flares, next to the horizontal blue flare. You can see a lot of this lens set on my channel youtube.com/@Eleni999 (not in 2024 videos, where I didn't have the Mars set with me). I use this set with the Pocket 4k. Camera and lenses match very well. Anamorphic clips get recognised in Resolve, the squeeze factor is automatically applied.

Regarding the difference of squeeze factor at different focus setting of the Mars set, I recommend this channel: https://www.youtube.com/@AnamorphicOnABudget
He also explains the inner working of cheap anamorphics compared to expensive ones that keep the squeeze factor. I watched dozens of his videos before I decided to buy the Mars set.

Regarding problems with anamorphic lenses and post stabilisation: I often use the 'Translation' and the gyro variant in Resolve. This keeps the flares mostly intact. Most of my investments in gear in the last two years was mainly focused on mechanical and motorised gimbals, and combinations of them (Arri Trinity alike). You really want to have a stable and fluid camera movement when working with anamorphic lenses. And use tripods. As Sean van Berlo said, post-stabilisation can get ugly. However, I'm only starting with the gimbal combinations. 2025 will be very interesting, once the snow has gone.