View Full Version : Excessive graininess on Letus Extreme


Shaun R Walker
September 16th, 2010, 06:29 AM
I've done a few shoots on my EX-1 and my Letus Extreme recently that had excessive graininess, especially using my 70-210mm lens. Aperture was set fully open on my EX-1 and about 2.8 on my lens. Any suggestions as to what could be causing this? I have attached a still from one of my shoots that shows the grain.

Bob Hart
September 17th, 2010, 06:25 AM
Shaun.


As I am attempting to assess based on a single frame grab my following comments are indeed just guesswork and should be regarded accordingly.


Except for using the zoom lens for dynamic effects, my personal preference would be to stick to prime lenses.

Before I can try to assess your sample image could you tell me a few more things.

What was your shutter speed ? ( I shoot "shutter off" when using the adaptor ).

What was your video gain setting on the camera ? ( From your settings and the image itself, it appears you were in a lighting environment that was becoming difficult. The image looks like it may have been adjusted with increased contrast as the sky is burned out and appears also that you shot against the dominant light source. Shooting against the light is not the strong suite of any groundglass adaptor. Shooting against high light bright overcast really brings out the worst in them).

Did you use a reflector to lift lighting on the face ? ( Your shot looks like you may have had to adjust levels in post to bring up the face. In that circumstance, I have observed "grain" to become accentuated. )

The following comment gets me in a bit of trouble from time-to-time, so again hold my comment to some level of question.

"Grain" in a groundglass sourced image ( for want of a more appropriate term ) may consist of two components, visible texture of the groundglass "seen" by the camera and sensor noise generated by the camera itself. Added video gain makes sensor noise more apparent.

The contention I get into is that I theorise that despite the movement of a groundglass which will de-resolve the texture of the groundglass, there are some points in the texture of the groundglass where what I call in nightvision speak, "scintillations" or momentary brighter pinpoints may occur randomly in the successive frames of image. This is one source of the filmic effect attributed to groundglass adaptors.

Separately, with vibrating or orbital groundglass movement systems like the Letus, a shimmering fixed pattern artifact is likely to become apparent when the aperture setting of the lens is at or tighter ( higher number ) than f5.6.

With some zooms, whilst the aperture wide-open may be f2.8, if this lens is not a constant aperture lens, at the telephoto ( narrow ) end, you may find that the effective aperture of the lens is a higher number. You might find that at the telephoto end, your effective aperture might be f4 which brings you into the zone where an artifact might become apparent.

I'll give you an example - the Sigma 50mm - 500mm zoom is f4 - f6.3. On the lens barrel this detail is described thus "50-500mm 1:4 - 6.3 APO DG HSM D". ON the Letus and P+S Technik Mini35-400 this lens is fine up to about 300mm then when zoomed furthur begins to provoke the "grain" effect.

I dont think this is your problem because the artifact in the Letus tends to have a slanted "grain of rice" look to it. In the P+S Technik Mini35-400 it is more circular with facets. I cannot rule it out for sure as I am looking at a single frame, not motion vision.

The full fixed pattern artifact issue appears to have a narrow transition phase from invisible to fully ugly. The circumstances you were shooting in might have just nudged the threshold.

Another factor which I also get into trouble with in discussion is that I suggest that I observe a small interaction between the camcorder iris setting and adaptor lens iris setting when the f5.6 threshold on the front lens is approached.

I could not find on the EX1 camera lens barrel anything which indicates that the camera's Fujinon zoom is not a constant aperture lens.

If it is not a constant aperture zoom lens, the zoom position you would be using, my guess from your posted image is about Z69 or more, (getting inside the chips in your prism if you are the owner who had one with a damaged prism). Therefore interactively, it is possible for you to be nudging the threshold of provoking the artifact if both lenses are at the tele end of their focal range and in effect, both are at the higher iris numbers.

I am not an industry professional so please do not accept my apparent observations as being any sort of technical criticism. I am not qualified to make such.

This clip was shot with the Letus Extreme on an EX3 via the EX3's own Fujinon lens. You will observe some visible "grain" or noise in some of the shots.

Junkyard Scene 1st Cut on Vimeo

I encouraged the cameraman to use all the lenses in the zone f3.5 - f4. In the hostile lighting conditions of high bright overcast and in some instances dominant light from behind subject, on reflection this might have been a bit unwise of me. Why I did this was because one of the primes they used was an f2.8 35mm so I wanted to keep it closer to its sweet spot and for the look of the other faster lenses to then match it. I also wanted to discourage them from shooting everything with a paper-thin shallow depth-of-field.

Contrast and saturation were brought up in post compared to what I recall the untouched images to have appeared like.


FOOTNOTE:

Another little small detail. If your editing system defaults frame blend "on" any "grain" or noise might become more apparent as a half-intensity occurrance. However it might also look worse unblended at its original intensity plus any effects you have applied.

Shaun R Walker
September 18th, 2010, 01:36 AM
Wow Bob, you're the man! Lots to consider. From memory the camera was set to 0dB gain, but I have a feeling the problem is perhaps the variable aperture of the lens. Unfortunately the lens is packed away at the back of my equipment cupboard so I can't get it out to check, but I would hazard a guess and say that it is likely the culprit. I am brand new to the Letus and obviously have heaps to learn. Unfortunately the only opportunities I have had to use it have been on actual shoots, which on reflection was probably a little unwise. I must take some time and do some more comprehensive tests based upon your recommendations before my next job.

I appreciate the time you have taken trying to get to the bottom of my problems Bob. You've helped me enormously.

Bob Hart
September 18th, 2010, 08:07 AM
("Wow Bob, you're the man") - I wish.


I could be sending you off on a goosechase.

It is usually a combination of subtle events that bring a groundglass relay image undone, a bit like accidents with airplanes actually. The more obvious single ones usually stick their head up and are corrected.

There's nothing wrong with going in off the high board and taking it on actual gigs, provided the project is an affordable loss or can absorb compromise. Life without risk is pretty dull.

A certain amount of operating in real-world situations is what will bring problems up and cause you to devise the solutions.

What is important in using the camcorder/Letus combination is to operate it enough times that you become intuitive and reflexive with it, learn the boundary limits and the benefits.

To return to the aviation analagy, that industry has a strong culture of maintaining recency in skill sets.

Camera operators should wisely do the same by digging machinery out and just operating if no gig comes along for longer than four weeks or so.

If you get a chance, try to find the DVD of the indie feature "Dear Wendy". It was DPed by Anthony Dod Mantle ("Slumdog Millionaire"). The P+S Technik Pro35 and a custom Sony HDCAM rig tethered to a recording system was their camera. The behind the scenes "extras" on the DVD are helpful.

The website

www.exilemovie.com

is also worth a look and there is a generous offering of behind-the-scenes material. They used a mix of groundglass relay ( P+S Technik PRO35) and direct-to-camera footage.