Letus 35 focus...a dead horse

[John Carrithers]

I may be out of my mind, but I feel like I should be getting better focus with my EX1, Letus Extreme and nikon lenses. Here's what I have done:

Adjusted the ground glass to 7.5mm using a caliper
Adjusted the back focus screws until a Siemens chart was "in focus" at 4ft with the focus ring was adjusted to 4ft on an 85mm nikkor lens

I feel like objects within a relative close distance seem very sharp. Everything beyond 4-6 feet seems to be less sharp. I may be expecting too much from this setup or maybe I have no idea what I'm doing. Either way I would appreciate some input.

I've posted some frames exported from FCP if my Siemens chart test here below.
Lenses used: 28mm, 55mm micro, 85mm and a 105mm

Another focus question- If I zoom my EX1 all the way into the ground glass of the Letus should "in-focus" objects appear sharp. While I was doing the Siemens test today a swung the camera around to shot an oak tree out side. I believe I had the 85mm on. I zoomed all the way into the ground glass and racked the focus ring back and forth and could find a point where the branches looked sharp.


I know this has been covered a lot on various forums and believe me I've read through many of them. I just feel like I"m missing something. Any help would be greatly appreciated.

Thanks,
John

[John Carrithers]

Here are some more images I exported from footage I just shot.
The 28, 55, 85, and 105 were shooting an exterior from my office window and an interior of my co-worker. I feel like these images should be sharper. Any thoughts?
Best,
John

[Steve Witt]

I don't think that this is an issue with your adapter. What aperture setting are you using on the lenses? Wide open aptertures= softer images. Try stopping down your lens a little.

[Ethan Cooper]

Originally Posted by John Carrithers

While I was doing the Siemens test today a swung the camera around to shot an oak tree out side. I believe I had the 85mm on. I zoomed all the way into the ground glass and racked the focus ring back and forth and could find a point where the branches looked sharp.

Wait a sec... are you saying that you're trying to focus using the camera's focus ring? If so, that might be your problem right there.

If this isn't what you're doing, then you may have given another clue to your problem. I believe that if you're fully zoomed into the ground glass, your image will be a bit soft no matter what you try. Someone correct me if I'm wrong please.

[Bob Hart]

John.


Phil Bloom is pretty much the EX1/Letus guru. Hopefully he will give you an analysis on your grabs. To me, it seems you have over-exposed the EX1 itself on the schoolbuilding wides and relay focus is possibly off in the interiors.

In meantime some casual comment from one much less crafty in the trade.

I am not sure how far along the learning curve you are, so if my discourse is dumbed too much please forgive and forget.

With a groundglass adaptor there are now two layers of everything in relation to image control except frame rate and shutter speed which can also adversely interact but more on that later.


The focus and zoom controls on your camera now become the "relay stage". Once set they are best left alone but monitored frequently for bumps or shifts after shutdowns and restarts.

On many cams, the autofocus can be used with equal or better ability to focus sharply on the groundglass than the human and can be used every now and again with confidence to keep the groundglass image sharp if you happen to bump the zoom or focus when it is in manual which is where I prefer to keep it.

Unfortunately the EX1 is not one of those cameras. You have to establish and maintain relay focus manually, an inconvenient added chore which inhibits intuitive camerawork and challenges confidence.

The Siemens chart is handy but I find I become hynotised trying to use it and am underconfident that I have actually found the sharpest point and keep going trying for better.

I find it easier to use a chart which has blocks of resolution lines on it. The camera may not be able to resolve the finest block but the block will return a moire pattern which means focus is at its sharpest.

The finest point of focus on the groundglass image seems to be actually within the thickness of the groundglass texture and not necessarily when the texture itself is at its apparent sharpest. I think trying for this final focus trim is good for about another 50 lines of resolution but that is purely opinion not necessarily real-world for everyone.

The focus control of the lens on front of your Letus becomes what I call "main focus". There is probably a more correct term for this.


Control of light - The iris on the lens on front of the Letus and the camcorder internal ND filter and internal iris control are all used.

The iris control on the lens on front of the Letus is primarily used for creatively selecting a choice of depth-of-field up to an iris closure of about f5.6 when groundglass artifacts are likely to appear.

The signature of the new user is iris wide-open on everything and extremely shallow depths-of-field until the novelty wears off. With a f1.4 lens, wide-open would totally blind film and burn out a video image. The groundglass adaptor is a little more forgiving but there are sensible limits.

A side-effect of too much light on the groundglass of any adaptor, not just the Letus, is that there can be some internal reflection diffusion across the groundglass texture which has an effect of softening the image and lowering apparent contrast when there is overall overexposure and flaring around strong highlights in normallly exposed images.

With motion picture film, this effect is defined as "halation" and dealt with by what is called an anti-halation backing. This is tech speak for an added surface which does not reflect back towards the lens like a mirror but absorbs most of the light which makes it through the film emulsion.

This layer is stripped away during processing so that the recovered image can be later seen.

As we need to see instantly through the groundglass, this solution is not available. Most recent innovations in adaptors now include surface coatings on internal optics and the groundglass itself.

It remains desirable to control the intensity of the light falling on the groundglass.

If a wide-open aperture is chosen for selective focus effect on the front-of-Letus lens, then an external ND filter in a mattebox might be more ideal than using the camcorder's internal light control functions alone.

A piece of ND optical gel in the space between the lens and the front glass of the Letus would also be next best.

This might seem to be a fussy business, however by choosing to use an adaptor, you aspire to move beyond the realm of home movies and happy snaps. The adaptor does not do it for you. It simply offers you more opportunities to manage your image in the manner of a competent DP.

In an ideal world, you would be able to keep the camcorder iris in the ballpark of about f4-f5.6 which seems to be the sweetspot for focus on the groundglass and control all light using the camcorder shutter speed, internal ND and the iris in the lens on front of the Letus.

However, most groundglass adaptors with orbital or vibrating movements do not tolerate shutter speeds beyond about 1/100th sec without the groundglass motion becoming visible.

Spinning disk systems are far more tolerant. As a rule of thumb, shutter speed seems to be limited to 1/50th - 1/60th sec.

This leaves you with the camcorder iris and internal ND for final exposure trim.

It is not a bad world but in low light you will find yourself moving into wide-open iris on both the front-of-Letus lens and the camcorder unless you light the subject artificially. Wide-open means potential image softness.

Some still-camera lenses are better than others but it is also a rule-of-thumb that wide-open means softer, the Noct-Nikkor f1.2 58mm and Nikon f1.4 28mm being two of the few exceptions outside of modern 35mm motion picture film lenses.


I find it fairly hard to judge the light when using the camcorder LCD screen and when setting up, switch back to camcorder fully-auto to get a near-enough reference then go back to manual for my creative preference.

With the Z1, I find it helpful to adjust the LCD brightness in the conditions I am viewing the LCD screen in by switching in the colour bars. The screen brightness should be raised until what is referred to as the pluge bar in the dark in the bottom right corner becomes barely visible, just a hint is all.

I was not able in the time I had with the EX1 to work out how to do this.

There is a sort of rule-of-thumb with video and reversal film of if in doubt, shoot under by about .5 to 1.5 stops. Competent cinematographers and videographers, please do not slap me over the ears too hard if I have recalled this incorrectly, but please post any corrections so that nobody goes down a dead-end on my account.


If you have good wide-aperture still camera lenses on front and you can back-off the camcorder zoom so that the frame of the groundglass is just outside of the edge of the underscan image, you can pick up some more apparent sharpness as the scale size of each "grain" of the groundglass becomes smaller relative to the frame size.

With care, you can even use limited camcorder zoom for creative dynamic effects or subtle reframing when moving the camera viewpoint is not an option due to urgency or inconvenience. It is a nice little bonus to have in the toolset.


My own rule of thumb with the EX1 is to lock the tripod, then frame with the zoom-out until I can just see the edge vignette of the Letus screen in the LCD, usually on the left, find an object in the image to use as a reference, about 6mm inwards from left edge of the LCD screen.

I then zoom-in until this object moves left and just clears the edge. This may not be valid for other individual EX1 camcorders and adaptors as variation in centering of the Letus screen when mounting up the adaptor to camera can occur.

[Ethan Cooper]

Bob - hell of a post.

John - I took a look at the larger versions of your pictures just now and they look a lot like what I get out of a cheap Nikon DX zoom lens when I have it mounted. I know that observation probably wont help much, it's just something I noticed.

Good luck.

[John Carrithers]

Thanks for the response everyone. I really appreciate it.
Steve, the aperture on the EX1 was open on both exteriors and interiors. The nikon lens were WFO on the interiors and at F8 on the exterior.

Ethan, I had the EX1 focused on the GG once I put the Letus on. Then I adjusted focus on the Nikons.

Bob, thank you so much for the informative response. There's a lot of valuable information in there. I really appreciate you taking to time to explain all that.

Thanks again everyone. I'll try to do some more test in brighter environments to see if the images appear any sharper. This is still new and I guess I need to experiment more.

[Ethan Cooper]

John - Let us (pun clearly intended) know what fixes your problem.

[Steve Witt]

Originally Posted by John Carrithers

Thanks for the response everyone. I really appreciate it.
Steve, the aperture on the EX1 was open on both exteriors and interiors. The nikon lens were WFO on the interiors and at F8 on the exterior.

John, If I understand you correctly, you are saying that you keep your camcorder's lens aperture wide open at all times. This would probably tend to soften the images as well. I don't know much about the EX1 but try stopping to around f2.8 on that camera and see what it does for you.

[Phil Bloom]

when you say you are zooming all the way into the ground glass, what is the z setting on the lcd?

[Leonard Levy]

You're right those pictures look very soft. Something is wrong.

You really need to give us more information about how you are setting focus though.

Bob's post is excellent and I fully agree that the sieman's chart is more or less a waste of time. I find if you get a lens resolution chart (and you can download one on the web just by googling it) and then go to kinko's and blow it up to 4' across then mount it on a piece of foamcore, you will soon learn an enormous amount about your lenses and your adapter and how to keep things sharp.

You can download that chart also at:

http://www.leonardlevy.net/Adapter_s...ster_Final.htm

The coolest thing about the EX-1 for me with this camera is I check the focus of my EX on the Letus screen by just using the red peaking. With your Ex-1 lens wide open just focus up any image with any 35mm lens that has a lot of detail in it ( trees or foliage work great). Then go back and forth between the 35mm lens and the Ex-1 lens until you get those lovely little red marks all over the place. (I use red anyway).

3 other thoughts:

1. Philip asks an important question, if you zoom in too much you may be making your image too soft. I usually end up around Z76 or a bit more.

2. Do you have a Letus with an EX-1 corrected achromat?

3. Are you having problems only when you focus on stuff far away? If so your back focus from the 35mm lens to the Letus screen may be off. I don trust the calipers method or the focus at 4' method myself because every single adapter I've ever bought had the backfocus off.
Look at your infinity focus for objects at least 100 yards away and preferably much longer. You should just hit infinity focus at or near the end of the lens throw. If you don't hit it at all , or if you never get to it, your back focus is off and you need to adjust the adapter or if you have their new Nikon mount you can adjust that.

Lenny Levy

[Bob Hart]

Leonard.


If you remove the < and > from either side of your link, this site will automatically convert the text into a clickable link unless you have intended it not to be, to limit traffic, in which case please ignore my comment.

You can blame me for the 4' recommendation. I found it more convenient for setting up in a confined workspace with an A4 sized chart, which can be framed at about 4' with a 50mm lens at about 4' or 1M distance. The average hacker like me often times only has a few of the darkness hours in which to indulge one's alternative interests.

Your infinity method is equally valid and I use this to maintain backfocus on an AGUS35 with three axis adjustment with the widest lens I have on front.


Phil.

Furthur back in a post, I may have sent someone down a dead-end when I referred to the number 42 for safe zoom-in for the EX1/Extreme without going in too far (estimated from the nearby 40 witness mark on the lens barrel). I no longer have the EX1 to check again. If you have time, could you check that and publish a refutation if I was wrong.

[Leonard Levy]

Bob,
Thanks I switched my link as suggested.

Truth is I haven't checked all my lenses to see if they really do line up exactly with the footage marks on the lens barrels ( If I have time I'll double check that with a few lenses - maybe I'll eat my words!)

I'm going from my own experience in a film rental house collimating lenses for years. We used collimators to check infinity. Then the assistants would check to see if the lens barrel marks were accurate and if they weren't they make notes for themselves.

I figure if you couldn't be sure cine lens barrel marks were exact why should we put so much faith in 35mm still lens barrels which aren't designed for anyone to care if the markings are precise.

I've learned that many of the manufacturers have been using your lens barrel method, but I've been shocked to find every adapter I bought ( 6) to be misadjusted, sometimes by enormous margins. Even when I thought I had them corrected them , one day I visited my old shop & checked the collimation. They were still all off. I learned there that my Nikon 80-200 tracked correctly if the back focus was right, so I now use that as a standard.

Since then I've talked to a number of the manufacturers and I think convinced some to invest in a collimator and make back focus adjustments easier.


Lenny Levy

[Bob Hart]

The autocollimator or projection is king for setting backfocus. It does require a bit of lens theory and handy skills to do as a backyard or kitchen table job.

I have used the projection method to check the witness marks on a lens by fitting up to a Mini35.

You take the lens out of an old slides projector and shine the raw beam in through the back of the Mini35 and through the lens you are testing onto a screen, the exact measured distance from the film plane. The groundglass texture is your reference. The projection is very dim and you have to take care not to cook the Mini35 by leaving the light on for too long.

The Mini35 is set up in the factory by projection, a calibrated lens and selective fitting of fine shims, so you can be assured it is correct unless somebody ran over it. I don't know if any of the alternative manufacturers are doing this.

I have actually used a PD150 to check the backfocus of the lens versus viewfinder screen on a CP16 film camera. Not exactly a precise and scientific method either but doable.

http://www.youtube.com/watch?v=pulx6otzs4c

There is variation between lenses for sure, an ugly business I have seen but I found the older metal Nikons to be fairly close across families.

In the ballpark of 4ft to about 10ft with the 50mm and 85mm lenses is the more critical zone in what we do. If you get the witness marks accurate over that range relative to the groundglass image, I feel that is the more practical course to take, although it is less than truthful if the lens witness marks are off and you put another accurately adjusted lens on.

The ultrawides are going to tell the truths here because they are far more critical for backfocus and are best outside of a proper calibration lens, provided the witness marks are honest, for setting up the mount backfocus using an infinity subject as you recommend.

[Leonard Levy]

Bob,
You are leaving me in the dust. That is one helluva way to check your viewfinder focus on a CP, at least i think that's what you were doing.

Can you explain the slide projector method in a little more detail. I have an old one laying around just waiting for something this nuts.

I'm just using my Nikon 80-200 now to check collimation because I know it tracks and it will only do so at the precisely collimated backfocus.

Lenny