Homemade 35mm Adapter

[Agus Casse]

Thanks Bob, you are my hero to keep this thread alive... actually the proyect is not really dead, the problem is that i cant find the right components to make a new and improved version. Works have been a bitch, but i think i can find the time to get back to the proyect in a couple of months.

Jim Lafferty, actually there is a old thread, which a guy put a razor and make the adaptor vibrating.., and actually my first version was using vibrations. also remember that my adaptor uses diopers, and achromats.


Quyen Le, i have been reading a lot of post about new adaptors, and i have received a lot of emails if i will make a commercial version, thats is why i am talking with a industrial designer, to find a way to make this thing cheap and with milimetric presicion.



see you guys soon...

[Jim Lafferty]

Ah... I thought you'd just used the credit card magnifier or the glass from a reading magnifier and not an actual achromat. I guess in trying to condense the history, I'm looking at the peaks and not all the smaller developments on the way. I'm pretty sure it was Brett who came to the table with much of the achromat research we all benefited by, and that's why I put his name there.

[Bob Hart]

Resolution Tests with an AGUS35 style device on a Sony HDRFX1 PAL.

The image yield occurred with the following path.

Test Chart >> SLR lenses at f1.8 >> AO5 groundglass (backpolished slightly) >> two prisms >> Century Optics 7+ achromatic dioptre >> HDRFX1 camcorder.

Conditions were adequate artificial lighting of the test card. Camera fixed on a sturdy tripod. The frame taken from the groundglass is 16:9 and about 26mm wide.

Today, I looked up the resolution scale printed on the back of a Lemac test card. I shot some lens tests with this card about two months ago. I had misinterpreted what these resolution indicators mean and wrongly assumed they all were invalid for assessing video.

The resolution of the FX1 LCD flipout screen may not be that of the camera system itself. In the screen, all but the "A" and "F" indicator patterns were visible. Whether the camera is actually seeing the individual lines in the finest visible "B" and "G" patterns or there is a moire interaction happening I don't know. But it seems from the LCD screen that the pattern is being detected.

For a moire interaction to occur it would seem that the CCDs of the camera should actually be seeing the fine patterns of individual lines. However it might not be necessary for the lines on the card to be seen at correct resolution for a moire pattern to be generated. It might happily occur if every second or third line was seen by the camera CCD for moire to occur. I simply don't know enough about the subject to make assured comment. I have not yet been able to get access to a high definition monitor to look closer at the recorded image.

That aside, if the camera CCDs are seeing every line in the "B" and "G" resolution indicator patterns, then according to the card, horizontal resolution is up to 864 lines and vertical resolution is up to 486 lines.

This hints at a better result than my guess of 700 lines from the EIA1956 test pattern from earlier tests into a PD150. The PD150 is apparently limited to 530 lines and that is where the separation of the four tapered lines ended. But there remained hints of the tapered line pattern outward into the 700 zone to about 710.

According to the card, the higher resolution "A" and "F" indicator patterns will not resolve on HDTV anyway. These are 1920 lines and 1080 lines respectively.

If my fairly rough and ready appliance with its home-made groundglass is truthfully yielding these results, then better results can be anticipated from the more thoroughly designed appliances with better groundglasses.

[Bob Hart]

A hint for anyone building an AGUS out of plumbers pipe and pipe caps.

The f1.8 lenses give you a nice wide area of light across the groundglass, larger than the 25mm or so the 56mm hypotenuse prisms will allow. When setting up the centres on the camcorder mount and the front lens mount to make the holes in the pipe caps, a little laziness may creep in and an adjustment to the front cap (ie., a remake of a new cap = lots of work and a wasted cap) one tends to leave the front cap as-is if the most vignette prone lens in the arsenal works okay.

You set up the backfocus across the entire image, by aligning the lens mount true to the rear camera mount, correct the backfocus again then fine tune it with the three adjustment screws while sighting on a infinity target, fuss around until all the interactive effects have cancelled out and are then well pleased with this beatific smile on your face.

Then you go and do a real-world test and find there's a soft edge or a corner. It's consistent across the entire range of lenses but more aggravated with the sides and turns up when the lens is focussed on something midway in the range. You check the device. All the glue locks are intact and nothing has moved. You didn't leave it in the sun to cook and nothing has bent or gone soft.

Then the dim realisation begins to pall. You stop the lens right down to get the darkest vignette and there it is. The lens is off-centre relative to the camera view, something the wide lenses will not accommodate, especially on closer objects.

On reflection, I would have saved myself a lot of time and effort and sacrificed pipe caps if I had made a facility for lateral adjustment of the lens mount position along the radius of the pipe cap, a simple slide and clamp arrangement and used the rotation of the cap itself on the image tube to provide the second direction of adjustment.

There is no room for unthoroughness and expediency in this business.


Another caution. If using small self-tapping screws to fasten into the PVC, take care not to screw them in fully home in one hasty journey to firmness. They generate enough heat to melt the plastic but also to cause a heat related stress rise where the head of the screw joins to the shank.

The plastic cools and sets on the screw threads. Later, you decide to dismantle, there is a gentle SNAP!! and the head comes off the screw. It has happened twice.

Grooved self-threading screws such as found in some players and videos, might be a better choice.

TEKs are no good as the drillbit on the end is too wide for plastic and will cut too much away for a secure fit.

Don't use the knurled screws which look like the threads are a sort of rasp or file to cut threads into the plastic. These will begin to strip the hole if fitted more than once as they are not intended to be unfastened.


Quyen.

If you are reading, I had a random thought about high definition resolution on your device. It might be worthwhile getting a few microscope slides from your nearest university medical faculty, high school or from a pathology service, a piece of aluminium cooking foil and some of Oscar Spiers wax formula.

I tried it wax between two disks but abandoned the idea because I was not able to control the thickness of the wax layer consistently around the whole disk. The image quality was by far the very best but the flicker was totally unacceptable = 1.5 f-stops.

Because your groundglass has a much much smaller movement than a spinning disk, consistency of the thickness across the groundglass should not be the same problem.

The problem would be locking the two glass panels together so that the wax and glass do not separate due to the vibration. The vibration and travel of the groundglass may be adversely effected by the extra weight and the wax composite panel of course will be vulnerable to effects of heat like any other wax gg. The lesser grain of the wax however may enable a smaller travel of the gg and make for a quieter device. - Just a thought.

[Jamie Roberts]

Hi all

my project to make an agus35 is officially finished. Its as the original one is, I gave a 10x macro lens between the camera and the gg, I have mounted a 40mm plano convex lens between the gg and the 35mm lens (closer to the gg), gg is sanded plastic CD mounted on a cd walkman motor, and i am happily able to shoot in widescreen on my gs400 without any hotspots, lightloss around edges etc. The money thats gone into it is about $60 australian and the results in my opinion look great in terms of acheiving shallow dof.

these threads have been invaluable in helping a clumsy novice like me make such an effective bit of kit! thanks alot

jamie

[Carl Jakobsson]

Jamie, I would like to see some footage. I also believe in sanded cds... :)

[Dennis Wood]

Jamie, I'd love to see some PAL footage from your GS400 with the adapter. This is from my GS400 with my spinner (now sold):

http://www.filefactory.com/get/v3/f2...4268f3d5e956b9

The GG is 1.5mm optical acrylic, fine media blasted on a rotating jig (at 1000rpm).

[Bob Hart]

Night Tests. AGUS35 to HDRFX1

For sake of some real-world testing, I went down to the local retail dragstrip and on into the city and shot some unassisted footage in the night. I stopped by the caryards in Vic Park on the way home. They are well lit and a cheap way of seeing how the system performs under good lighting. As usual, the roving security people give you a strange look and passing fitness walkers do likewise when they see the weird gadget on front.

The light loss of the system is most apparent on the cityscape.

The direct-to-camera images are bright enough for the overhead scud cloud to be just visible.

The light loss through the system is very noticeable, especially through one lens, the 12-24mm zoom which is f4. However, and this is weird, there was not the apparent two-stop increase in the brightness from the f1.8 lenses I expected to see. I can't work that one out at all.

From the relayed images, all you get is the pinpoints and dim window rows on the towers. Direct-to-camera picks up spill from streetlighting onto the sides of buildings. So for night footage of cityscapes, the relay system might better be abandoned in favour of direct-to-camera.

Another weird aspect is that I forgot to turn the disk on for the first two lenses and there is no groundglass texture to be seen.

The longer lenses perform better, mainly because any bright light sources such as signage occupy a larger portion of the image and are recognisable.

Under the bright lighting of the caryards all the lenses performed well. Except for the increase in brightness and better contrast, there was little discernable increase in apparent image sharpness between the relayed and direct-to-camera images.

Again, there was very little discernable difference between the f4 lens and the f1.8 lenses in this lighting environment. On a high definition display it likely will be a different story but I have no acccess to one.

I also shot one test with a single lens in a streetscape after light rain. The area was moderately lit. With the gain fixed to keep the blacks less noisy a pleasing and realistic image can be had.

I have tried to put motion files up on www.putfile.com but have not been able to download them to view them so have no idea if they are downloadable or not. Can somebody give me some advice on the best compression system for placing motion files there?


Jamie.

Congratulations on completing your project. Now may likely begin, the obsession of trying to make it that little bit better.

If you have a chance sometime, could you post a frame grab somehere. I'd be interested in seeing how the condenser performs in your arrangement.

For field of view without a condenser, my arrangement frames the 16:9 Lemac chart in a 25mm wide frame on the groundglass with a Nikon f1.8 SLR lens set at 1.0 metres from chart to groundglass.

[Dennis Wood]

Bob, try http://www.savefile.com/ for uploading files.

For sample footage at full frame size, windows media 9 (MPEG4) is hard to beat. Encoding takes a long time, but at 3000Kbps, the quality is better than max bitrate MPEG2, IMO.

http://www.freefunfiles.com/software...encoder-9.html

For NTSC HQ anamorphic footage 16:9, I resize to 872x480 square pixels, and encode at 3000Kbps. That's about 17MB/minute. All of my uploads for adapter footage tests are done that way.

For NTSC lower quality anamorphic footage 16:9, I resize to 436x480 square pixels, and encode at 500Kbps.

Hope that helps. If you want a few .wme files email me. They can be loaded into the encoder and will have all the settings done for you. You just will need to change source and destination file locations.

Hope that helps. If you're on a Mac, H264 is the MPEG4 to use...and I can't help you :-)

[Jamie Roberts]

i am going to do a little staged shoot tonite here at home using artificial lighting.

will put something up as soon as i have something.

cheers

jamie

[Bob Hart]

First look on a high-definition plasma screen.

Today I took a FX1camera and a component patch lead across to our neighbouring Harvey-Norman and they graciously permitted me to playback tests in the high-definition realm.

I have to confess that despite my scepticism, the playback looked a lot better than I had expected.

The pitfalls of the home-made relay device soon became apparent.


Sharpness of the image.

The AO5 finished groundglass texture can deliver an image of adequate sharpness. Maintaining the sharpness of the image delivered to the groundglass is another matter.

Focus failed on two levels, the relay and prime lenses, both due to operator skill or lack of it. In a production environment, a high-definition monitor would be essential. Being short-sighted, I occasionally overlook soft-focus given that it is my natural normal vision state.

I also found that making lens changes often meant the camera focus ring was bumped. This needs to immobilised after relay focus is set as another participant at this site has already observed.

Light levels onto the groundglass also seem to have an effect. The camera's ND filters amd manual apertures have to be used more and with more precision. Some interactions occur between the prime lens aperture adjustments and the camera aperture adjustments which can add to or subtract from sharpness.

An image containing a strong large area backlight and a large area underlit foreground was almost impossible to resolve without visual artifacts from the groundglass disk appearing.

This was most apparent with a f4 12mm-24mm zoom which only leaves one f-stop of adjustment before the f5.6 rule is broken. If lens aperture was adjusted to reduce the overexposed area, artifacts appeared in this area. If the camera aperture was opened up, the artifacts moved to the underexposed area. The f1.8 lenses fared better but it is a fine tapdance trying to find the best combination of both aperture selections and the ND filter.

The scene was views along a roadway lined with thick forest under light 7/8ths overcast with patches of clear sky. A graduated ND filter might be the only way to resolve this image if it possible at all. It is certainly a scene I would avoid.

"Magic-hour" shots looked great, especially with light from behind camera, in the direct light or in open shade. The sharpest images were achieved in this circumstance. A one-stop manual under-exposure seemed to yield the best sharpness and colour rendition.

The colours of backlit foliage in the "bokeh" areas was an excellent effect, likewise at night, the various colours of lights when focus was moved from very close to on-subject with the 85mm and 105mm lenses.

Another effect on sharpness was observed and has been previously commented on. There were occasional slight momentary focus shifts due to end float on the disk motor allowing the groundglass to move off focus. There is only about 0.25mm in it but it is enough in the HD realm to spoil the image. Up until today, I thought I had eliminated the problem.

A possible solution might be to fix a large metal washer very close to the motor side of the groundglass with a large enough centre to surround the hub and using the magnetic piece found in some CD players which clamps the disk. The magnetic field may be enough to keep the motor armature back against its rearmost limit without adding the friction a home-made thrust bearing would.

The night footage seemed softer and flare in the relay images was more apparent as one would expect from the many extra pieces of glass in the path. Except for inferior contrast, there seemed to be less difference between the direct-to-camera images and the relayed images than during natural daylight.

There was one incident of chroma separation on the right side of the image as viewed. I attribute this to the prism path as it is not uniformly spread around the entire outer edge of the image. It showed where a bright tripod leg intruded into a darker area of the image which was also slightly out of focus.

The earlier tests which made apparent the misalignment of the groundglass and lens centres were woeful to see in the HD realm on a large screen.

I found I had some eye tiredness after 50 minutes of viewing. I experienced this eye tiredness to a lesser extent after watching "Star Wars Episode 2" and on a par with that I experienced after watching "Open Water" , both movies in a theatre.

Overall, I think one could make a low budget feature with this combination but it would require a stressful degree of close vigilence and avoidance of some lighting situations, notably, large untextured areas of overlit background and underlit foreground of more than two or three f-stops of difference.

There's a way to go yet with this project.

For now, the FX1 goes back to its own home with my fingerprints on it and it's back to the PD150 for a while.



Dennis.

Thanks for your hints. I'll give that a shot.

[Jamie Roberts]

Hi all

heres a link to a short clip that i did up today that shows my agus35 adapter in action!

http://www.filefactory.com/get/f.php...3a59388db778dd

Inside stuff really didnt have enough light but i kind of liked the way it looks anyway. Outside clip at end is to show how it looks in good light.

Im not after perfection. I will definately use it for close up shots where i think it really comes in handy. I would like to now continue on im my efforts to actually make a decent little movie!

I can see how easy it would be for it to become a hobby in itself just working on improving a 35mm adapter but thats not me, unless some new simple idea comes along that a child could make (so I can!), i'll be content with this for now.

Jamie

[Bob Hart]

Jamie.

You've done well it seems, going by the clip playback.

There's no more needed to be done for your purposes, beyond perhaps some fine adjustments, unless you want the convenience of a flipped image.

I observed no variable density flicker in any of the images so it seems you have the disk well sorted.

No evident endfloat of the spindle or misalignment of the disk on its hub is evident.

Interiors are as bright as I would expect them to be.

Corner to corner, the light intensity seems to be even, I guess the benefits of a condenser in the path at play here.

In the exteriors there is a hint that your disk might not be square-on to the focal plane or the camcorder centre axis might be slightly off, but it is usually the disk alignment itself which is usually more affecting to the image.

As I see it, the left edge of your disk may be about 1.5 mm rearwards and the top edge of the disk about 0.75mm rearwards, relative to the centre. This is really only wild guess as I am going by the soft areas in the backdrop of the hedge and the nearby roof eaves which seem off in the right and upper portions of the image as viewed.

This is however very much an assumption. It may be that the hedge is not square-on behind and the roof eaves in a genuinely out of focus area so don't take too much notice of this comment.

The interiors seem square-on as far as the disk alignment goes so you might be getting a bit of flex with the CD-R case if that is still what you are using.

A little circular piece of plywood fastened to the inside front of the CD-R case with screws and silicone sealastic to stiffen the front structure might be the way to go if flex is happening.

This is the beauty of Agus's original design, a result tantalisingly close to the P+S Technik benchmark for a miniscule fraction of the cost and well within reach of the budget and construction skills level of the enthusiast.

The one and only motion video file at www.dvinfo.net/media/hart was shot on a non-flip version ith a plastic disk. Feedback suggests preference for the image from this plastic disk compared to the image from the glass disks.

[Jamie Roberts]

thanks for that Bob.

You are probaly on the money in terms of the disc being slightly out. I also actually used an ultra thin sanded plastic disc that i found in a spindle of dvd's. maybe its too thin and wobbles a little. it would be half the thickness of the std clear plastic cd's u get.I was hoping the thinner one would allow more light.

thanks also for the good idea about reinforcing the front of the housing.

Jamie

[Bob Hart]

Jamie.

Your disk is running true so I would not mess with it. The DVD clear spacers are good. They are optically true and run true if they are the ones you can see a slight trace of the guide tracks on (looks a bit like opal against the light). This is the surface you sand or grind away at.

You may well have hit upon a good combination. The DVD spacers or split DVD disks seem to be made of a harder composition than the older CD-R spacers. The DVDs don't seem to dress as easily using the abrasives I use for the glass ones but when you get a successful one, they seem to be as good as the glass. They also run true because they are thin and can bend and the centrifugal force takes care of any runout in the hub.

The more aggressive sandpaper with fixed grit might be a good solution for homebuilders. Once the abrasion has started, then a dressing in a fluid abrasive might be viable for a finer finish.

I made one and the image was more filmlike than the glass but it got a circular scratch on it from a screw-end sticking through the motor mount. When I took it off the hub to dress it again, I furthur damaged it and I have not been able to successfully make one since to replace it and am back to glass which I know.


With your arrangement, if there is mis-alignment, it will be in the fixed position of the disk/motor/motor mount assembly. I am assuming in all this that your condenser and SLR lens are both correctly centred relative to each other.

How have you mounted your disk motor? Have you provided for any adjustment? If you have screwed it straight onto the rear cover of the CD-R case from the outside and have mounted the hub and disk on afterwards on the inside, there may be some flex going on there in the case. If that is what you have done, then a round piece of thin plywood, screwed onto the case surrounding the motor and three screws with the ends filed off flat so they don't puncture the plastic case, threaded through the plywood at the points of an imaginary triangle, each point about 25mm out from the centre, pushing against the plastic to bend it slightly when you screw them in, might distort the plastic of the case enough to give you that final trim adjustment. It would be a cheap and nasty fix which would drift over time as the plastic slumps.

A better method is to mount the motor on a plate and fix this plate to the case. I use three small gutter bolts fixed to the case with nuts, then a small stiff coil spring around each bolt, the motor mount plate sitting on those, then another nut and washer on top to screw down to press the mount plate forward or loosen off to release backward using the spring pressure. In your arrangement, you might need to fasten the bolts to the motor mount and have the adjusting nuts on the outside of the case where you can get at them, as you don't have the same workroom to fit them around the disk inside the case as I have with the pipe caps so you are limited to mounting and adjusting from the rear of the case. It is a better solution for adjustment in the field as you don't have to take a cover off like I do.

If you can get hold of or photocopy a Lemac chart with the four Siemens circles in the corners, these make an excellent back-focussing and alignment aid as you can adjust until all circles are the same sharpness.