Indie Film Live

Friday, December 23, 2005

Discussion on using the P+S Mini-35 with HD100

While trying to keep up with everything Spoon related I noticed two of the filmmakers discussing the camera setups and using the P+S Mini-35 over on As they do refer back to here, I thought I should link back. Here is the thread.


  • Hi David,
    This is copy of a the posting I made to a thread on You can find the link here Threads such as these tend to be very technical and my understanding for this blog is to lay open some of the more general aspects of film making on an indie budget. However rather that re-writing the whole piece a thought I’d just plonk it here and let the blog take it’s own direction henceforth.

    Most of the following relates to the use of the new (late 2005) JVC GY 100 camera in a series of test being used with the P & S Technics mini 35mm converter. I'm the one standing by the Mobile Wafian Mobile (MWM) in my fast red jacket.

    “It is about time I broke the seal and made a posting on this site. It has been and continues to be my pleasure to be working with Simon Hansen and Sharlto Copley as DoP on 'Spoon'. We are all taking a bit of time off for Christmas readying ourselves for the 'Big Push' come Jan 6 th - I'm off doing a bit of R and R out here in the slopes in Switzerland yet my dreams are all still filled by Spoonic imagery. It's funny how this industry gets under your skin; she's always there tugging at your elbow or beckoning your thoughts away from level living. A strangely unrequited love affair in which we all find ourselves vying for the attentions of a beautiful yet slightly aloof mistress. Ah but I find myself wandering off on a euphemistic tangent – it is probably something to do with all the Charles Dickens and Oscar Wild that the BBC love to show at this time of year…

    Anyway, back to the point.

    We've been testing the hell out of the JVC GY 100 pushing and pulling it in all sorts of direction but it has been all a bit different to the kind of testing I am used to. I suppose there has probably been some sort of perverse desire to find a magical combination of settings that will give us 'that look', some as yet untried combination of menu tweaks that will to pull us out of the quagmire that is video and bring us nicely into the realm of film making. Perhaps ultimately it is all a bit ego driven but I have to be honest in saying that I have found this testing very testing. What I have found so particularly taxing is the almost infinite number of menu permutations that could potentially be employed to some how maximise the image capture. I have always had a somewhat discomforting feeling that those techy chaps at the video camera factories don't actually know very much about film making and so this testing has been about confirmation that the tech-heads that made the cameras know what they are doing; to be satisfied that the factory standard settings are not just set up for John Doe making his crappy family holiday video.

    So far this is what we have decided upon;

    Master Black.

    In some situations it can be useful to boost this above 0 or normal. Naturally the image comes out looking very flat as there are no true blacks and so it has to be graded but by crushing the backs back in the grade the effects are pretty much indistinguishable from shooting with master black at the normal setting. I'm not too sure why this is and was rather hoping the final results would be similar to shooting film with a Varicon. I have even tested a Panasonic Varicam and a Canon XL1 with a Varicon against simply adjusting the master black and found the results to be so similar that the extra effort and cost of using the low con filter system seemed to be unworthy of the effort. There was a qualitative difference to the way the Varicon treated the shadows as compared to pushing the master black but the results always seemed to come back to parity in the grade. Which ever way I tried to get more info from the shadow areas of the scene I would always end up with a graded image that was over run with pixel noise. One idea we have had is to shoot a very flat image and grade the blacks when we go to film print. We are waiting for our first set of test to come back from the labs and who knows maybe there will be some lovely combination of digital capture and chemical grading that will get us away from those ugly chunks of pixelated shadow detail we all hate so much.

    My personal goal with all these low con tests was not to find a way to mimic the contrasty and de-saturated look of a bleach bypass or ENR grade as that is very easy to do in a digital grade. Rather it was to find some way of eeking out an extra stop at the top end by being able to underexpose the image more. But so far I don't seem to have come up with the goods. Bah.

    Black Stretch.

    All this seems to do is push up the gain in the shadow areas. Useful but only at the expense of a noise increase.


    At one point we thought we had found something really interesting were we could really make a difference. Turing the detail function off gives us a true native image of what the CCD chip is actually reading. This somehow seemed to be a good thing. The detail function applies some sort of algorithm to the image before it is captured artificially increasing the contrast between individual pixels to create an image that 'appears' to be sharper. Since it is effectively a post effect we figured we would by-pass that process in camera and apply our own sharpening to the image during the post process that way gaining full control. Theoretically a cool plan but we have yet to find a plug-in for either Combustion, After Effects or Premiere that increases the apparent sharpness as well as the in camera detail function. Our first series of test with the Mini 35 converter were all shot with the detail function off and the results were far too soft.

    It may seem like a slightly long winded way of trying to achieve sharpness but there was another reason for trying this route. We have been working very closely with Cineform in trying to get the most out of their wavelet compression algorithm and one of the things we discovered was that shooting with the detail off reduced the data rate of the encoded files. Whilst the CineForm codec is effectively visually lossless we were finding odd artefacts appearing when the detail was on. It seems now though that CineForm have solved this issue by offering higher bit rates to which the codec can write.

    The other aspect that bothered us about the detail function was that it increases the level of the background CCD chip noise levels. All chips produce a certain amount of noise – if you shoot a plain coloured wall and zoom in on the image you can see the attributes of the digital noise. It is around these areas that I have always felt video really falls down. There is a truly chaotic randomness to the type of noise (grain) that you see when you examine a film negative that is not present in a video image. CCD pixels are always in exactly the same place conforming to a grid pattern which is why we get problems with aliasing and jagged diagonals. It is this fact that has largely given rise to the need for artificially boosting apparent sharpness, problems that are especially noticeable in static wide shots. Turn detail off and all the fine detail has a soft edging to it. Turn detail on and the apparent sharpness is increased but there is a latent ghosting to all aspects of the image. We are essentially down to choosing between the lesser of two evils still and this really bothers me.

    So at the moment we are back to that standard solution of turning the detail function down by a couple of clicks but living with the standard results obtained by all and sundry unless anybody is aware of a post process that replicates the in camera detail function.

    In terms of using the P & S mini 35 I would have to say that in general I am off a similar mind to Charles Papert who wrote:

    1). "Sharpness" is a complicated term when used to compare film and video. I think that the slight softness that the Mini35 creates is part of what makes it look more pleasing to the eye; I find straight HD in general to sometimes be overally sharp in appearance…

    The Random nature of the 'noise' we are getting from the ground glass of the mini 35 is going a long way to eradicating those problems I have just mentioned. There are however a whole ton of other problems thrown up by using the Adaptor.

    Again I quote you Charles:

    "Regarding which lens to adjust exposure, it is standard practice to set the cine lens at the desired aperture for depth of field purposes and adjust exposure via the relay. You should not see an optical difference with the relay wide open vs stopped down. When you say that the image was cleaner when the relay was stopped down, in what way? If the backfocus of the Mini35 is not properly adjusted, you will see a focus shift as you stop down the relay (it will improve), so remember that backfocus should be adjusted with the relay wide open and the cine lens stopped down to make the groundglass texture more apparent (Mini35 motor off, of course)."

    Yet our ground glass was very carefully set – I even recorded the process to tape in order to be able to review the results again. Yet by stopping down on the relay lens we were in fact getting a sharper image. Actually a difference of only 2 full clicks on the relay lens made a very marked difference to the quality of the image. As well as the increase in sharpness there was a very noticeable change in colour cast. There was a warm yellow hue when shooting wide open that disappeared when the lens was shut down.

    So now I turn to Obin Olson who I think has hit the nail on the head:

    "It's very true that the relay lens will make your images softer when it's open, the reason? as far as I can tell a 1/3rd inch ccd is so small in size that your almost at the OPTICAL limit of glass to get above a 720x480 3ccd image, with the dvx100a and it's pixel offsetting we are getting 1540x984 HD frames from the camera, It's very true that the relay lens will make your images softer when it's open, the reason? as far as I can tell a 1/3rd inch ccd is so small in size that your almost at the OPTICAL limit of glass to get above a 720x480 3ccd image, with the dvx100a and it's pixel offsetting we are getting 1540x984 HD frames from the camera, however when you open the dvx lens all the way IN HD ONLY it's lens goes soft, not so with SD, that looks great...not enough resolution to show the softness. "

    The hardest part of lens grinding is tending to the outer most regions which is why it is always advisable to stop down by at least one stop from wide open. It is nearly always the 'thin bits at the edges' that give aberrations and soft focus issues. A cheap fast lens will usually be thin at its edges as compared with a higher quality lens which in effect is just the central part of a much bigger lens. Take a look at the relay lens as provided by P & S and it just looks too small. I really don't know much about lens making but there just does not seem to be enough glass there. And as such it is not surprising that when used wide open we are getting focus issues.

    Under ordinary situations the solution would be to simply stop down and deal with it but 'Spoon' is a dark film and with the mini 35 we are so starved for light at the CCD that it isn't really an option for us.

    There is also the issue of nasty blue chromatic aberrations around any high lights that are probably as a result of poor achromatic lens coating. (I am guessing here). The Fujinon lens that comes with the camera also has nasty green/violet aberrations. Since the image is split by a prism and captured on three separate CCDs (red, green and blue) digital lenses have to be very carefully coated to ensure that there are no aberrations. I can only assume that this is probably a lengthy and expensive trial and error process to calibrate each lens to the specific attributes of each digital camera. As far as I know the relay lens used to connect the mini 35 to the JVC GY 100 is exactly the same glass that it used to connect the Cannon cameras and maybe it needs to be coated differently.

    Anyway, putting aside any light loss issues of using a min 35 my feelings at the moment are that the relay lens really is a weak link in what is other wise a very finely engineered piece of kit.

    I now turn to Dennis Hingsberg's posting;

    " Your problem however seems to be the opposite of all of this and having shot in night interiors and exteriors all I can say is shooting wide open (f1.2 or f2) on a 35mm lens poses some extremely challenging focus issues. For me mainly finding focus but also pulling focus if necessary for the shot. When directors want to dolly in or out under low light I cringe since usually I pull my own focus, an experienced focus puller would obviously help the situation but my point anyway is that you will get a wider depth of field with a lens that is stopped down to f3.5 or f4 which could make for some easier focusing but of course require more light. :

    Bang on mister. On set focus is one of my biggest fears at the moment. In fact it makes my bum go funny just thinking about it. We will have to shoot a lot of the film wide open T1.3 with a lot of work being done on tighter lenses such as 60mm and 85mm. Even a set hardened seasoned veteran focus puller would think twice about taking on this gig. Add to that fact the very low quality of the eye piece view finder and I see clouds on the horizon. My first reaction to this was to find some way of taking a split from the eye piece monitor and sending a signal to a small HD monitor which my focus puller could use but this would almost certainly need a remote focus assist for which we simply don't have the money. If anybody out there has some good suggestions…

    Dennis, you also wrote;

    "It might not make a lot of sense at first, but I've often suggested on some productions not to use the mini35 adapter for low light shots. Often depth can be achieved through good lighting and placement so perhaps consider that as an alternative option."

    We have some pretty big night exterior set ups to get through and we have discussed using the native Fujinon lens. I completely agree that we could create depth in the lighting and so I am slightly less worried about the massive disparity between the 35mm DoF and that of the 1/3 inch CCD with it's native lens. However the Fujion lens has a very different look to it, for one any bright spots at night give off a star effect which looks very video like where as the mini 35 has a look somewhere between old black and white films before the introduction of anti halation backing and a rather heavy handed use of white pro mists and low con filters. If we can find some way of treating the image to match it to the mini 35 this is still a path we might pursue.

    It is late now and it is new years eve and I have to go out and cause some mischief. But before I do I just want add a quick note to something I have only just touched upon. We have been working closely with David Newman, David Taylor and Jeff Yuel from CineForm to really get the most out of their Prospect codec and I can only say that it has been a real privilege and honour to work with such a dedicated and helpful team. Thanks guys. They have set up a blog site for us that deals with the specific issues we have encountered in working with their truly excellent coded. Either visit and search for a blog called 'Indie Film Live' or try opening this link. I'm not really very good with this inter-web thingy so it might not work. Let me know if it doesn't.

    And now I am off to watch the fireworks and take stock of another years worth of life lessons. We grow old so quickly yet I wonder when I might be able to call my self a grown-up.


    By Blogger Joe de Kadt, at 4:08 AM  

  • The HD100 has a noisy chip!!!. It has been quite depressing to see our images close upon a good monior. On the CRT's that looked pretty cool - on our HP 3552 thaey look very grainy.

    I have played with grain removal in post, all the settings everything I can think of to improve it - nothing. Not to mention that the P+S also add grain. So we have all this noise happening yet the aesthetic is so close to great. There are compromises here and there but overall the look we are getting is awsome. The truth is the chip is just too small - or is it.

    I had this suspicion about HDV for a while and today I ran the some tests and I think I get it. Can hear all the guys on DVINFO saying sure obviously we knew that. But I just got it now. The comression is removing a lot of grain. As I understand compression - it tends to preserve detail and sumerise flatter areas into few vairients. Essentially summarising the low detail areas while providing more detail for high areas. At that is how JPG works doesn't it.

    I rember David Newman saying that cinform would perform better with the 35mm lense because of the limited depth of field - is thie related.

    When I captured to HD100 tape at mpeg2 and cineform fillmscan 2 I got two file types of the same thing and sure enough the grain in the Mpeg 2 was way less but the detail still pretty good. I merged the two to produce two files indesinguishable form the Filmscan original other than the fact that some of the grain noise was removed.

    There may be something in this as far as a clean up tool is concerned. Surley with the right compression alternate layers or final piecees can be rendered using compression which while it decreases the latitude in the file it decreases the grain without loss of detail. For a final master no more tweaking need to be done so the resultant image could be enhanced for final output or as a trnasfer layer in th composite.

    Am I nuts or could that work?

    By Blogger simon, at 3:49 AM  

  • Nice article as for me. I'd like to read more concerning that topic. Thank you for posting this material.
    Joan Stepsen
    Gadgets and gifts

    By Anonymous Anonymous, at 7:30 AM  

Post a Comment

<< Home