Mei Lewis

I think you should both be very careful using any of the resources you've linked to. I assume you're interested in the camera for video (given the forum we're on) but a lot of the information available about ISO relates to stills. What is best for stills shot in RAW and what is best for video might be very different. Video is captured in a compressed, 8 bit form with locked-in color temperature, similar to jpeg stills. There's a lot less information in each frame than is in a still RAW image. The link you posted Jason http://www.cryptobola.com/PhotoBola/Canon7D_ISO.htm specifically says "The conclusions are relevant only for those, who are working with the raw data; the considerations are different for the "JPEG shooters", choosing a setting, which is ideal for raw data may cause the JPEG image to become too dark or too bright. " The vimeo videos don't seem to say if they're about stills or video, I assume because they're moving it's video. They've exaggerated the results though to make them visible and they're only looking at noise which is not independent of dynamic range (see below). ----------- AFAIK in stills general it's the whole number non-expanded ISOs that are 'best' e.g. on the 5D2 that's 100, 200, 400, 800, 1600, 3200. 'Best' here is in terms of noise AND dynamic range. I believe some intermediate settings are less noisy but at the expense of dynamic range. As JPEG and video have less dynamic range to start with it may be that some of the intermediate ISOs are better there as the loss of dynamic range isn't apparent but there is less noise.

That makes a lot of sense, think of it as CINEMAtography.

He must feel really threatened.

This is just an argument about semantics. 'Cinematography' used to mean filming things in the real world, but now general usage of the word has loosened to include animation and CGI. Words change their meanings over time. Saying that what was done in Avatar isn't cinematography is like saying the word 'gay' doesn't mean homosexual.

Ok, I'd like to help you out if I've not got other work on. Can you email me so we have each other's contact details? I'm: Mei Lewis 0781 685 7450 mei@missionphotographic.com

One way to do it: Have them miming to the track as you record the video, and record the audio they're miming to along with the video as a guide track. This audio doesn't need to be high quality. Take the video (along with its audio guide track) into your video software, along with a hi quality copy of the original song. Pan the original song audio hard left, and the guide track with the video hard right. Manually slide the video around until it's guide track syncs with the original song. Having the audio panned left and right makes this easier.

Thanks. I guess some people just oversimplify and say it's discrete, wheras there actually is some energy in the 'gaps'. Looking at that graph I'm not surprised I can't get good color out of it.

I mighty have got that wrong then. There are numerous sources on the internet that say the flourescents spectrum is discontinous (eg http://www.idigitalphoto.com/dictionary/discrete_spectrum) and I'd assumed that was correct because of my own experience trying to color correct properly shot stills under fluorescent lighting.

I don't really understand the OP's point. "Strobist" isn't an artistic movement or a philosophy, it's a self-description term for still photographers who use mostly small, battery powered flash, mostly off-camera. The word only exists because one of the people who popularised the style (David Hobby) wanted to have a made-up word he could have some ownership and control over. Most people who would describe themselves as 'strobists' might be a bit precocious, but all they're trying to do is learn how to use light to make things look nice. And that's a big part of cinematography too.

I think this is probably the same issue as you get shooting standard fluorescent tubes, although they are much worse. With those, there are 3 problems: 1) The light is greenish. That's easily fixed by setting the correct color temp in camera. 2) The light flickers. Fluorescent (and HMI?) tubes cycle in power and color temp slightly every time their power supply cycles. You're probably seeing this. As the shutter moves across the frame different parts of the frame are seeing slightly different brightness and color temps. For ordinary fluorescents the 'safe' shutter times are half the power supply cycle time, the whole power supply cycle time and any whole number multiple of that time. E.g. in the UK (50Hz supply) the safe shutter times are 1/100, 1/50, 1/25, 1/10, 1/5 etc. My guess is that HMIs cycle at much higher rate, so that movie cameras with a typical shutter speed of 1/48th or similar, don't notice it. But at 1/400th you have. 3) The spectrum of fluorescents is incomplete. There are gaps where colors are simply missing from this supposedly white light, and anything in the scene which includes those colors won't be represented quite correctly. I don't know of any proper solution to this. For me, no matter how careful I am about (1) and (2) there's usually a slightly odd feeling to the pictures resulting from this incomplete spectrum.

So what's is the 1.333 ratio all about?

The 5D2 shoots 24p now after a firmware update. I think the stills and video world are getting closer, but there's so much that separates them apart from the camera they're not going to merge anytime soon.

That's called focus stacking right? Do you have a link to more info?

Do you know when you're filming?

That's great, thanks Jay. Hal, thanks for the tip. I don't think the production I saw used toughened bulbs, they broke one during the show!

Okay, now I'm trying to understand the difference between Bayer and co-site. After reading around, it seems both have an array of photosensors overlaid with a grid of color filters. In the case of co-site sampling, to make a single image several exposures are combined with the sensor being moved between exposures, so that every point is scanned at least once in each of the three colors. With the Bayer system a single exposure is made so that each point is only scanned once. Colors are interpolated from neighbouring photosensors? The Bayer pattern repeats on a 2x2 grid, so a naive first guess would be that however many photosensors there are on a Bayer chip you'd have to divide by 4 to get the number of color pixels it produces. I'm sure the number 4 there is wrong, but I don't know what the correct number should be. I'm pretty certain it would have to be strictly greater than 1 and no more than 4 but that doesn't narrow it down much! To put some numbers on that, say there was a 1000x1000 photosensor co-site system. Then the number of full color RGB pixels you'd get out would be 1,000,000. If you had a 1000x1000 Bayer system, then the number of full color pixels you'd get would be 250,000? Or maybe 333,333? (i.e. 1/4 or 1/3 of the number of photosensors.) I'd always thought that when a manufacturer quoted a camera as being 10MP say, there would have to be roughly 3 or 4 times as many actual photosensors, i.e. 30 or 40 million.

There's a book "Painting with Light" about this sort of thing (amazon link), has anyone read it?

Hope someone can help me with this. I was at a theatre show a few weeks ago and one section was lit by a couple of the performers who held lightbulbs on the end of electrical cable and swung them around slowly in circles. The bulbs were clear and the filaments so dim I could look straight at them without hurting my eyes much and see the glowing red wire. Here's a photo I took of one of them in use: (The bulb is blown out because of the limited dynamic range of my camera, not because it was very bright, in fact it was very dim.) I thought they must be special bulbs, but I talked to one of the performers and he said they were just ordinary clear household bulbs on a dimmer. He hadn't made the setup himself though so I couldn't ask him exactly how it was done. Does anyone know how I can make such a thing? Can I just buy the right sort of bulbs, a bulb holder, a dimmer and some cable from a DIY store? I'm in the UK with 240V mains and I think this option might be a bit dangerous.

I have a related question someone may be able to answer. In Premiere Pro CS3 all of the project presets relating to 1080 HDV (which I assume is just short for High Definition Video) have non square pixels, "Pixel Aspect Ratio: HD Anamorphic 1080 (1.333)". The two 720 options have square pixels. Why is this? I was trying to make a project with some files from my Canon 5Dmk2 which I shot at full HD 24p, which I think is 1920x1080, progressive, with square pixels, and actually 23.976 frames per second prgressive. I wanted the output to be the same as the input and had to make up a custom preset. Is the non-square pixel issue something to do with a broadcast format?

The box set has recently come out in the US and there's an audio interview with the actors who played Dutch and the main cop's wife here: http://www.blogtalkradio.com/back-by-midni...gets-away-clean

That's great.

Okay thanks for replying so quickly and clearly. -- In that case I don't see that there's a problem with what Red is saying. Few people buying a digital stills camera would think that the number of pixels directly translated into the amount of detail the camera/lens/processing system could capture. For example, a 6mp phone camera is going to be nowehere near as good as a 6mp Nikon or Canon DSLR with stock lens. I would expect anyone considering spending tens of thousands of dollars on a digital video camera would know that, right? Or was Red claiming something else?

Yes, I live and work in Cardiff. I'm a stills photographer and In the last year or so I've got really interested in doing stills work on films, which is why I'm on this forum. Unlike seemingly everyone else with a recent DSLR I don't want to start using it to shoot video, I like static photos too much, but I do want to learn about the approach of film making and apply it to my stills work. Some of my film still work is here: http://missionphotographic.com/commercial/unit-stills/ I occasionally come down to London to visit or sometimes for jobs. Looks like I'm going to be working on a short shot in Farnham next month which should be fun.

Can someone tell me if I understand this correctly, I'm a bit confused. RED markets its cameras by stating the 'resolution' of its sensors i.e. the number of RGB pixels you get out of it. That's 4096x2304 pixels or put it another way about 9.5 megapixels. If the Red 1 had been a digital still camera it would have been described as 9.5MP sensor in its marketing. Is that correct? I know there are ather shooting modes, like 4480x1920, but my basic question is do the pixels that Red is talking about equate to the kind of pixels manufacturers like Canon or Nikon are talking about in their stills cameras? Thanks.

Hahahaha! Love it!