Axel

With raw photography, with a conservative exposure, 'ETTC', you also get a misrepresantation and end up in having to change the exposure in post and eventually reduce some noise. Why? Because what we have here is a 12-bit image. Without any curve applied to the quantization (for the preview!), an 8-bit image can only show less than 5 stops of latitude (256-128-64-32-16, anything below 16 becomes indistinguishable on any common 8-bit display, and in reality none can show values above 240 as well). This is an extremely low dynamic range, and even with a new tone-mapping, the values visible then are a fantasy and by no means represent the true relation between, say, 1354 and 11433 (just above 3 stops in the infathomably depth of 12-bit). What we are demanding when we call the washed-out images a 'flaw' is effectively a display with an upstream LUT to let us see an incorrect, very much compressed image that we are used to in our tiny 8-bit universe. Something like this is probably built into an Alexa or the like. The workaround to shove the ISOs down also (in the preview) just grabs a very narrow range of what we actually record. The same is true for external devices, if their monitoring is not capable of processing 12-bit raw (which is?), for whatever reason, like Germys Atomos Samurai, which would be suitable for 10-bit, for which the built-in histogram should be reliable, but of course not any LCD viewer. As a digital projectionist from 2000 until 2011, I very often compared the MJEPG2000 DCPs (typically 12-bit) to my own 8-bit ambitious amateur stuff. And no matter if I had used opendcp to convert it or just plugged a laptop to the scaler, the differences were huge. It never was an issue of spacial resolution! Three deliberately provocative theses on where we stand with our appreciation of raw: 1. All DSLR bodies have a lot of knobs, buttons and dials as well as complicated menu options and presets. All of these gadgets are phony ways to cripple (potentially) high quality images to some pleasing-looking consumer-jpegs or mpegs. White balance, ISO, looks for different conditions: Crooks! There are two things that count, and neither of them live in the camera's body: Focus ring, aperture ring (EDIT: Well, one: 'shutter'). 2. We still record, grade and view in 8-bit. To shoot raw in order to get a better 8-bit video seems comparable to killing a whale for a cup of cod liver. And we hunt the whale with blindfolds. 3. ETTR will have highlights and some midtones. That the graded images look naturalistic although there are no real shadows could very well be limited to 8-bit. Germy1979 wrote: Unfortunately, afaik, Neat can not neat raw video. Removing noise (useless, redundant data) should be the first step. To do everything right, there is probably no alternative for the BMPCC buyers to buy a full version of Resolve as well. Or ETTR for 8-bit.

  I really hope there will be a new wave of influental independant films that deserve this name. Even worse than being dependant on the politics of major production companies is the situation in my country. It has the "FFA", an acronym for film promotion association, founded in 1968, and it effectively finances german theatrical films with tax money, but approving the standards, which can only be described as uncinematic and boring. It's a federal association that decides if scripts are worth the investment, and you can easily imagine what that means for creativity and originality.    I particularly hope, that indie filmmakers will stop to think retro. If they want to reinvent cinema, there is no frigging 'cinematic look' to be followed, no grain filters, no exaggerated DoF-gimmicks, no oange-teal-'blockbuster'-grading, that would be embarrassing. There has to be a good and revolutionary story that dictates it's own aestethics, preferably clean and straight.   Don't dream about fancy imagery, you've got a camera with sufficient color depth to look good on a big screen. Now it's time to tell something!

Not explicitly. but since ISO 400 and ISO 200 are each a full stop down, you can do the maths: Here is the original BM thread. The overexposed images look as if there were more than just two f-stops off a center exposure (it could be though, but with 13 stops latitude ??? ) So maybe it is still the flaw/bug you described in your initial post, that in raw you can't get a decent monitoring. I'd like to test it myself, but unfortunately this is still theory for me in two aspects: My BMPCC didn't arrive yet, and also there still is no raw mode :( EDIT: Perhaps combing the lower ISO with pulling down brightness on an external device will be acceptable.

You didn't read the post above or saw the clip? The BMCCs (and therefore the BMPCCs as well) always record at ISO 800 in raw. They also record at ISO 800 in ProRes, only that then the brightness, DR and noise changes caused by the setting are baked into the compressed file.

Probably the same Tom says on the BM forum: That sets my mind at rest: In raw mode, you can use every ISO you want, without changing the exposure of the clip.

In the comments to the ETTR video on vimeo, Majerski elaborates: Because of how the Canon DSLR's have analogue gain - the dynamic range response is different with the ISO values and although much of the core explanation is the same, there would have to be some additional caveats in order to prevent any mis-information and or confusion. I thought the only difference was the nature of the compression, that common DSLRs with mpeg4 codecs used internal raw interpretation before they compressed everything to a noisy mud. Am I wrong? I am more confused than before.

Good to know. Thanks. Weird, that most cameras bake in values that could actually be better refined in post, isn't it? I never again shot a jpeg after learning about raw photography. It's just fabulous that I will get this for video too. I used to be skeptic about the raw hype, looks like I become a believer.

I have researched on the topic, but there is confusion about it everywhere. It seems reasonable to assume the following (please correct this, preferably with a link to your source, and excuse my awkward english):   1. Native ISO in a digital camera equals the analogue charge, at which the sensors record nothing in complete darkness and are fully loaded in maxiumum light (whatever that means). If it is marked as "800", this says nothing about what's actually happening, but is a rough hint what to expect compared to analog film.   2. If the camera is set to ISO 1600, analog gain happens, some double electronic load or so (arguable, some say it's not, see point 4). Since now everything is enhanced, the dynamic range is lower.   3. The DR also is worse at ISO 400, 200. I am not sure why.   4. Here is the problem: Some state that all increments of ISO dividable by the native ISO are analog, whereas others were digital, but that the smaller values were digital too. Others seem to say that all non-native ISOs are derived by digital processing. But what would this mean for raw?

      I'd like to know more. In layman's terms, are there actually two concepts how to understand ISO when dealing with raw? The BMPCC manual says:     ISO settings are helpful when you are shooting in a variety of light conditions. The optimum ISO setting for the Pocket Cinema Camera and Cinema Camera is 800ASA. For Production Camera 4K choose the lowest ISO for the available lighting conditions. Depending on your situation, however, you may choose a lower or higher ISO setting. For example, in low light conditions 1600ASA would be suitable but may introduce some visible noise. In bright conditions 400ASA would be best to record richer colors.    So it seems there is a weakened signal ("in an analogue way") when I choose i.e. ISO 400 before recording, and it gets "baked into" the raw clip. And then when I raise exposure in post, I will see noise as if I had recorded at higher ISO?   Something confuses me here. Help me to let the penny drop (if this is the proper idiom).     This is nothing. You should see the flamewars elsewhere, especially dealing with platforms. Andrews moderation is moderate. Certainly less 'personal view' from his side would generate more traffic, but also attract more trolls. It's a question of delicate measure.    

  ISO is an analogy for the sensitivity of film stock. The use of this term is *helpful* for estimating the relative sensitivity of a sensor with a certain gain chosen. But actually vague and incorrect. Native ISO means the signal is neither amplified nor attenuated.

  You are right. It is a default Photoshop preset applied to every raw file. And setting it from "25" to "0" practically eliminates what is described as "noise". Keep in mind, that camcorders with intelligent compression algorithms also develop raw images under the hood, that they add gain, NR and sharpening. But sharpening must be applied to the BMPCC images, obviously. Perhaps there are better methods (.i.e. in Resolve, as a last step of grading then, there are also discussions everywhere whether to apply NR first or last, plainly the right answer is first, the node model makes it easy to get the order right - only that there is no denoiser in Lite :( ), but I guess it just needs some understanding on how the tools work:       This is from here. Hystery begone!

I only now realized (on a bigger display), that the plants were no timelapse thing but CGI. Watched the making of. This is very cool stuff. Just one suggestion: Some of the CGI is too sharp, which betrayes the trick. What do you think?

Bardzo dobrze! Interesting camera moves, nice photography, thanks for sharing.

  I say, let's wait for some moving images. Noise is hard to evaluate from a single image. Yes, if it was a stills camera ...   It may be bad, it may actually look good. You know, all the high bitrate - small GOP hacks of the GH2 never added any real detail to the images, they dithered obscure parts in a way like film did, creating so called 'temporal samples'.   Yes, I know, iso noise is fundamentally different from film grain. But, like Andrew wrote, it still is something people add voluntarily to make their videos look more organic (they add random film grain simulation, which like iso noise, has nothing to do with the pixels in the image, it's an artificial layer).   If it really turns out to be an 'issue', there is a good chance that tools like Neat (or the pro version of Resolve?) can fix it easily.   Raw photography is also not free from noise. Adobe Raw within Photoshop has it's own NR filter. The noise you detected on the ship is nothing I would have found in original size (1080 viewed on a 1080 monitor from a viewing distance appropriate to, er, cinema), my guess is that it is negligible for video in most situations. In this photo, the noise disappeares completely with a mild 18% luma NR (visible in 400%), no big deal, since one has to treat the raw with several agents anyway.   What is more, the sensor is said to be comparable to that of the BMCC. The first Brawley demos also showed some noise, but it isn't discussed as problem anymore, afaik. Maybe it can be avoided altogether, maybe not. No reason to despair.   Paramount for me, I like the images, I am not for missing the wood for the trees. Or worse, to ignore the path underneath your feet! 

Here, used lens: SLR magic 12mm. Funny colors in parc sky are CA. Or what?

No interlace! Interlace needs to be deinterlaced when viewed on a progressive device (which are all modern TVs and monitors). Hobbyists (if they don't have 50p/60p) sometimes prefer it, because doubling the motion phases smoothes their shaky camera to some extend (smears the details, that's what "i" does on progressive displays). The better solution is: Avoid shaky camera, this also improves resolution (veeery much, complicated to explain with my limited english)!   24p 24M (the latter probably the data rate in mbps) is the way to go, and 60p (if it looks good) for slomos (don't know though if iMovie treats this well). Use a shutter of 1/48 or 1/50 always to get the right amount of motion blur in your video (1/100 for 60p).   About the PAL dilemma: It never was one, as long as you don't work for your local broadcaster. It applied to DVD only, and even then it was not a problem at all. PAL DVD players always play back NTSC DVDs also ("as" PAL60). 24p are NTSC conform. For SD (I don't know if iDVD will make problems, but I don't think so. It's big brother, DVD Studio Pro, just said, "the frame rate does not match PAL", so you had to change the project properties or the frame rate).   24p are also BD conform, no matter which country.   Viewing on a computer: Most monitors have 60Hz frequency. Now, even if you played back 25p on this, the very light swallowing of frames never disturbed anyone, it's hardly noticeable. Also the other way around: If the refresh rate is 75Hz, a 24p movie will still look fine.   As of now, 50p/60p are not supported by youtube, vimeo (they transcode it tacidly to 30p, in case of 50p originals with stuttering) and most BD hardware. 

Owner of the G6 since yesterday. Colors are definitely nicer than those of the GH2. Now I am all for grading, not because I detest good video out-of-cam, but because it's fun. An image that shines does so because of vivid skin tones. So it's not just a matter of shoving colors in post to cover the so-called skin line, but mainly because a broader spectrum around that line was recorded in the first place (must be the Canon magic). Otherwise your motifs will look doughy. I don't know yet, how good the colors actually are compared to EOS. With raw, you shouldn't have to worry about colors, if post-grading raw photos is comparable. I will have the Pocket (hopefully it won't take too long), and it will have a complete, well-thought-out rig from the beginning, and some day raw, for the price of a Mark III (body only).

You could be kind and post your experience here afterwards. Imho the risky part is, when you are tempted to improvise to get a really nice shot, but deviate from your experiences with the camera. For example, when it seems faster to change exposure by rotating the ND fader, and you realize only later, that there was a terrible color shift within the take hard to balance in post. Shot in question: The pair prepares to leave the church, the doors are opened (you follow with a pan), and the light explodes (I've done this two times before with the manual aperture of the XH A1, and it was a soft transition, one time I was not told that a flock of white doves was to be set free to welcome the pair. I comped them in - actually they were seagulls from an old holiday video, nobody noticed). You have no time to think of alternatives, everything is live, irreplaceable and unrepeatable. That's why I wrote in another thread that wedding videography can easily be underestimated.

Yet I admire your chuzpe too shoot a wedding with a BMCC. But not raw?

As you wrote, ND faders tend to produce side effects like unexpected, irregular vignettes (which I can confirm, I have a Heliopan in exchange for a LCW, same thing). So if you know you will always need an ND, it seems the wise thing to look for a better solution. I talked to a photographer the other day, who told me, if he was limited to 1/48 (is that 180 degree on the BMCC?), he'd really have to use strong NDs. Isn't this really, where a matte box was in order? You minimize the risk of reflections, you can have one 4x4 for all your lenses with different filter mounts, you can change them fast. Unfortunately, good 4x4 filters like Tiffen are very expensive. Sorry for not directly answering your question, I just share my thoughts. Having lived with a lot of compromises the recent years, with the decision for a BMCC there should be less. It was adequate to get a camera like the GH2 to produce more cinematic looking video by adapting funny lomographic lenses. But this is a cinema camera, time to reverse gears.

As shown here, there is actually an even cheaper card for ProResHQ, the Transcend 64GB: What nobody knows so far: What cards will work with the compressed DNG? A friend of mine bought the TV Logic Alphatron EVF, and it's not as heavy as it looks bulky. Contineo cage (just the cage, the rest can be assembled with cheap enjoyyourcamera stuff or parts I already own), rods, follow focus, small matte box (for the inevitable ND fader). Can still be a compact setup, and with the Pocket is about the price for a 5DMIII body (which also needs rigs).

The Z1 is an HDV camcorder from 2005, the pro version of the F1 I worked with then. It is the first HDV cam. The resolution is 1440 x 1080i, which doesn't quite reach 720p effectively. It pushed the quality for SD-DVDs tremendously, but for modern expectations ... There are more issues. Because of the interlace video, it always looks videoish. Back then people deinterlaced to get rid of the "i", but that looked convincing then. The tape recording forces you to capture prior to editing. That said, the form factor, the perfect automatic program, the decent lens with perfect servo zoom, the phantom XLR inputs, made it the perfect camcorder for event videography. And it still will be - if DVD is your target. I don't know the Nikon. In the german slashCAM DSLR & camcorder comparison it only gets a "c" in general, but only a "d" in 'subjective image quality'.   I had the 7D. In the comparison, it got an 'a' for handling, but only a 'b' for image quality (that was in 2009, I doubt it would get more than a 'c' today. The subjective image quality is not bad, perhaps a little soft, but the actual resolution is about 720p). Compared to your Z1, all VDSLRs would get a 'd' for handling (for video). Be aware of that. To overcome this (in part), one has to configure a *personalized* rig to be able to stabilize the camera. Since there is no reliable auto focus with a DSLR (even if it's fast enough, it can't know where you want the focus to be - search for 'shallow depth of field'), you are forced to focus manually, and only with a wide angle lens and hyperfocal distance to the motif you can use a flycam at all. I wrote *personalized* rig, because only you know, if you deem the naked LCD display enough for critical focussing (hardly, and if so, only with > peaking, but think of sunlight hitting the tiny screen), prefer an external monitor (this will extend the length of the rig and you will wish to have a  > follow focus) or press one eye to a > viewfinder, which requires additional add-ons, because VDSLRs don't have an EVF (some mirrorless system cameras have, I would personally recommend a Lumix). You see, this is hardly a thing to solve in a fortnight. I forgot to mention that in comparison DSLR audio tends to be abysmal, another problematic area. There are modern camcorders like your Z1, using cards and mostly AVCHD, with progressive frames and close to fullHD. I am not well informed about what gives you the most bang for the buck there.

Explain please.

Another thought: If there is no AWB, how do you set white balance? Thungsten or daylight? Seems to be a little too imprecise. EDIT: I see Not so easy, since I learned a camera "sees" differences roughly in steps of 300°K, the human eye doesn't ...

Two things shocked me: Hot pixels, dead pixels? What a horror. I once had a hot pixel on my 7D (which had much more pixels to be able - in theory - to skip the bad ones) and it still appeared occasionally after the Canon service did something about it. I complained again and got a new body. The description of the display sounds much worse than I had expected. I didn't bookmark the link, but I read somewhere about a month ago, that Sony SDXC cards 95MB/s were usable too. The Sandisc Extreme costs almost twice as much. Can someone confirm this? Handheld is completely impossible, but my understanding is, that there are so many ways to stabilze a camera, that I can as well use all. A pistol grip with a Z-Finder will be the most basic rig I will test (see avatar). People kept telling me it didn't give enough stability, but they just don't know where or when it works or how to hold in the first place. And maybe it won't be long until China gets the mini MOvi clones right and affordable ;-) I actually muse about an external viewfinder (there are at least two with 'retina' resolution by Cineroid and TV logic). Many BMPCC clips on the vimeo channel look soft, but only because the focus is set to infinity while the main motifs seem to be less than thirty feet away, which lets them look unfocused, especially with the comparatively long lenses used. For me this is proof that peaking is unsufficient for these situations.