http://www.youtube.com/watch?v=TIGN3XdWDUw
I can't find a Vimeo link, but this was shot by Matty_G, RED @ 2K, Zeiss S16s. It's a completely viable option... but, you have to pretty much to commit to it (like any other creative choice, though).
Digital Bolex weeks away
- Thread starter combatentropy
- Start date
http://www.youtube.com/watch?v=TIGN3XdWDUw
I can't find a Vimeo link, but this was shot by Matty_G, RED @ 2K, Zeiss S16s. It's a completely viable option... but, you have to pretty much to commit to it (like any other creative choice, though).
Darren Levine
Veteran
now, if DB did make the sensor upgradable, and if the hardware they got in there was designed to handle more than it's current spec, then all they'd have to do is get their hands on the sony 1" 5k sensor 
now, if DB did make the sensor upgradable, and if the hardware they got in there was designed to handle more than it's current spec, then all they'd have to do is get their hands on the sony 1" 5k sensor
There is an inevitable tradeoff in pixel size vs pixel quality. More is not necessarily better. A different brand of non-CCD sensor would require a completely different set of electronics on the backside. Cheaper to make a new camera.
now, if DB did make the sensor upgradable, and if the hardware they got in there was designed to handle more than it's current spec, then all they'd have to do is get their hands on the sony 1" 5k sensor
Everything but the sensor and electronics is cheap. The sensor, its calibration and the electronics and programming are probably 80% of the cost of the camera, so there'd be no savings. On top of the fact that if you made sensor/chip changes you'd probably want the freedom to change the internal structure of the camera too. The CPUs on cameras improve so quickly from generation to generation that it probably doesn't make sense to hamper a new sensor with an older backend to save 10% on costs, but potentially lose as much in performance.
The only kind of modularity that sortof makes sense is what Sony does with the F5/55, where the sensor/brain of the camera is one bit, but things that can move from camera model to camera model, like the battery and recorder, are modular.
The better single chip 1080 cameras all rely on oversampling. A 3 chip 1920x1080 design will produce close to full 1080 resolution with no in-band aliasing, and no colour aliasing - that is not true of a single 1920x1080 chip with Bayer filtration. The latter effectively gives 960x540 resolution for red and blue - hence significant in-band colour aliasing. To improve results, you have to either go to 3 chip, or oversample.
The general ball park figure is that the best deBayering algorithms give luminance resolution up to about 80% of the photosite count - so to get 1920 resolution, you need at least 1920x1.25 photosites horizontally - about 2400. It's no coincidence that the F3 has very close to that figure and the Alexa has considerably more.
I'm not saying that 1920x1080 therefore must mean bad results. The figures suggest better than 720p resolution - and the original Varicam had a good reputation "in spite" of "only" being 720p. But times move on, the new Varicam will be 4K, and there is little interest in buying 720p only cameras now. (Even if such may still be being used.)
As far as the comments about cropping mode on the F5, then the super16 sensor at about 12mm x 7.2mm is close to a quarter the area of s35. Hence for a total count of 3840x2160, a super16 crop will give an uncompromised 1920x1080. That approach gives all the options of a super16 1920x1080 chip - and a huge amount more besides. The options of s35 4K RAW, or full s35 4:4:4 1080 if a 4K workflow is just seen as too much.
Nothing wrong with what Digital Bolex are doing - just that technology moves on, and the long delay means that what may have been cutting edge at conception, now looks fairly run of the mill at shipping.
The debayered RGB resolution for 1920x1080 with the more sophisticated 4 vector algorithms, luminance-red-green-blue, is nearly equal for all three color channels. The blue channel in a debayered image does not consist of only blue filtered pixels nor do the green and red channels consist of only their pixels. 80% is probably a good figure, which would put optical resolution in the 750-800 vertical line range, equal to or better than most under $10k HD cameras of any kind. Better video and color resolution than the compressed video coming from any DSLR, or any 4:2:0 AVCHD format camera. From seeing uncompressed 4:2:2 broadcast AVI encoded samples on a high end 3-chip DLP projector, this camera will do very well indeed on big screens.
Plus the camera shoots at native 2048 x 1152 and 2048 x 1080 2k digital cinema format as well as 1080p. The max resolution in 4/3 full sensor format is 2336 x1752. That will come with firmware updates and still lens mount options in the near future. So the potential for oversampling is there.
No - but my point was relative to coloured aliasing, and here the effective 960x540 nature of the red and blue is highly relevant. Imagine the difference between a 3 chip and single chip design being presented with a pattern corresponding to limiting resolution - say alternate white/black vertical lines of a resolution corresponding to a line per column.
In the 3 chip case, then (theoretically) there's no problem - output as input, and alternate white/black lines - each white line is stimulating R,G, and B photosites on all three chips. In the single chip case it means that alternate columns get input - but then they will either be all red and green photosites or all blue and green. In the extreme, you could (theoretically) get all cyan or all orange output, and nothing the best debayering mechanism can do about it. Practically, it's not going to be such a "perfect" case as that - but you do see exactly that type of colour aliasing at around exactly that spatial frequency. That's why if a Bayer filtration, the chip resolution should be considerably higher than system resolution. For 1080/2K output, you want a chip with considerably more than 1920x1080 photosites.
It's a bit messy, but the 80% figure is generally taken for LUMINANCE resolution - the CHROMINANCE resolution will be lower. Reasoning is along the lines above, and the 960x540 effective resolution of the red and blue photosite matrices.
The actual figures will depend on subject matter to an extent, but think of an example when the only illumination is saturated light - red or blue. (Which may indeed happen in such as a stage show.) The resolution may well be 800 lpph with white light illumination, but drop to about 500 lpph with (say) saturated red or blue light. In other words 4:2:0 is roughly what is coming off the sensor related to a 1080 system. (Which doesn't take into account the 80% of luminance, the "4" if you like - so maybe equivalent to "3.2:2:0")
As for DSLRs, then yes, I would expect it to be better than such, though by how much is heavily dependent on the way the DSLR reads the sensor. (And some are much better than others.) And yes, probably true for the sub-$10k market - though not for any 3 chipper with 1920x1080 chips - you'll get better from anything like an EX1, HPX250, XF300 etc. I say again - I'm not by any means saying the specs suggest the output of this camera will be "bad", the most I'm saying is "not now cutting edge". And that's because it's been delayed. 720p used to be "cutting edge" - it's certainly not now, even though it may still look OK. Same comment applies to factors like the max framerate. OK, it'll do 30p - but nowadays people have got used to more, even if only for overcranking.
The 2048x1152 may help a little in 16:9 mode, and if it can manage 2336x1314, then that's getting a lot better. But then aren't we getting away from the ethos of super16 if the frame is too big for 16mm lenses? If we need to use still lenses, then why not just go for an s35 camera, or a 4/3 for that matter?
groveChuck
U-matic Member
But how does it look? :huh: :laugh:
Samuel H
Major Contributor
Modern debayering methods make all those calculations of resolution really difficult, since they use all kinds of awesome stuff to interpolate the missing information (edge detection, for example).
Best way to approach this, by a very long shot, is: point a camera at a zone chart, see what you get. You'll soon be able to do that
Best way to approach this, by a very long shot, is: point a camera at a zone chart, see what you get. You'll soon be able to do that
starcentral
Veteran
But as I understand it, D16 shoots RAW and therefore the camera does not perform the debayering correct?
So what and who's software in post is being used to debayer, etc ?
So what and who's software in post is being used to debayer, etc ?
starcentral
Veteran
what I wonder is when Sony calls a sensor "effective" do they say that because they are deriving it from MORE or LESS actual pixels. As most people know, most camera manufacturers use sensors with larger mega-pixels than would equal the 1920x1080 resolution for example. So are those extra pixels really cropped and chopped off, or are they using MORE pixels to get a better HD image for example with MORE resolvable lines of resolution.
I have not put my Sony F55 on a resolution chart for myself to see, but I'm curious to know its resolvable lines in both HD and 4k mode.
I have not put my Sony F55 on a resolution chart for myself to see, but I'm curious to know its resolvable lines in both HD and 4k mode.
philiplipetz
Veteran
They include Pomfort Clipware, and it looks like a really good total solution for handling everything from debayering, color correction, sum checked backups, to viewing dailies and on set.
Effective pixels are those that are actually recorded as part of the image. There are always boundary pixel regions on a sensor that are not recorded. With Bayer sensors, the boundary pixels may be interpolated for greater debayer accuracy of recorded pixels at the edge of the image, but cropped out of the image itself.
demoboy
Well-known member
There's no way D16 will survive if the price of the cam is $3000. AND 16mm lens coverage, and somewhat scarce, is the only real ideal sense to represent the image. with that being said, if D16 proprietary lenses are made for this cam there might be hope. using lenses with a bigger flange is the reason why DSLR videomaking is all cooky. as long as lenses fully cover the expected sensor size the prettier the image. This is what I like about MFT, DX, and Full Frame capture. you get the image of what the lenses was calibrated to with respect to 35mm. Or also S35 coverage and Cinema Lenses. same thing. I hate cropped images of using a lens that's bigger and not designed for the sensor. sure it may sharper, but that's not what the lens coverage was designed for. this is the whole reason why lens manufacturers continue to improve on chromatic aberration or consistent sharpness on the edges. most of you would agree. full coverage of the image is the most important thing. If I can rent a D16 kit with its desired 16 cinema primes and proven the best calibration range then I'll definitely do it.
resonance_image
Active member
16 and S16 vintage lenses are readily available on Ebay or other used sources like Visual Products or Alan Gordon this is why the DB S16 makes much more sense than the compromised neither here nor there sensor size of the BMCC.
Darren Levine
Veteran
Advancing tech always comes into play. imagine if the gh3 hack team manages to get raw? or the rx10, or the nex5/6/7? the only two tricks the db has that keeps it treading water are raw, and arguably the CCD sensor. i say arguably because you have to consider the vast amounts of high end cameras shooting brilliant CMOS images. rolling shutter on cmos cameras has been very well tamed on most new releases.
So ask yourself: if the DB shot only avchd 24mbps, how much of its appeal would it lose? do you really like the form factor? do you really like the feature set?
First gen is always the rockiest, and if the db team decide to go for a v2, i would bet on them putting out a respectable competitor, seeing as they do take in much of the input the users give them. trouble is, by the time they do that, blackmagic will be in v3 and have worked out their kinks. and the majors will have released dozens of new interesting cameras. fading into history is far too easy these days, and thats where RED has it spot on with their modularity/upgradability.
So ask yourself: if the DB shot only avchd 24mbps, how much of its appeal would it lose? do you really like the form factor? do you really like the feature set?
First gen is always the rockiest, and if the db team decide to go for a v2, i would bet on them putting out a respectable competitor, seeing as they do take in much of the input the users give them. trouble is, by the time they do that, blackmagic will be in v3 and have worked out their kinks. and the majors will have released dozens of new interesting cameras. fading into history is far too easy these days, and thats where RED has it spot on with their modularity/upgradability.
So only cameras that are dirt cheap and just like the cameras that everybody else is already producing will succeed? I think you are wrong about that.
A reasonably priced high quality replacement for 16mm film cameras is exactly what many of us have waited in vain for from other manufacturers. It is a superb general purpose production format, large enough for a decent range of creative DOF control, and small enough that fast high quality wide range zoom lenses are light and practical. Great documentary format.
There are more lenses made for C-mount than any other lens mount in the world, from $20 security camera lenses to $3k cinema grade lenses. You can even still buy new Bolex Switars.
Interchangeable lens mounts let you use just about anything you want to. PL, Canon, m4/3. But the camera is primarily designed for cine grade lenses in a classic film format.
The D16 is a very robust high quality camera: premium grade components, all metal construction, sealed weather resistant body, ready to shoot out of the box needing nothing but a charge and a lens.
Its image quality will stand up well against anything anywhere near its price and it can produce a native 2k digital cinema format image with full DCI P3 wide gamut color if you want to take it down that path. The image has a very smooth organic film like quality free of moire and aliasing under conditions that would make BMPCC video unusable. With excellent lenses, resolution and detail are close to as good as HD gets. It looks fantastic on a cinema screen with a high end 3-chip DLP projector.
It records the best quality audio of any camera on the market at any price with excellent studio grade low noise preamps, phantom power and XLR connectors.
Right now the closest thing to the D16 you can actually buy is the SI-2k if you want S16 format, or maybe a Red Scarlet or Sony F5/55 when they get the 2k crop mode implemented.
$3500 is a serious bargain for a camera like this.
There will be a significant market for film schools looking to replace their 16mm cameras since they can use the same lenses and accessories with the D16. It is an ideal camera for teaching fundamental cinematography skills since it requires full manual operation and is calibrated for use with light meters for exposure. The ISO range closely matches the range of typical film stocks.
Last edited:
demoboy
Well-known member
Well I remember hearing about the Bolex more than a year ago and I gave up on them because the development kept on lagging, but I guess that's the culture of kickstarters. As the D-Bolex is an already niche product, it becomes even more niche and non pro when enthusiasts have to resort to eBay or old lenses. Let's be honest, it's do-able but not practical. I just guess that this cam was $3000 or I read it somewhere, and I would think it to be around $2000...Now it's being confirmed that it's around $3,500. The problem with the Digital Bolex is that the philosophy has been contrained to S16 ''purists'' and therefore a scouted sensor to cover old lenses. You realize it's hard for someone to scout a consistent 3 lens kit at the very least to match this unit. Had they just branded it the Digital Bolex without any ''S16'' marketing with a sensor perfect for MFT or APS-C (NEX) I would love to use this cam with Zeiss's Tout series, but no...
While it's true that digital cinema cameras are a niche product (that is to say, the market for them is much, much smaller than for, say, smartphones or even digital stills cameras), and the Digital Bolex represents a niche of that niche, I agree that there's every reason for it to succeed. As Razz16mm said, there's a need for a digital 16-mm/S16 analogue. The greatest danger to the DB's survival is the DB itself -- spotty availability, missing but obviously useful/needed features, poor reliability, buggy implementation of hardware or software, and failure to meet announced specs. In other words, the issue that have plagued the BMCC. But despite being announced before the BMCC, the DB will begin to ship well more than a year later. Ideally, the extra time has been sufficient for turning out a nearly flawless product in reasonable numbers.