Camera industry in the dumpster - article

[jjj]

As I mentioned before I never had a cable from Apple break.

I've never had a heart attack, so should I not believe they really happen?  

Apple replaced everyone's magsafe cable because it was they were forced to do a recall on it, two out of three of my iPhone cables is doing the standard Apple fray and if you want general info on fragile Apple cables. Actually the third cable is showing signs of impending breakage in the exact same place and the two already fraying ones tend to be left in situ, i.e the plug/charger and yet have still failed. It's a shit design, looks come before functionality, unless of course if the function is the need to be expensively replaced on a regular basis.

[jjj]

In the Interwebs I'd rather use (for example) Twitter for one thing, Instagram for another, Vimeo for yet another rather than an all-encompassing blob like Facebook. But that's me...I'm a fan of the modular approach to most things.

But then I just think of FB as a different [and very useful] tool to all the other ones you mentioned.

The film industry already has an appropriate form of crowdsourcing: ticket sales.

Which sadly is only really a reflection on how well films are marketed, not how good they are. It's all about first weekend gross nowadays.

[Telecaster]

Which sadly is only really a reflection on how well films are marketed, not how good they are. It's all about first weekend gross nowadays.

That's certainly true to a large degree, but once a film is made it's made. It may not find its audience initially but if it's good it will eventually. Sometimes a film mangled by studio meddling gets a second chance too if there's a good film there to begin with (and all the footage is preserved). Blade Runner comes to mind. And not all films are 1st-weekend-gross dependent. Some directors have built brands for themselves and have audiences who'll reliably turn out for their films. Like Alexander Payne and Nicole Holofcener (two of my favorites).

-Dave-

[Ray]

I'm not looking at avoiding clipping any highlights. I'm looking at the maximum exposure that does not clip any essential highlights.

Exactly! Good point!

I used to frequently bracket exposures when using Canon DSLRs, in order to give myself more options during processing. The choice might be between a correctly exposed background sky, as seen through the gaps in the trees, but with very noisy shadows in the main subject of the composition; or completely blown patches of sky, but much cleaner shadows without banding.

If the subject was fairly static, one might be able to get the best of both worlds through selection/copy & paste, or merging to HDR if one had been using a tripod, or had a very steady hand.

However, I found that Nikon offered some additional options for this type of process, such as bracketing ISO, which allows one to maintain the desired shutter speed and aperture for the circumstances, and an AF-ON button which completely separates focus from exposure whilst in full manual mode, allowing one to lock focus on one part of the composition whilst one determines an appropriate exposure for another part of the composition, such as the sky, by moving the camera or focusing square to that other part of the scene, then recomposing.

The other advantage of Nikon DSLRs is the wider DR at base ISO, which makes it less of a priority to fuss around trying to get the perfect ETTR.

Nevertheless, when I compare shots taken with my 24mp Nikon D7100 with the same scenes shot with my D800E using appropriately longer focal lengths, the lower shadow noise and smoother mid-tones of the D800E are very obvious. There is great scope for improvement here.

I want a DX camera, or Canon EOS, with the performance of a D800E in terms of resolution, DR and SNR. Are you listening, Nikon and Canon? 

[bcooter]

I'm equipment agnostic and not to take this to the computer brand wars, but sometimes I think Apple is just arrogant.

I have twenty apple computers and two new Imacs, was going to buy the new macpro for an editing station, but it doesn't have pci slots, thunderbolt 2 doesn't allow for external pci boxes with 16 lane access, (which kills of my RED Rockets for debayering footage and for most use, except transferring data isn't not really any faster than the top of line Imac which also doesn't allow 16 lane pci cards.

Since I move studio to studio, workstation to workstation, the newer macs are a problem with thunderbolt and I'm going to have to start working in usb 3 to make my drives backward compatible to older machines.

With apple killing off fcp 7, going to the totally different mindset of fcp X and then limiting the macpro to our way or the highway business model is crazy.

Had I known this was coming I'd have moved our workstations to windows machines a long time ago. 

So I guess this relates to new cameras.   If you buy a new camera and your on the apple system you usually have to go to a new operating system, then a new computer.

I wish they had gotten away from the proprietary file system a long time ago.   

The beauty of my Leica is it connects with usb, it shoots a dng file that will run in computers and software that is 4 years old.

IMO

BC

[dturina]

I don't think anything useful could be accomplished by including ETTR as a metering mode, for several reasons. First, it's an advanced technique, used by a fraction of the ones who shoot raw, who represent a fraction of the market. Second, people who use advanced techniques prefer to have a hands-on approach and judge the histogram personally. So basically it would be similar to expecting macro photographers to use a "flower" scene mode. It'll never happen.

Also, I must admit that I tried using ETTR and gave it up, and it wasn't even a difficult decision because the digital cameras have such a vast tonal reservoir in the shadows, and are so fragile in the highlights, so sensitive to clipping, it only creates problems. Of course, a decade ago when digital cameras usually had very small and noisy sensors it made sense to avoid the shadows as much as possible, but with anything close to a modern camera with a large sensor you'll never see a problem ETTR is meant to solve. Not at low ISO, and if you raise ISO in order to push exposure planning to pull in processing, you'll be better off lowering the ISO and exposure in the first place.

To illustrate what I mean by exposing dark (Canon 5d):

[jjj]

That's certainly true to a large degree, but once a film is made it's made. It may not find its audience initially but if it's good it will eventually.

Except films tend to disappear from cinemas pretty quickly if they don't hit pay dirt on first weekend. Sleeper hits do not happen in cinema anymore and is why hollywood stuff is tending towards summer blockbuster/spectacle type films these days all year around and trying to maximise first weekend grosses. Destroying entire cities/half the planet seems to be the norm now. Though Feb/March tends to be when 'better' quality films are briefly around as all the more serious Oscar hopefuls get released around then.
The cinema I go to, when it was owned by the French Cinema company UCI, used to show a lot of interesting world cinema too. Sadly since it was taken over by Cineworld it now shows mainstream Hollywood + Bollywood films with very rare excursions into anything else.  

Sometimes a film mangled by studio meddling gets a second chance too if there's a good film there to begin with (and all the footage is preserved). Blade Runner comes to mind.

Which was made 30 years ago, is a very, very rare exception and only had a second chance because those involved were still very big names years later and the studios wanted to cash in and make up for the poor box office first time around. Still didn't really make any money and even the various Director's cuts were not how film was originally intended. It's worth reading Paul M. Sammon's book 'Future Noir, The making of Blade Runner' on the travails of making the film and the later failed attempts to do a definitive version.

And not all films are 1st-weekend-gross dependent. Some directors have built brands for themselves and have audiences who'll reliably turn out for their films. Like Alexander Payne and Nicole Holofcener (two of my favorites).

Yet Nebraska quickly disappeared from cinema before I got to see it. It got a single extra screening run due to it doing so well during award season and luckily I was free at tea time on that one day.
The British art-house director Peter Greenaway, used to get funding for his films relatively easily, despite their somewhat esoteric nature. This was because he made them cheaply [by film standards] and they always grossed enough to pay back costs - he was particularly popular in Holland I believe.

[jjj]

I'm equipment agnostic and not to take this to the computer brand wars, but sometimes I think Apple is just arrogant.

I have twenty apple computers and two new Imacs, was going to buy the new macpro for an editing station, but it doesn't have pci slots, thunderbolt 2 doesn't allow for external pci boxes with 16 lane access, (which kills of my RED Rockets for debayering footage and for most use, except transferring data isn't not really any faster than the top of line Imac which also doesn't allow 16 lane pci cards.

Since I move studio to studio, workstation to workstation, the newer macs are a problem with thunderbolt and I'm going to have to start working in usb 3 to make my drives backward compatible to older machines.

With apple killing off fcp 7, going to the totally different mindset of fcp X and then limiting the macpro to our way or the highway business model is crazy.

Had I known this was coming I'd have moved our workstations to windows machines a long time ago. 

The claim that the new MP is for working pros is a bit of a joke as a smaller box with less options is not what many pros want or need. The internals are now external, so the tiny new machine is no smaller once you add the previously internal content back with added spaghetti. Plus connectivity and customisation options have been reduced and price dramatically increased over previous version. At end of the day it's more like a really powerful Mac Mini, aimed more at consumers with deep pockets than working professionals.
This neat but sadly non-existent idea kind of points up the poor form factor.

[jjj]

It's extraordinarily rare that a customer is going to come up with a game-changing idea for you, it's all incremental stuff, usually very tactical, like Michael's idea. "Here's a thing you could do that would solve this problem I have" roughly. Often what's much more interesting is the description of the problem, rather than the proposed solution.

Yup it's extremely unlikely that millions of end users who actually use the product in real world conditions, can think of something that a 2 or 3 people sitting in a design studio won't have already come up with. [/sarcasm]
The problem with too many products is that they are made by [usually very talented] designers, where is too often a big disconnect between someone who is a good designer and an end user. A designer by nature of his job is very unlikely to be say a professional user of his end products so may not grok the reality of real world usage of say a pro camera, even if they say take photographs themselves.
Take for example this cycling jacket aimed at commuters. It has a small led placed within back of jacket to let other road users see the cyclist. A pretty good idea, except there's one big problem. If commuting/using a bike around town a backpack is often used, which will then cover light as it is much higher up jacket than it needs to be and make it stick painfully into your spine.  
If recreational cycling, usually done without a backpack, then the light's position is just fine. But you'll probably use a much snugger, less baggy top for that sort of riding and not a commuter style jacket. Light should have been placed below waist level to clear backpacks, where it will still do it's job nicely and also be out of the way, should you crash and land on your back.

[Manoli]

I have twenty apple computers and two new Imacs, was going to buy the new macpro for an editing station, but it doesn't have pci slots, thunderbolt 2 doesn't allow for external pci boxes with 16 lane access, (which kills of my RED Rockets for debayering footage and for most use, except transferring data isn't not really any faster than the top of line Imac which also doesn't allow 16 lane pci cards.

Not that I have any idea of motion but are these alternatives ?

http://www8.hp.com/us/en/campaigns/workstations/z820.html
http://blogs.windows.com/windows/b/extremewindows/archive/2012/11/13/video-editing-supercomputer-z820-red-edition.aspx

" Windows offers an advantage in this context because you can stack an arbitrary number of RED ROCKET cards (based on how many PCI slots are available) which is not possible on other platforms. Since it takes a dedicated RED ROCKET card per layer of 4K/5K video being transcoded or decoded, that means that Windows offers the most powerful capability for viewing and editing multiple layers of 4K/5K footage on this kind of workstation. "

[amolitor]

Yup it's extremely unlikely that millions of end users who actually use the product in real world conditions, can think of something that a 2 or 3 people sitting in a design studio won't have already come up with. [/sarcasm]

With all due respect, it you're going to quote me, please make some effort to understand what I am actually saying first before getting all snotty.

Thanks.

[bjanes]

I don't think anything useful could be accomplished by including ETTR as a metering mode, for several reasons. First, it's an advanced technique, used by a fraction of the ones who shoot raw, who represent a fraction of the market. Second, people who use advanced techniques prefer to have a hands-on approach and judge the histogram personally. So basically it would be similar to expecting macro photographers to use a "flower" scene mode. It'll never happen.

Also, I must admit that I tried using ETTR and gave it up, and it wasn't even a difficult decision because the digital cameras have such a vast tonal reservoir in the shadows, and are so fragile in the highlights, so sensitive to clipping, it only creates problems. Of course, a decade ago when digital cameras usually had very small and noisy sensors it made sense to avoid the shadows as much as possible, but with anything close to a modern camera with a large sensor you'll never see a problem ETTR is meant to solve. Not at low ISO, and if you raise ISO in order to push exposure planning to pull in processing, you'll be better off lowering the ISO and exposure in the first place.

To illustrate what I mean by exposing dark (Canon 5d):

Your downsized image looks OK, but that does not prove your point since downsizing reduces noise. I downloaded your image and it is 1024x683 pixels. The 5D resolution is 5920x3950 pixels, so the ratio is 5.78:1 by linear image size and 33.4:1 by area. The signal:noise due to photon statistics (shot noise) varies as the square root of the number of photons collected. Your downsizing is effectively pixel binning and you are cramming the number of electrons into a space 33.4 times smaller, improving the SNR by a factor of sqrt(33.4) or a factor of 5.8. Why don't you post a full size image and then we could see the real SNR.

Regards,

Bill

[BJL]

Some recent discussion of ETTR sounds like bracketing exposures, choosing the best frame, and if necessary adjusting its levels.  You can get a three or five frame bracket in a fraction of a second, far quicker that examining in-camera histograms (and worrying about their unreliability), and hundreds of such bracketing groups ft on on a single card (and cost a fraction of a cent to store on hard drive even if for some reason you keep them all).

So I am not sure that I see much value to adding a special ETTR light-metering mode for this.

[dturina]

Your downsized image looks OK, but that does not prove your point since downsizing reduces noise.

The full-size image looks more-less the same as this, there is no noise. I made a completely smooth B2 sized exhibition print out of the file. I actually can't figure out why you would expect there to be noise, it's ISO 100 on Canon 5d, it's not made with a smartphone.

If you want to count photons, I'm not giving that one away, but I do have one technically similar for those purposes - exported from LR 4:
https://drive.google.com/file/d/0B_V_KsUHLnURUi1ZWG8yQnE4enc/edit?usp=sharing

There are many things wrong with it, but noise and shadow tonality aren't among them.

[Telecaster]

Except films tend to disappear from cinemas pretty quickly if they don't hit pay dirt on first weekend. Sleeper hits do not happen in cinema anymore and is why hollywood stuff is tending towards summer blockbuster/spectacle type films these days all year around and trying to maximise first weekend grosses.

All I can say is: DVD, Blu-ray, On Demand, streaming. The Stunt Man, a film that did nothing at the box office, is available in a really nice Blu-ray release with in-depth supplemental material. Cimino's notorious Heaven's Gate (I'm a fan) is out via Criterion Blu-ray with a new director's cut, again with good supplemental stuff. Fassbinder's World On A Wire, a 3 1/2 hour mind-f**k extravaganza shown once on German TV in the 1970s, is also out via Criterion Blu-ray. I missed Nebraska in the theater too but watched the Blu-ray on its release day. Once a film is made it has a shot at surviving, and if it survives it has a shot at being seen. It's not a perfect scenario by any means but at least good stuff still gets made.

-Dave-

[bjanes]

The full-size image looks more-less the same as this, there is no noise. I made a completely smooth B2 sized exhibition print out of the file. I actually can't figure out why you would expect there to be noise, it's ISO 100 on Canon 5d, it's not made with a smartphone.

If you want to count photons, I'm not giving that one away, but I do have one technically similar for those purposes - exported from LR 4:
https://drive.google.com/file/d/0B_V_KsUHLnURUi1ZWG8yQnE4enc/edit?usp=sharing

There are many things wrong with it, but noise and shadow tonality aren't among them.

Thanks for the upload. I did download and look at it and does look very good. Of course, it does have some minimal noise visible in the smooth out of focus areas in the upper left portion of the image--due to photon statistics noise is inevitably present. The DR of your image is not that high and ETTR was not necessary to get good results. However, in a high dynamic range situation ETTR is still necessary to obtain the full DR of the sensor. I do agree that ETTR is less critical with modern full frame sensors, but it still has a place even at base ISO.

Regards,

Bill

[hjulenissen]

So I am not sure that I see much value to adding a special ETTR light-metering mode for this.

Bracketing is a band-aid to many photography troubles. I would prefer accurate, predictable tools that would aid in setting automatic/semi-automatic exposure rather than "spraying and praying".

-h

[Hans Kruse]

I don't think anything useful could be accomplished by including ETTR as a metering mode, for several reasons. First, it's an advanced technique, used by a fraction of the ones who shoot raw, who represent a fraction of the market. Second, people who use advanced techniques prefer to have a hands-on approach and judge the histogram personally. So basically it would be similar to expecting macro photographers to use a "flower" scene mode. It'll never happen.

Also, I must admit that I tried using ETTR and gave it up, and it wasn't even a difficult decision because the digital cameras have such a vast tonal reservoir in the shadows, and are so fragile in the highlights, so sensitive to clipping, it only creates problems. Of course, a decade ago when digital cameras usually had very small and noisy sensors it made sense to avoid the shadows as much as possible, but with anything close to a modern camera with a large sensor you'll never see a problem ETTR is meant to solve. Not at low ISO, and if you raise ISO in order to push exposure planning to pull in processing, you'll be better off lowering the ISO and exposure in the first place.

To illustrate what I mean by exposing dark (Canon 5d):

I think you missed my point: The way I shoot as described previously is to have to look minimally on the LCD when shooting and come back with optimal or very close to optimal results in terms of exposure. In Lightroom I will choose the best exposure and I can tell you that I get much overall results by using this method than chimping which I did in the past. I still have shots from the old 5D than are suboptimal and I wished that I had used the method I use now. Also using the 1Ds mkIII and D800E I can say the same although for the D800E it is mostly to avoid highlight clipping. I now use mainly 5D mkIII and although many shots are just perfect with no bracketing and EC, there are plenty where this is not the case. I find chimping severely distracting in my creative process and actually unnecessary.

You image does not prove anything as you kept shadows really dark. That can be fine in some scenes but in others it does not work, at least for me

[Hans Kruse]

Exactly! Good point!

I used to frequently bracket exposures when using Canon DSLRs, in order to give myself more options during processing. The choice might be between a correctly exposed background sky, as seen through the gaps in the trees, but with very noisy shadows in the main subject of the composition; or completely blown patches of sky, but much cleaner shadows without banding.

If the subject was fairly static, one might be able to get the best of both worlds through selection/copy & paste, or merging to HDR if one had been using a tripod, or had a very steady hand.

However, I found that Nikon offered some additional options for this type of process, such as bracketing ISO, which allows one to maintain the desired shutter speed and aperture for the circumstances, and an AF-ON button which completely separates focus from exposure whilst in full manual mode, allowing one to lock focus on one part of the composition whilst one determines an appropriate exposure for another part of the composition, such as the sky, by moving the camera or focusing square to that other part of the scene, then recomposing.

The other advantage of Nikon DSLRs is the wider DR at base ISO, which makes it less of a priority to fuss around trying to get the perfect ETTR.

Nevertheless, when I compare shots taken with my 24mp Nikon D7100 with the same scenes shot with my D800E using appropriately longer focal lengths, the lower shadow noise and smoother mid-tones of the D800E are very obvious. There is great scope for improvement here.

I want a DX camera, or Canon EOS, with the performance of a D800E in terms of resolution, DR and SNR. Are you listening, Nikon and Canon? 

You must have used a very old Canon like e.g. the 5D. All Canons since ca. 2007 has an AF-ON button, before that folks remapped the *-button for AF. I always used AF separate from the shutter.

The issue with the D800E is more to avoid highlight clipping than the consequence of underexposure although noise will creep in in such cases and sometimes in distracting ways.

If I was to think of a sensor for my liking, it would be something like:

35mm full frame, 50MP resolution, each pixel consisting of 2 pixels for PDAF and one for exposure time. No mechanical shutter. Exposure start by resetting all pixels, end of exposure would happen when either the electronic shutter ends the exposure or for individual pixels the well is full and the exposure time pixel will record the exposure time which could be shorter than the overall exposure time. Then the RAW format would contain the light levels for each pixel and exposure time and the RAW converter could convert this into a final picture. This could be Bayer type sensor or eventually full color at each pixel. Such a sensor could almost infinite DR if the overall shutter time was long enough. Such a sensor would never clip highlights. Unless the shutter time was chosen too short there also would be automatic ETTR.

Well, well I'm probably dreaming.

Besides that I would like a very compact body and a single lens 14-300mm f/4 fully corrected in HW (in the lens and not in SW) and a resolution to match the sensor. The lens should be fixed on the camera to avoid issues of dust etc. It should be very compact and light and made of light materials. LCD should be touch screen and with automatic DOF and also manual DOF by pointing on the screen from where to where the DOF should exend to. Here assuming traditional technique as we know from today of how lenses work wrt. DOF.

That would be my dream landscape camera. It could possibly weigh just one KG or less.

[dturina]

I think you missed my point: The way I shoot as described previously is to have to look minimally on the LCD when shooting and come back with optimal or very close to optimal results in terms of exposure.

Well, I must admit I don't spend much time fussing with exposure in the field; I take a look to see whether I clipped something and usually make minor exposure adjustments in raw conversion, but other than that, I have a "knack" for nailing exposure from years of shooting chromes, which are quite unforgiving in both shades and highlights. Compared to that, digital is so easy to get good results with I'm just not motivated to geek it out.
Also, most of the stuff I shoot doesn't lend itself well to ETTR principle; if I shoot handheld in variable lighting conditions, it's difficult to nail it as it is, but if I add another variable it spirals out of control. For instance, this:


Go even the tiniest bit to the right and it's ruined. It's so vital to guard the highlights, nothing else matters, especially with a modern large-sensor camera that renders blacks so velvety smooth.