The next crop of cameras is coming down the chute

[ErikKaffehr]

Hi,

It may just be that the sensor manufacturers do know what they are doing?

Best regards
Erik

shame nobody thought to tell the sensor manufacturers.
all the while the design and manufacture of sensors is not in the hands of the camera designers photographers will have to put up camera's that are less then ideal.

maybe canon,fuji or pentax will change this.

[eronald]

Hi,

It may just be that the sensor manufacturers do know what they are doing?

Best regards
Erik

I'm sure the ones at Sony and Nikon know, and also the guys at Dalsa and Kodak who do the military sensor chips. I would be very surprised if either Kodak or Dalsa can be bothered to put a top engineering team to work on MF camera sensors and allow them to generate new designs FROM THE CELL UP; rather I think they have some team player repackage the leftovers using bits and pieces from other projects.

I mean, I don't wanna be rude, but the top guys get sent where the money is, and MF is not exactly a big application these days - the whole sensor market for MF is probably worth less than $ 5 Million a year, which translates into about a million dollars R&D, not a lot if you need to pay a ten-person team in Europe or the US.

For a new sensor design, I'd expect you'd need two cell designers, two IC layout integration guys to handle clocking, layout, pin driver amps, global control circuitry, all the stuff you need to set up the array and make it talk to the world,  some guy to do the color filters and microlenses and cover glass, a couple of people to make the dev electronics for physical prototypes and test gear, and of course some support infrastructure and technicians.  

A new cell design would probably be quite slow, because the geometry is crucial, you need a process simulator at geometric level that models the physics precisely, simulations would need validation by baking a prototype. So I don't think one could go from zero to full chip in less than a year.

Of course, if you employ existing cells and existing control schemes and peripheral circuirty, then you can probably do an initial design of a sensor chip in a couple of weeks, and finish the product in 3 or 4 months with little innovation but also few surprises. Unless a company is feeling philantropic it will probably choose this route, rolling off an MF chip when all the components have already been generated for a project that is heavily funded for R&D.

Edmund

[fredjeang]

Can someone tells me then, how Red One, wich is a zillion time a smaller company than Sony, Canon and co is the only one to currently produce an affordable raw video (there are others but more expensive), with a great or at least a good design, pro featured, top imagery etc etc...

Because if a bunch of folks at Red One are capable of doing what their are doing, there is then a simple conclusion: big manufacturers are fooling us.

[Fritzer]

Can someone tells me then, how Red One, wich is a zillion time a smaller company than Sony, Canon and co is the only one to currently produce an affordable raw video (there are others but more expensive), with a great or at least a good design, pro featured, top imagery etc etc...

Because if a bunch of folks at Red One are capable of doing what their are doing, there is then a simple conclusion: big manufacturers are fooling us.

Hold on a second : do you own a RED ?
How much is it for a complete system ?
It's easy to talk the talk when you don't have to put up your own money.

As it was mentioned above, the Canon 5DII has somewhat revolutionized access to budget film, if not on purpose .

As for the MP race : I think it's alive and well, and rightly so .
I for one take high resolution at base ISO over high ISO quality any day; not that it's mutually exclusive, sensor design usually improves in all areas .
Each his own, but I actually have clients ask for more when I shoot 30MP - only magazines, though, who would have thought ...

I can't wait for the NEX-7; MFT I've never owned, but I just can't believe a GH2 will be even close to any NEX in terms of file quality, and my NEX 5 is about as low as I dare go.

[TMARK]

Fritzer said:

"Hold on a second : do you own a RED ?
How much is it for a complete system ?"

Yes, and about $80k, which includes lenses, Arri mattboxes, monitors, storage, supports blah blah blah.  That is set up for commercials, high end production.  You could probably get away with $30k for a bare bones kit.

[eronald]

Fred,

- Obviously Red exists (Our friend the Cooter says they sold at least half a billion dollars of gear to date, who am I to disagree). So your "if" question become a "why" question. RED were fast and the incumbents seem hobbled. I think RED gained 3-4 years just because of how they started.

- Red got underway fast because it was launched and funded by a billionaire, who recruited some top talent, and identified both the niche of digital pro film-making, and the enabling technology of recording wavelet compressed Raw.  Physically they needed to design a box with a hole. So they had none of the startup process inertia which a japanese firm has in setting up reviews for a totally new product.

- Once RED were set up they came fast out of the gate on development because they could pay cash for innovative R&D (the Mysterium sensor and the wavelet compression codecs) and camera prototypes. As I pointed out in a post above, an entirely new sensor design only costs around a million or two dollars, but  what you get is only as good as the guys you put on the project.

- The question of why big camera companies are not making similar products is a business case study. It's obvious that Sony and Canon and Nikon could make a RED lookalike in a year. I believe that camera company management everywhere are protecting the cash cows of existing TV camera ranges. Kodak couldn't move as a company from film to digital although thy invented a lot of digital imaging technology. IBM was incapable of capitalising on the PC technology which it launched.

Can someone tells me then, how Red One, wich is a zillion time a smaller company than Sony, Canon and co is the only one to currently produce an affordable raw video (there are others but more expensive), with a great or at least a good design, pro featured, top imagery etc etc...

Because if a bunch of folks at Red One are capable of doing what their are doing, there is then a simple conclusion: big manufacturers are fooling us.

[fredjeang]

The cost of a Red system is (without the robocoperies), on line with a high-end medium format system, wich considering the target and investment-return is to my eyes what I call "cheap", or at least reasonable.

We are talking in Spain about those numbers:

the MX is sold at 20.000 €

the Epic M complete: 58.000 €

Epic X brain: 35.000 €

Rear and slide modules: 1.500

the remote: 750

etc...


a 40 MPx Phase back: 17.000 €

the big boys backs are more than 40.000 € (don't remember the exact number but it's somewhere there). Tell me about that!

and all that without the lenses also, without body, without anything.

In that context Red is cheap (or MFD are expensive) depends the point of view.

But, back on the robocoperies, that's also true and specially true with any dslr like the Canons: matte-boxes, microphones, monitors, storage, supports etc...

If you get that new Canon, it will be 7000 just the body, no raw video and obliged to get all the accessory saga we where talking about, plus, what will be that new codec?  Will Canon writte fast an AMA script so we can link native? that remains to be seen. Or are we going to be obliged to chase the version point___ of Premiere Pro or Avid, or even worse, a third-party little application created in a hurry that would do the job. Like if we didn't already have enough mess. The 12 fps I.S is already been done by the prosummer Sony and I'd like to see how many seconds it will handle that, and how many seconds it freezes the buffer when it's full. Then...  to be fair with Canon, the video bitrate announced is really good. The low-light will be very good, as expected, and I don't giive a damn if it's a 18MP or 25MP reso, it's more or less the same story really. The old Leica R 10MP digiback could still smokes some bigger sensors in print to date. Those reso are mostly marketing claims.

About the GH2 vs NEX _ ? ... I've tried the Nex, it clips way more the highlights than the Pana that already clip a lot, to me (maybe I didn't get well the user manual) and I'm absolutly unable to see the differences in footage in a blind test, except at higher isos where I found the sony a little more workable. I could work with both of those cameras. But I've already seen a GH2 on a big set manipulated by experienced videographers, never saw any NEX, so the Panasonic is probably not that bad...IMHO.

About the clients that ask for more than 30 MP ? Yes, some of you might experience that, but in my case, all I hear since about a year or 2 is "make it smaller". (even for printing output).

About Red, I don't own one at the moment (it's in my plans), but regularly edit R3D and it's one of the most friendly workflow I've ever seen. At least to my way of understanding what a good workflow should be. Raw video is great indeed. (without talking about the performance and qualities and flexibility that those deliver).

Correct me if I'm wrong, but it seems that R3D will be like a PS format, readable between generations, no? Because tomorrow will have an AVCHD # 4, 5 or 6...that nothing will read, as always.
And even in the case there where issues in the future, RCX will read-writte those for free on Mac or Windows.

I have 2 set-ups with the red workflow, both are rock solid and free of platform issues (wich is not the case in the Alexa because of ProRes). I use one kind of workflow using Avid Media Composer via AMA, and another one using Red-Cine X and Edius 6 in a dpx workflow. Then, send all that into Nuke. Absolutly no hassles.

[BernardLanguillier]

shame nobody thought to tell the sensor manufacturers.
all the while the design and manufacture of sensors is not in the hands of the camera designers photographers will have to put up camera's that are less then ideal.

maybe canon,fuji or pentax will change this.

I guess the Nikon D3s never existed... :-)

Cheers,
Bernard

[eronald]

Made by Nikon for themselves, and not by an MF company.
Not one of the bought-in leftovers which are fed to us by the MF houses.
I am sure that Dalsa and Kodak do very good reconnaissance sensors, though.
Has anyone noticed that Kodak had 35mm full frame before anyone else, back when they were in the photo business?

Edmund

I guess the Nikon D3s never existed... :-)

Cheers,
Bernard

[BernardLanguillier]

Made by Nikon for themselves, and not by an MF company.
Not one of the bought-in leftovers which are fed to us by the MF houses.
I am sure that Dalsa and Kodak do very good reconnaissance sensors, though.
Has anyone noticed that Kodak had 35mm full frame before anyone else, back when they were in the photo business?

I see.

I tend to be more and more application focused these days and had forgotten this was an MF thread. The artificial siloization of photography based on sensor size makes less and less sense to me.

But yes, I guess that this is still an MF silo and I should have guessed that this comment was about Kodak and Dalsa.

Perhaps Michael should just get rid of this and instead create a high image quality silo? When you think about it, the very structure of this forum is designed to support the idea that MF cameras really have a specificity to them, photography itself is less the focus that gear used to create it... per design.

Cheers,
Bernard

[eronald]

I see.

I tend to be more and more application focused these days and had forgotten this was an MF thread. The artificial siloization of photography based on sensor size makes less and less sense to me.

But yes, I guess that this is still an MF silo and I should have guessed that this comment was about Kodak and Dalsa.

Perhaps Michael should just get rid of this and instead create a high image quality silo? When you think about it, the very structure of this forum is designed to support the idea that MF cameras really have a specificity to them, photography itself is less the focus that gear used to create it... per design.

Cheers,
Bernard

Well, in a way the concentration on gear frees us to talk about the other stuff by allusion.

Let's face it: eg. in fashion, there are photographers who are good at selling to the customer; there are photographers who are good at getting the project done, no hassle, without direction; there are those who are very good shooting exactly what the client wants, without ever intruding into the client's work process; there are those who are remarkable at imagining evocative settings, where there is nothing in the picture, but still a mood; there are those who are good at finding locations that are full of little details, and lighting them; there are those who are really good at working the model and getting an expression; there are those who the best models love to work with and the name of the model makes the shot; there are those who know how to make the clothes look more beautiful than the woman. There are many sorts of fashion photographers, but they all use cameras - and talking about cameras gives them an excuse to talk. And we haven't even tried to discuss landscape yet

Edmund

[Sareesh Sudhakaran]

RED RAW is not the same thing as photography RAW files, not even close. They just call them 'RAW' (who can stop them?) but it's already compressed data: http://en.wikipedia.org/wiki/Red_Digital_Cinema_Camera_Company#Compression_and_workflow

The closest codec to a still-camera RAW file is the new ARRIRAW (on the Alexa). It's not cheap. Red is not cheap either. But that's another story.

A gang of bright engineers with access to money can buy sensors from Kodak and build a MFDB, or even a 2K video camera. It's one thing to build it, but another to wade through the murky cesspool of patents, copyrights, licenses, legal fees, marketing, sales and customer service to form a profitable company that caters to professionals with the highest demands in quality (or at least they think they do!).

One of the strangest things I've come across is the fact that most moviegoers have been watching films in cinemas (in 30-100 feet screens) for a 100 years at 1K resolution (or less). Nobody has ever complained of the lack of sharpness or the low resolution. Most images are viewed on newspapers, billboards or the internet, where resolution, color and detail are at their lowest.

This is the market Sony, Kodak and Canon cater to, I suppose. What the niche must realize (which I'm sure they do) is that if they can buy a 50MP sensor from Kodak for $3,500 it's because Kodak sells millions of smaller chips to mass camera manufacturers. I can't blame anybody for it, except myself maybe for always wanting more.

[eronald]

At the rate we're going, startup innovation is going to be a thing of the past.

I don't think it's possible to market *any* piece of software or hardware in the US anymore without infringing on some patent. Whether that patent would stand up to scrutiny or be invalidated is a moot point, because you couldn't afford the legal fees anyway, unless you happen to have big pockets like Red's founder.

Edmund

A gang of bright engineers with access to money can buy sensors from Kodak and build a MFDB, or even a 2K video camera. It's one thing to build it, but another to wade through the murky cesspool of patents, copyrights, licenses, legal fees, marketing, sales and customer service to form a profitable company that caters to professionals with the highest demands in quality (or at least they think they do!).

[fredjeang]

RED RAW is not the same thing as photography RAW files, not even close. They just call them 'RAW' (who can stop them?) but it's already compressed data: http://en.wikipedia.org/wiki/Red_Digital_Cinema_Camera_Company#Compression_and_workflow

The closest codec to a still-camera RAW file is the new ARRIRAW (on the Alexa). It's not cheap. Red is not cheap either. But that's another story.

A gang of bright engineers with access to money can buy sensors from Kodak and build a MFDB, or even a 2K video camera. It's one thing to build it, but another to wade through the murky cesspool of patents, copyrights, licenses, legal fees, marketing, sales and customer service to form a profitable company that caters to professionals with the highest demands in quality (or at least they think they do!).

One of the strangest things I've come across is the fact that most moviegoers have been watching films in cinemas (in 30-100 feet screens) for a 100 years at 1K resolution (or less). Nobody has ever complained of the lack of sharpness or the low resolution. Most images are viewed on newspapers, billboards or the internet, where resolution, color and detail are at their lowest.

This is the market Sony, Kodak and Canon cater to, I suppose. What the niche must realize (which I'm sure they do) is that if they can buy a 50MP sensor from Kodak for $3,500 it's because Kodak sells millions of smaller chips to mass camera manufacturers. I can't blame anybody for it, except myself maybe for always wanting more.

That's correct, Red Raw is in fact already compressed, slightly compressed. Should it be compare to Raw still? Yes to a large extend, no in the absolute. The workflow is very similar to any raw dev and when you correct, the footage stands up, and more importantly, it's way easier to get the same look (wb, expo etc...) when multiple cameras are involved, because that is a real mess. There are not 2 canons who shoot the same! Always differences.

Arriraw is almost not used. I don't know one Alexa prod that shoots Arriraw, very few do because they are so happy with the 444 prores that they don't consider the extra cost a necessity, and, as you wisedly point, we are watching movies for decades in a lower reso and nobody complains. Totally agree.

In fact, and I know some differs on that, the only good point of having a lot of reso is when you downsample, it looks really great. For ex, a full HD downsampled to 720 looks generally better than a 720 native.

I'm aware that this statement is very discussed, but I find the differences visible on screen.

[BernardLanguillier]

At the rate we're going, startup innovation is going to be a thing of the past.

I don't think it's possible to market *any* piece of software or hardware in the US anymore without infringing on some patent. Whether that patent would stand up to scrutiny or be invalidated is a moot point, because you couldn't afford the legal fees anyway, unless you happen to have big pockets like Red's founder.

Very true, IP protection has been turned around and is now another entry barrier created by multi-billion US$ corporations.

I do understand the willingness to protect the IP that genuinely created for a really product developement purpose and as a means to protect they creations from competitors with low morale, but the massive buy over of IP we are seeing today and the cost of litigation is all too favorable to existing businesses and does clearly endanger genuine innovation.

Cheers,
Bernard

[Radu Arama]

I'm sure the ones at Sony and Nikon know, and also the guys at Dalsa and Kodak who do the military sensor chips. I would be very surprised if either Kodak or Dalsa can be bothered to put a top engineering team to work on MF camera sensors and allow them to generate new designs FROM THE CELL UP; rather I think they have some team player repackage the leftovers using bits and pieces from other projects.

I mean, I don't wanna be rude, but the top guys get sent where the money is, and MF is not exactly a big application these days - the whole sensor market for MF is probably worth less than $ 5 Million a year, which translates into about a million dollars R&D, not a lot if you need to pay a ten-person team in Europe or the US.

For a new sensor design, I'd expect you'd need two cell designers, two IC layout integration guys to handle clocking, layout, pin driver amps, global control circuitry, all the stuff you need to set up the array and make it talk to the world,  some guy to do the color filters and microlenses and cover glass, a couple of people to make the dev electronics for physical prototypes and test gear, and of course some support infrastructure and technicians.  

A new cell design would probably be quite slow, because the geometry is crucial, you need a process simulator at geometric level that models the physics precisely, simulations would need validation by baking a prototype. So I don't think one could go from zero to full chip in less than a year.

Of course, if you employ existing cells and existing control schemes and peripheral circuirty, then you can probably do an initial design of a sensor chip in a couple of weeks, and finish the product in 3 or 4 months with little innovation but also few surprises. Unless a company is feeling philantropic it will probably choose this route, rolling off an MF chip when all the components have already been generated for a project that is heavily funded for R&D.

Edmund

Your assessment is most likely correct (I am not qualify to judge and you are so I trust your opinion) but this was valid before Pentax got into game. Don't forget that Pentax makes the 645D from January 2010 in at very least 500 units per month (with strong suggestions that they upped the production to 700 after they failed to meet demand in Japan last year). Even more important those at least 6000 sensors per year are identical rather than various dimension and cell size (and even different manufacturers in Hasselblad case). Until 645D will be discontinued is possible they would of bought close to 15K sensors from Kodak.

IMO Pentax with its Japanese strong connections is in pole position to order two CMOS sensors (most likely to Sony) based on the same cell size and with 33x44 mm and close to full 645 size (since they made even the huge 25 mm for the whole 645 film size circle). In the 2012 they will already have 3 new digital lenses on the market (55, 25, the new portrait one), tethering application and lots of FA 645 lenses in production. I see them passing 1000 units per month with two models that means 24K sensors in two year's time. Even at a modest combined price of 700 USD per sensor (taking into consideration that naturally the larger size will be made in smaller numbers) I think a 16.8M USD deal is nothing to be ignored by Sony or other Japanese CMOS company.

My money is on a 6 micron cell with a 40 Mp 33x44 mm variant and a close to 80 Mp 5x(x)4x mm Sony made sensor and if not at CP+ early 2012 I think they will announce it at latest at Photokina in September 2012. It makes sense and Pentax has the money, connections and the will power to be the first MF manufacture with such a sensor.

Radu

[eronald]

Your assessment is most likely correct (I am not qualify to judge and you are so I trust your opinion) but this was valid before Pentax got into game. Don't forget that Pentax makes the 645D from January 2010 in at very least 500 units per month (with strong suggestions that they upped the production to 700 after they failed to meet demand in Japan last year). Even more important those at least 6000 sensors per year are identical rather than various dimension and cell size (and even different manufacturers in Hasselblad case). Until 645D will be discontinued is possible they would of bought close to 15K sensors from Kodak.

IMO Pentax with its Japanese strong connections is in pole position to order two CMOS sensors (most likely to Sony) based on the same cell size and with 33x44 mm and close to full 645 size (since they made even the huge 25 mm for the whole 645 film size circle). In the 2012 they will already have 3 new digital lenses on the market (55, 25, the new portrait one), tethering application and lots of FA 645 lenses in production. I see them passing 1000 units per month with two models that means 24K sensors in two year's time. Even at a modest combined price of 700 USD per sensor (taking into consideration that naturally the larger size will be made in smaller numbers) I think a 16.8M USD deal is nothing to be ignored by Sony or other Japanese CMOS company.

My money is on a 6 micron cell with a 40 Mp 33x44 mm variant and a close to 80 Mp 5x(x)4x mm Sony made sensor and if not at CP+ early 2012 I think they will announce it at latest at Photokina in September 2012. It makes sense and Pentax has the money, connections and the will power to be the first MF manufacture with such a sensor.

Radu

I have no reason to doubt your marketing numbers, and Sony could of course just recycle an existing cell design into a large sensor, thereby giving Pentax what they need. Whether Sony would be happy creating themselves a competitor for their own flagship 35mm full frame is probably the main question to decide whether this happens.

However, Pentax could also commission a sensor from Samsung; I think this is a more probable resolution.

Edmund

[ErikKaffehr]

Hi,

Sony already makes CMOS sensors for the competition, namely Nikon (D3X, D300, D7000) and Pentax K5. Very much possible that customer devices are different from Sony devices, Nikon is said to have tweaks and at least the D3X is much better in some areas then the Sony Alpha 900, which is using pretty much the same sensor.

If Sony would develop a larger sensor is another question, but I guess it is more about economics than policy.

Best regards
Erik

I have no reason to doubt your marketing numbers, and Sony could of course just recycle an existing cell design into a large sensor, thereby giving Pentax what they need. Whether Sony would be happy creating themselves a competitor for their own flagship 35mm full frame is probably the main question to decide whether this happens.

However, Pentax could also commission a sensor from Samsung; I think this is a more probable resolution.

Edmund

[BernardLanguillier]

IMO Pentax with its Japanese strong connections is in pole position to order two CMOS sensors (most likely to Sony) based on the same cell size and with 33x44 mm and close to full 645 size (since they made even the huge 25 mm for the whole 645 film size circle). In the 2012 they will already have 3 new digital lenses on the market (55, 25, the new portrait one), tethering application and lots of FA 645 lenses in production. I see them passing 1000 units per month with two models that means 24K sensors in two year's time. Even at a modest combined price of 700 USD per sensor (taking into consideration that naturally the larger size will be made in smaller numbers) I think a 16.8M USD deal is nothing to be ignored by Sony or other Japanese CMOS company.

Hum... it really depends whether their current equipment enables them to manufacture such large sensors. If they don't, then the cost of machines alone would eat up a significant part of the margin.

I would love this to happen though.

Cheers,
Bernard

[yaya]

Hum... it really depends whether their current equipment enables them to manufacture such large sensors. If they don't, then the cost of machines alone would eat up a significant part of the margin.

I would love this to happen though.

Cheers,
Bernard

Exactly, plus unless they've cracked the low yield issue, the cost per sensor (larger than 24x36) is going to be very high....possibly to the point of not being profitable...and nobody makes camera for charity...