I have an EX1 for a weekend. Anyone want me to run tests?

[tnle2]

Well, with all due respect, I don't think Jim would disagree with me at all. There are obviously compromises involved in the Red design, specifically to meet a price point. [SNIP] Part of the goal of the Red design, from day one, was to make it affordable so that guys like us (DVXUser members) could buy it. That was always part of the design philosophy. Price was always a foremost consideration in the Red design.

I doubt choosing between CCD and CMOS would have changed the price that much. It's inexpensive not because of the choice of sensor, but because there are no moving mechanical parts (like a spinning shutter), no optical viewfinder, and no huge corporate overhead. Time and time again the RED team has said they set out to built the most badass camera they could--a camera capable and worthy enough to shoot the most high-end applications imaginable, like Hollywood movies. Although cost probably was a design consideration at some level, I sincerely doubt they intentionally "crippled" it with a rolling shutter just to hit a price-point that guys like DVXUser members could afford. Doing so would make the camera look bad in those high-end applications like VFX work and anywhere the rolling shutter would cause a problem.

How is this inconsistent with anything? Of course they're going to use this thing anywhere and everywhere they can now, they developed it, they have it, why not use it? Unless the market has a backlash against it, of course.

You said that when faced with a choice, every manufacturer that offers both CCD and CMOS has, as far as you know, always chosen CCD for the high end, and only offered CMOS in the lower-cost end of their product line. So I'm just saying Thomson has chosen CMOS for all of their cameras going forward.

Barry, it seems like the core of your argument is that these companies are shoving us this rolling shutter CMOS business and its limitations because they're inexpensive chips and it's saving them money.

My argument is that it's not necessarily about saving money but there is something else driving this move to CMOS in very high-end imaging applications. Maybe it's the low power consumption, maybe it's image quality, maybe it's the format flexibility, maybe it's something else...but I doubt it's cost because these high-end cameras are hugely expensive.

I honestly don't know what it cost to make a high-quality CMOS sensor but I doubt it's much cheaper than CCD like you keep repeating. Maybe it's cheaper with mass-produced, low resolution, low quality cellphone CMOS sensors, but to get high-quality CMOS sensors that can compete with CCDs, there is probably much more to it. DALSA's own paper seems to agree:

The production cost per unit of processes silicon does not strongly favor one technology over the other (as originally thought). The extensive process engineering and number of fabrication steps to bring CMOS image quality to levels comparable with CCDs require much more expensive wafer processing than was originally projected. Cost is often more strongly influenced by the business economics and competitive motivations of a particular foundry, rather than by the choice of the technology itself.

 

[ullanta]

Folks, keep in mind also that "cost" is relative. Saying "CMOS is cheaper" doesn't mean CMOS is "cheap" or a CMOS chip is cheap for a given application. It means that it tends to be cheaper than CCD for the same application. That is, a 35mm-equivalent sensor is not "cheap" either in CMOS or CCD - but the CMOS may have enough of a cost advantage to make a product feasible that wouldn't be feasible with CCDs. And, it's not just the chip - it's the support hardware, the power/heat management, board complexity, etc.

------

A history question - I have (well, in a basement 3,000 miles away) an OLD consumer camera. It's a JVC, from the era when the VHS deck was slung over your shoulder and the handheld camera attached by cable - early 1980's. I'm pretty sure it was CMOS. Was there not an era when CMOS was more popular for video imagers?

 

[ullanta]

Although cost probably was a design consideration at some level, I sincerely doubt they intentionally "crippled" it with a rolling shutter just to hit a price-point that guys like DVXUser members could afford. Doing so would make the camera look bad in those high-end applications like VFX work and anywhere the rolling shutter would cause a problem.

Argh! The old "intentional crippling"! Stop it!

Cost is always a design consideration. As is sensitivity, noise, heat. EVERY design is a compromise. Not making a device unrealistically perfect is NOT intentional crippling. No camera can be perfect. Especially at reasonable prices. Design trade-offs are not "intentional crippling". It's the opposite - a "heroic" struggle to find a way to make a better product that will be reliable, at a reasonable cost.

There aren't many CMOS cameras out there with global shutters. Clearly some of the trade-offs involved are considered unacceptable at this point. Would you trade 2 or 3 stops of sensitivity for a global shutter? 4 or 5? Would you pay $100,000 rather than $20,000, and would that be worth it? The most "bad ass" camera for indie filmmakers certainly implies cost as a MAJOR factor. A $300,000 camera would not be at all useful for most of the target market.

 

[Barry_Green]

It's inexpensive not because of the choice of sensor, but because there are no moving mechanical parts (like a spinning shutter), no optical viewfinder, and no huge corporate overhead.

I'd say the main thing that's lowered the price is the distribution mechanism. Not having dealers and regional sales companies eliminates two points of markup. But will this only-sold-in-L.A. model "work" for them long-term? We'll see.

I sincerely doubt they intentionally "crippled" it with a rolling shutter just to hit a price-point that guys like DVXUser members could afford.

Your words, not mine. I don't think they "crippled" anything, I think they set out to make the very best system they possibly could.

But here's the thing: at the time the Red was designed, did anyone even know what kind of impact a rolling shutter would have on the footage? Or, do you think that if they had it to do all over again, they'd make the same choice? I doubt it, but only they know for sure.

I still see people asking "what is a rolling shutter?" and certain posters who like to present themselves as all that knowledgeable still think it's related to a little slanting here and there. Skew is, IMO, the least objectionable of the rolling shutter artifacts.

In some ways the EX1 will be a test bed for whether rolling shutter is a problem or not. It's the first mass-market camera that someone might try to do some actual work with. So as more of them hit the user's hands, over the course of the next six to 12 months we should see some results as to whether it's much-ado-about-nothing, or whether it's a real issue.

Nice thing with Red is that it's modular and upgradeable. If the rolling shutter turns out to be a problem, they can design a new sensor and substitute it. Don't know of any other systems that were designed with that capability in mind, and it's something that makes me a little less hesitant about accepting the existing rolling-shutter design.

Doing so would make the camera look bad in those high-end applications like VFX work and anywhere the rolling shutter would cause a problem.

You are very correct on this, and that's what I'm worried about. Unlike a lot of people, I'm supposed to plunk down $50,000 within the next 30 days on one of these systems, and I am quite concerned about it. Unfortunately for me, my reservation's coming up so quickly that the camera really hasn't had a chance to be thoroughly vetted yet, which leaves me in the position of having to jump into an unknown system. Not a comfortable position to be in.

But if you think I'm just gonna say "well, Red's a big company and so there won't be any problem at all", you're nuts. There are issues or there aren't. I have no reason to think the Red won't be affected; in fact I've seen some examples on it already. But if anyone's managed to get a handle on it and control it, it's probably Red. Even Russ Andersson said that he expects Red to have solved the problems. But if they haven't then yes, Red would indeed look bad in high-end apps like VFX work.

So I'm just saying Thomson has chosen CMOS for all of their cameras going forward.

I see what you're saying. I don't agree, of course -- they can say they intend to use it, but let's see if anything hits the fan over the course of the next six to 12 months. GM might have thought they were gonna sell a lot of diesel cars, too, but it didn't turn out that way because the engines melted. If the customers reject rolling shutters, you can bet that Thomson and others will get off that train real quick.

Barry, it seems like the core of your argument is that these companies are shoving us this rolling shutter CMOS business and its limitations because they're inexpensive chips and it's saving them money.

Not really. I mean, yes I do think that unwitting customers are getting shoved something at them that is designed primarily to save the manufacturers money, yes, but that's the main aim of, say, the HV20 or EX1 or other mass-market lower-cost product. I don't think the Red chose it because of that, I think they designed the best thing they could and, at the time, they didn't necessarily even know what impact the rolling shutter would have on 'em. And now that they have seen what happens, they've taken active steps to minimize it as much as they can.

Instead, the core of my argument is that people are praising the CMOS sensor as the second coming of all holiness, when in reality it's just another alternative, one that's got several compromises in it, compromises that do not best serve the high-end market. So we're seeing CMOS rolled out primarily in low or lower-end gear.

Regardless, the crux of my point in even bringing this up, and trying to educate people, is this: these new camcorders do not behave the way the old ones did. There has been a fundamental change. And that change can result in some very unexpected artifacts in your footage, including the ability to ruin shots. If you know about this and willingly choose it, go right ahead. But if you don't know about it and you end up getting screwed by it, you're going to be quite angry. So I'm just letting people know what can happen, and under what type of circumstances it can happen.

I really, really, really, really don't see what's wrong with that...

 

[harddrive]

A history question - I have (well, in a basement 3,000 miles away) an OLD consumer camera. It's a JVC, from the era when the VHS deck was slung over your shoulder and the handheld camera attached by cable - early 1980's. I'm pretty sure it was CMOS. Was there not an era when CMOS was more popular for video imagers?

My memories of that era are that the cameras such as you describe were tube based - a single vidicon if I remember correctly. My feeling is that it was the emergence of viable CCD technology that meant it was possible to form a one piece (consumer) camcorder unit - it was the tube technology that drove such as JVC to make it as a two piece unit.

I can only say that if I'd been handed an EX1 in the mid 80's, I wouldn't have refused it on the grounds of a rolling shutter........

 

[Barry_Green]

Well, yeah -- because the old tube cameras pretty much behaved the same way as a rolling shutter... :thumbsup:

 

[tnle2]

I was being sarcastic about the "crippling," hence the quotes. I design products for a living so I am well aware of the trade-offs and compromises that engineers make when designing a product. Cost is only one of many considerations that need to be prioritized. In some applications (mission-critical ones) cost is very low in the priority. There you design it to do what it has to do it just cost what it has to cost. Examples that spring to mind include aircraft parts, spacecraft and very high-end movie production equipment. I'm not saying RED did not consider cost at all, but for other guys like ARRI, I don't see it being a big factor when their target market already has very large budgets. One thing I noticed about RED was they did not prioritized weight reduction early on.

Regardless, the crux of my point in even bringing this up, and trying to educate people, is this: these new camcorders do not behave the way the old ones did... I really, really, really, really don't see what's wrong with that...

Nothing wrong with that and I agree that everyone should be aware of the limitations of their equipment. But that's not what this recent debate was about. Your comments a few posts back painted a picture of CMOS as being not well suited to high-end work.

Comments like..."[CMOS] Makes a lot of sense for cheap camcorders/cell phone cameras. Makes less sense for something like the Infinity"..."What top-of-the-line cameras use CMOS?"...."EFP cameras? No, not buying it"...."Nope, CMOS is higher noise. Higher sensitivity? Sorry, no, CMOS is a lot slower"..."every manufacturer that offers both CCD and CMOS has, as far as I know, always chosen CCD for the high end, and only offered CMOS in the lower-cost end of their product line."

Some of these statements appear to be based on inaccuracies and/or outdated information so I chose to correct you.

Btw, for some time now, Ikegami has been selling CMOS cameras that are used in EFP sports applications and it seems no B.S. has hit the fan since Bexel and mobile trucks are buying more of them. I think Thomson will be alright.

 

[ullanta]

My memories of that era are that the cameras such as you describe were tube based - a single vidicon if I remember correctly. My feeling is that it was the emergence of viable CCD technology that meant it was possible to form a one piece (consumer) camcorder unit - it was the tube technology that drove such as JVC to make it as a two piece unit.

Actually, it was a 3-piece unit! Camera - VHS-Deck - Tuner. I can't track anything like it down on the web. However, after you said "Vidicon", the word "Saticon" popped into my head as something that was written on it; but none of the Saticon cams on the web are the one I had. Now I'm just obsessed - anyone know a model number or anything to help me track this down? The camera connected to the VHS deck with a high-pin-count (I'd guess around 30) round connector...

Anyway, yeah, it seems certain it was a tube, though I do think it said CMOS... but who knows for what?

 

[tnle2]

It's either a vacuum tube camera or a CCD. CCDs came right after tube cameras. It couldn't have been CMOS, which came much later. I believe the first consumer CMOS video camera was the [SIZE=-1]Sony DCR-PC1000 in 2005.

Some retro camera history: http://www.digicamhistory.com/1970s.html

Maybe you saw something else on the camera like MOSFET or something like that.
[/SIZE]

 

[ullanta]

It's either a vacuum tube camera or a CCD. CCDs came right after tube cameras. It couldn't have been CMOS, which came much later. I believe the first consumer CMOS video camera was the [SIZE=-1]Sony DCR-PC1000 in 2005.

Some retro camera history: http://www.digicamhistory.com/1970s.html

Maybe you saw something else on the camera like MOSFET or something like that.
[/SIZE]

Nice page... but still not there! Yes, I'm pretty sure it must've been a Saticon tube. But, you know, I am and was a techie, and wouldn't confuse CMOS for MOSFET... I think... I just wonder if somewhere other than the imager they were advertizing some sort of CMOS tech... just curious. The camera itself was small, held like a modern small video camera rather than with the pistol grip on a lot of the cameras of the early 80s... it was more modern-looking than the boxy things pictured, too.

 

[Illya Friedman]

I honestly don't know what it cost to make a high-quality CMOS sensor but I doubt it's much cheaper than CCD like you keep repeating.

I also don't know all the costs are for sensor manufacturing, as that's not my area of expertise. I do know that irregardless of sensors type, features (like rolling and global shutters) as well as performance and fill factor will greatly affect manufacturing costs.

I've heard (and I hear my share of both substantiated and unsubstantiated talk) that CMOS sensors >can< be far less expensive to manufacture. And while CMOS performance has increased in recent years there's some very good reasons why the current best digital cinema sensor in the world is a CCD.

A high fill factor is extremely important for performance, so it might be good to know the fill factor of any sensor your interested in evaluating. I make a point of asking the question whenever I meet a manufacturer who knows, or ought to know.

There's some very good reading here for those that are interested:

http://www.dalsa.com/markets/ccd_vs_cmos.asp

I.

 

[rockoblm]

I'm typically a reader, and not a comment person (I guess a lurker), but I gotta say, the pissing matches on this forum are down right annoying!

 

[tnle2]

I'm typically a reader, and not a comment person (I guess a lurker), but I gotta say, the pissing matches on this forum are down right annoying!

Sorry about that. Looking back now, you're right, it was annoying. Although my intent was to correct factual information or provide new information, the back-and-forth debating just created a lot of noise.