February 27, 2015, 02:46:41 AM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - UrbanVoyeur

Pages: 1 ... 3 4 [5] 6
* if you have any doubts that Canon is ripping you off, just look at the video side of the market; BlackMagic is selling (still not shipping, but that will come in august) a camera that puts all the canon, sony, panasonic, etc, models to shame, for $3K (with $1500 of awesome software included); Canon could do the same, with a bigger sensor, for $4K; yet they prefer to sell it for $40K; good for them, but just don't expect me to buy it (ok, the C500 has some things that are better than the BMC, but some are worse too; definitely not worth the price difference unless $30K is just small change for you)
Yeah, I think between the Black Magic, the lower cost RED's and the Sony NEX line, the whole digital video market is in upheaval. I don't think it makes sense to buy into any of these until things settle down bit. But I will be renting the Canon 4k when they are available. :-)

The center point was an f/4 cross, the others required f/2.8 or faster lenses. With every 1-series digital body and with the 1V and 3 film cameras, an f/4 lens gives one and only one cross point (except certain specific f/4 lenses on more recent bodies), and an f/5.6 lens gives you no cross points at all.  On the 1D X and 5DIII, an f/4 lens means 41 crosses, an f/5.6 lens means 21 crosses, and an f/2.8 means 41 crosses and 5 dual crosses.  Old tech and AF sensor redesign, sure.
On the 3 & 1V, yes there were 7 standard cross points usable with f/2.8 or greater lens. Yes, only the center one worked with lens with a max ap of less than f/2.8. Starting with the EOS 1Ds MkIII in 2007, 19 additional cross points were added.
Is the AF on the 5DKIII better than the EOS 3 and 1V of 14 years ago? Yes.
Is it new technology? No. Just a modest refinement of old tech.

We are nit-picking here. You asked if Canon had AF with >19 point and at least 1 point cross-point workable at less than f/2.8 in the past. I showed you several, some dating back over 14 years which met this criteria.  Yes, the system has improved over the years - that's to be expected. And yes, the 5DMkIII benefits from that refinement. But no, it is not a major breakthrough and to me, not worth the premium Canon wants.

These 'facts' are making me really curious. While the EOS-3 was launched in 1998 having the 45 point AF, why Canon never implemented this in any of the cameras other than the 1 series? Was it the issue of cost of production being high or was Canon simply hiding this up its sleeve all this while?
I don't know.  I don't think it is that expensive to implement.  From what I understand of the AF technology, the number of focus points and cross points  is more a matter of software analysis rather than major changes to the AF detector. And of course, a focusing screen withe LED overlay.

Hackers showed that it was possible to unlock the more focus points on lesser models cameras, like the 20 & 30D  and even configure more AF cross points.

Canon also never brought back eye focus (which I loved) or for most of its line, slaving the 2% spot meter to a selected focus point.  Though here again, hackers showed that the spot meter to focus point link existed in the firmware. And that it could be made into a 1% meter on some models.

Edit note: I was right about the number of AF points (45) in the 3 and 1V, but wrong about the number of cross points (max 11, not 19 or 21) . I corrected it in my post and I wanted to make sure you knew of the change.

Another interesting 'fact' is that EOS-3 was launched at a price of 185,000 yen which translated to roughly USD 1500 in 1998. On current exchange rates, these yen translate to USD 2,300. That's a whopping $ 800 hit on forex alone. 
That may have been the retail price. Film cameras typically sold for much less. I paid about $800 for the body at Adorama around 1999/2000.

Anyhow, I don't see the reason why Canon shouldn't be recovering its R&D cost even now, when it put this system in a very select group of products in the first place.
I guess, but I think they recovered the cost of all the legacy technology (that shared with film bodies) a long time ago. Overall, I'm more concerned that people think the AF, FPS and prism coverage are expensive tech and justify a price increase, or even maintaining the old price. They don't. They are essentially give aways.

Wow, so Canon had an AF system with many f/4-sensitive crosses?  They had AF sensors with >19 points and even one f/2.8+f/5.6 dual-cross, much less five? Which models were those again, I must have missed them?
Yes, the EOS 3 and 1V film cameras, had 45 focus points, 11 of which were cross 7-11 of which were cross depending on CR settings and worked from max aperture of f/1.0 down to f/5.6, depending on the lens. Adding more cross points was done very early in the EOS 1D line. So yes, this is very old tech.

You also claimed that Nikon has outsold Canon for the past few years. From what orifice do you pull out these non-facts?
I never claimed that. Where did I say that? I don't know who outsold whom in the past several years, and in any case it is not germain to any of my arguments.

BTW, what makes you think the sensor is the only restriction on frame rate?  You don't suppose they might...just might...have needed to redesign the mirror assembly to achieve 6 fps vs. 4 fps for a FF camera, do you?  Nah, that conflicts with the non-facts pulled from your...well, I'll quit while I'm you're, ummm, behind.

Again, a misquote. I said that the sensor had nothing to do with the increase in frame rate from 4-6.  I said that the size of the high speed RAM buffer on the camera controlled the FPS. More buffer RAM = higher FPS.

The mirror system would not need to be redesigned. Canon EOS film cameras have been capable of 6 FPS at the lowest model end for over 20 years. Nearly every EOS film camera that took a winder could do 6 FPS with the stock mirror. I know. I own them. Same mirror, same flip action. That problem has been solved. Canon does use a more robust reflex system when going for 11+ FPS, but again, the timing issues and mechanism were worked out at least 20 years ago. It's just a part number for them now.

I understand that these aren't important features to you, but for many people the new 61-point AF and faster burst rate are huge upgrades.
Actually, the 61 point auto-focus is a very important feature to me - as is the option of slaving the spot meter to the focus point, which I don't know if the 5D MkIII custom settings allow. It is one of my favorite features from the EOS 3 and the 1D line.

I do take exception to the 61 point auto-focus being touted as some great advance or new feature when it's really a freebie. As is the increased FPS from 4 to 6. To me, there is nothing in the MkIII that justifies the price.

I started this thread wondering how Canon could have ended up here, with a marginal upgrade at an inflated price that is completely out classed by an aggressively priced competitor with a massively improved sensor.

My thought was that Canon probably tried, but given the deadlines in manufacturing, just couldn't get their comparable sensor together in time for release. I do believe Canon has a high MP, low noise, high DR sensor at the same price point - there is nothing magical about the Sony sensor.

I hope Canon pulls it off in the near future and is able to manufacture it at scale, or they are in for a bumpy ride as lower cost derivatives of this Sony sensor filter through Nikon line up. I think we are seeing the first couple in the D3200 and upcoming D600. 

At that point, I don't think anyone would have dreamed that Canon would put a 61-point AF system with 41 cross-type points into the 5DIII, but they did, and now all of a sudden the camera is a POS? Anyone that suggested that the 5DIII might have 41 cross-type points at that stage was considered an idiot and flamed profusely.
I would like to point out that 61/41 point/cross point AF is old technology that Canon perfected in the EOS line over 15 years ago. It costs them essentially nothing to put it in any camera - no software development cost, no new hardware to develop. The marginal cost of the 61/41 point screen and AF sensor is essentially nil.  So yes it is an important feature, but hardly an advance of any kind.

Same for the ever so slightly higher frame rate (4 MkII vs 6 MkIII). That is a function of the size of the high speed memory buffer. It has nothing to do with the sensor. It is very old technology, with very little cost associated with the modest frame rate increase. It may have even cost less, given the drop in price of fast RAM since this MkII introduction.

The processor would need to be part of the sensor package, reliability would be poor, and you would be locked into old technology which would likely stop being made after 4 years.
Not really. Senors are just arrays. The conversion to image is math. If you present the array to the software, along with the measured characteristics of the sensors and the location map of the pixels, then the processing to an image is a calculation, independent of the sensor. A firmware upgrade would be needed with a new sensor, but in many cases, probably not a new processing chip. The DIGIC 5 is much faster than DIGIC 2, even though the processing chip used is only slightly faster.

I'd much prefer to get a whole new camera every three - four years.  switches wear out, LCD's develop bad pixels, the body wears and gets scratched up, it isn't just about the sensor.
That is more an issue of the build quality. More expensive bodies have longer duty cycles. I've burned out Rebel shutters in as little as a year, where my EOS 3 shutter is still perfect after 12 years. Same with LCD's. Scratches? Well, those happen.

You cannot just change sensors from one type to another, they have totally different technology, different pin-outs, different numbers of readout channels, etc.
Pin outs yes - that of course would need to be standardized.
Technology no. Sensors in DSLR's these days are almost all CMOS. CMOS sensors essentially all work the same - light enters, voltage comes out. They are all designed in a rectilinear or hexagonal array. They all map each pixel to a single physical location and to a hard wired pin out. They all have fixed physical dimensions (FF, APS-C, 4/3, etc). They all must output RGB. They all have similar power and cooling requirements (they use nearly the same batteries and fit in the same relative body sizes) The rest is software interfacing.

Wish granted.   :P
Close. Maybe the ideal is not that far off.

I wonder how long it will take before we come full circle and are able to upgrade/ change sensor packages and sensors become like film. After all, full frame lenses have the same minimum circle of coverage by definition, as do to all APS-C and 4/3’s lenses.

Not necessarily upgradeable by the end user – sensors still require precise alignment & a dust free environment – but say by a repair shop for $300-500 - what they charge in labor to replace a damaged sensor on a DSLR. (Though a very clever inventor might develop a way to auto align a sensor.)

Then you could buy the body features you want and select from either the OEM sensors or get an after-market. Sensors could be selected on size, MP count, DR and noise without regard to body make. Sony sensor in Canon. Canon sensor in a Pentax. No different than Fuji film in an Olympus.

Frames rate would be function of the amount and location of high speed memory and the speed of the digital processing chip, more than the sensor itself. High frame rates would still require expensive high speed RAM and processors – as they do now.

Other than standardizing the sensor pin-outs and some of the firmware API calls, there’s no serious technological hurdle. All sensors are mapped arrays like LCD screens or scanners. The power and cooling requirements are all roughly the same across comparable sensors, because ultimately, they all have to work with the same size batteries and bodies.

Metering and auto focus are sensor independent.

The various image processing chips are more about the firmware than the actual semi-conductors. They all have to do the same functions. Performance appears to be mostly a matter of software - from DIGIC 3 to DIGIC 5, the chips had only a modest increase in clock speed. And as hackers have shown, the chips and software are fairly standard across a given OEM’s models – it’s only a matter of unlocking features in the firmware.

You could even go further and make the digital processing chip upgradeable, though I suspect you would have to continue using a Canon processor in a Canon body.

The obvious place to start would be the 4/3’s cameras, which already maintain standards and interoperability.

Why would the camera manufacturers do this?

It would free the body development cycle from the sensor development cycle. Since the days of film, camera makers always made money from selling bodies, and this is no different. Canon for example, could refine its bodies and digital processing in response to market demands much more quickly, without having to wait on a new sensor.

Manufacturers like Nikon already source their sensors from Sony, allowing Nikon to focus on bodies and processing software, and letting specialized companies with more resources focus on sensors. This is just another step in that direction.

Camera makers could continue to make money selling various sensor and digital processing upgrades. When you figure that most of the retail price (not cost!) of a given SLR is digital components, breaking that out into its own product line makes sense.

It also opens up the possibility of putting higher end sensors in lower end bodies. There is no physical reason why a Rebel Ti, a 60D or 7D is limited to a APS-C size sensor other than the sensor itself and the choice of prisms.

10 years?

Perhaps, just maybe, consider the fact that your opinion isn't representative of 'many/most' users.
Umm. Ok.

Canon's market cap is much bigger than Sony's if I remember correctly
The market caps of the two companies are about the same but Sony's semi conductor business is much larger than Canon's.  see wikipedia and yahoo finance.

And there are no other sensors with the same pixel density as the D800--so how could Nikon be paying Sony for the best of each run?  They're paying for all of these sensors that exist, at least for now.
Creaming is a common practice in the semi conductor business. Intel does it every day with chip Pentium chip speed, as does Apple and its ARM chips.  In a given run, the chips that past the most stringent tests get the highest speed rating, and thus command a premium price. Chips that pass the tests at lower speeds sell for less. The same also applies to quad vs dual vs single processors. The best become quads, and so on down the line.

The sensor Nikon gets is the last step in long manufacturing process for Sony, which starts from silicon wafers.

Like every semiconductor part each of those steps has a yield; in gross terms good, bad, perfect and junk are produced in each run.

Nikon pays for the low noise, high DR, full frame 36 MP sensor - that's the only chip they want. In that same run, there may be low noise, high DR version, but only if you ignore the outer sensors - APS-C @ 22 MP.  Or you my get a full frame with great noise and DR, but you cannot use adjacent pixels - FF @18 MP.
My examples are crude, but I think you see the point. There are many cameras models from Sony and other manufacturers that don't need 36 perfect MP's on a FF chip, which may present an opportunity to sell chips Nikon rejects.

* Most people don't want the 5d3 to have any more resolution and I'm sure Canon knows this because they will have asked
Until it comes out. Then everyone wants it. Nikon users are no different than Canon - both companies hear the same things from their customers, and both companies have a pretty good idea of how the market will react to a given upgrade.

Given a choice between more MP OR less noise and more DR, many/most users choose the latter. When given the option of getting all three (more MP, less noise, more DR) nobody turns it down.

Editor's Note: I am retracting the "many/most" statement. It is factually incorrect (not being based on anything but my conjecture and opinion) and is a distraction to the discussion.

The 5d3 sensor will have been being worked on for probably 2-3 years now.
That's kinda my point. Canon has been planing this upgrade for 3-5 years and working on a new sensor for that long. They just couldn't pull it all together in time for the MkIII.

EOS Bodies / Why did Canon Release the 5D MkIII (pure conjecture)
« on: May 03, 2012, 10:22:57 AM »
I don't think Canon planned to release the 5D MkIII as it exists today. I think Canon intended to put a much higher MP, lower noise sensor with greater DR in it, but it wasn't ready in time. This is all pure conjecture.

I think Canon was fully aware of the Sony sensor and Nikon's plans, and physics and electronics being what they are, was able to produce a sensor equal to the one in the D800 using Canon technology. There's nothing revolutionary about the Sony/D800 sensor. But I don't think Canon could get it produced in the quantities they needed.

I suspect there was a quality control/yield issue. Yes, the sensor worked, but not enough of them were coming off the production line that met their standards - too many rejects in each batch.

This left Cannon with some choices:
- Release no update to the 5D MkII, already 3+ years old until the sensor was ready, and in so doing, appear to cede that market segment to Nikon.
- Put the top of the line 1D series sensor in the MkIII - an improvement, but the only way to maintain the price point would be to take a loss/break even on every body.
- Put a slightly tweaked MkII sensor back into the MkIII along with the other planned feature upgrades and basically mark time until the new sensor was ready.

I think they took the last option - issuing a minor upgrade so as not to be seen abandoning the market segment to Nikon.

Nikon took advantage of the situation by knocking down the price, and I think that, more than anything else caught Canon off guard. Canon would do well to take a price cut on the MkIII.

Why did Sony and Nikon succeed with the new sensor where Canon did not?

I'm not so sure they did.  First, I think Nikon may be creaming the sensor production - paying a premium to Sony for the very best of each run.  They don't need that many, since the D800 is a relatively expensive, low volume camera.

A company Sony's size may be able to afford production runs 100x or 1000x larger than Canon. Their yield of the very best sensors may be no better than Canon, but they may have buyers for the lesser sensors - other camera makers and other Sony models that may be happy with fewer pixels, more noise, lower DR or even, via trimming, smaller sensors.

Or Sony can afford to take a loss on initial production runs, knowing everything will work out in the near future. Sony did this with the PS 1 and PS 2 gaming processors.

It could also be that Canon's sensor production facilities for this model were in Thailand, were flooded out, and the replacement factories are not fully on line yet or tweaked for higher yields.

I think Canon's next full frame camera will be very informative. If it is in the  18-24 MP range, but retails near the 7D, then I would bet that these are based on the same sensor that was originally planned for the 5DMkIII, but represent the lesser quality production yields.  (18-24 usable low noise, high DR pixels vs 36-45)

I bet that within a year, Canon updates the 5D MkIII type camera with a much higher MP count, lower noise, higher DR sensor. They may not call it the 5D, but it will be what they had originally planned for the MkIII.

The D800 represents a reversal in the megapixel situation with Nikon surpassing Canon, and no change in the DR situation, with Nikon still out front.  How else could Canon interpret that?  Canon will assume the problem is MP, revert to their former more-is-better strategy, and we'll see a high MP body with no greater (or perhaps even less) DR.
I wish that Canon would take up the challenge and push past Nikon in MP count, DR and noise, but I agree with you - Canon will probabley stick to the MP count. It's easier.

Did you account for the fact that Canon doesn't count dollars to determine profit, but that they count yen instead?  That means exchange rates are a huge factor in pricing (the strong yen accounts for much of the recent price hikes, for both Canon and Nikon), and also the ever-shifting landscape of trade tariffs, luxury goods import duties, etc., play a major role (which is why Canon gear appears to cost more in the EU, on a dollar for euro comparison).
Nope, didn't consider it at all, as it is irrelevant. I don't care what the camera costs to produce or how much profit Canon makes. I only know what Canon chooses to charge me in dollars and how that stacks up to the previous version and the competition.

The absolute price isn't an issue for me in this case. At issue is whether the Mk3 offers enough features to justify the price of the upgrade compared to what is possible (Nikon) and what Canon has offered in the past (Mk2). It does not.

If the Mk3 offered the same MP, DR, and low noise as the current tech leader, Nikon, the Mk3 would be worth the upgrade. I'll wait until what Canon is offering @$3500 is more in line with what is possible, the D800 @ $3000.

If you want the camera, you pay the cost.  If the cost is too high, you don't buy the camera.  I guess on a personal level, it's pretty simple after all.
That about sums it up.

Wow, there are just so many misassumptions there that I'm not even sure where to begin.  You don't, by any chance, do cost analyses for the US government, do you?   :o

No I don't :-). But I have been buying and using the Canon EOS system for over 20 years. I still own several EOS film cameras, and I can match them just about feature for feature with their their digital counterparts.  Build quality, lens mount quality, sealing, shutter MTF, % VF, auto-focus points, metering, etc.

I've come to know how much of a price increase there was when Canon "improved weather sealing", "added metering modes", "increased VF %" or "increased the number or AF points" between one model year and the next. Answer: essentially $0.

I also know what each of the film cameras cost new, and in many cases, what it cost to have all or parts of the film transport replaced or repaired. I've been pretty hard on some of these cameras.

So when I look at a Canon digital camera, I also look at it's nearest film sibling.  The difference in price between the two, minus the retail parts cost of replacing the film transport, is the "digital premium" Canon is charging me for that camera.  I am not talking about what it costs Canon to make the camera. I don't know and I don't care. I am only concerned with what Canon is charging me for the digital features.

So if the Canon EOS 3 ($800 new) and the 5MkIII ($3500 new) are nearly identical in build and features. And I was charged $275 in parts for replacing a destroyed film transport, then retail price of the non-digital parts of and EOS 3 type body I can assume is $500. NOT what it cost Canon, but what they are charging me for it.

Likewise, if an EOS 3 body is like a 5MkIII body, then the digital premium - what Canon is charging me for the sensors, chips and software - on the 5 MkIII is $3000. ($3500-$500)

Do the same analysis for  the Nikon D800. It has a film sibling that sold for around 800-900. Assume the film transport price is the same. The Nikon digital premium is $2500, for 36 MP and near medium format performance.

So the question for me becomes, is that $3000 Canon digital premium worth it. My answer, in this case, is no, not when compared to the MkII and not when compared to the value offered by Nikon.

The analysis would be just as valid if I took off nothing for the film transport. The digital premium would still be there, and Canon would still be offering less value than Nikon, and an insignificant improvement over the MkII

Pages: 1 ... 3 4 [5] 6