New Canon on 22nd March?

Started Mar 7, 2013 | Discussions
(unknown member) Senior Member • Posts: 2,853
Re: New Canon on 22nd March?

noirdesir wrote:

howardroark wrote:

The reason many of us refer to them as "war" is similar to the reason there was an HD format "war" between HD DVD and Blu Ray around the same time. There were two ways of thinking and they were like the old saying "there are only two kinds of people in this world" because there were very few people who were on the fence about the issue of more pixels versus better pixels. For a long time improvement came in the form of more pixels. There were very tiny technological advances in the pixels themselves but big manufacturing advances in getting more pixels on a chip. People wanted higher resolutions, but we also really, really wanted better image quality in the form of lower noise/higher useful ISO's (not to mention better dynamic range).

If your argument is that the quest for smaller pixels led to less investment into improving QE, saturation capacity and read noise, you might have a point. But then some of the same underlying technology improvements improved both pixel count and the signal quality.

A lot of companies were making huge jumps in pixel count but the noise got worse and worse.

Could you give us some examples? Because I don't remember any but then I did not follow things as closely ten years ago.

It seemed like DSLR's were heading the same way. Notice how many camera models over the last few years had fewer pixels than their previous iteration?

Again, the only example I remember was the Canon G11.

S1, 2004, 3.2MP, 1/2.5" (5.744 x 4.308 mm)

S3, 2006, 6MP

S5, 2007, 8MP

SX1, 2008, 10MP, 1/2.3" (6.17 x 4.55 mm)

SX20, 2009, 12.1MP

SX30, 2010, 14.1MP

SX40, 2011, 12.1MP

SX50, 2012, 12.1MP

I put the sensor size next to the number of pixels for a reason:  pixel density is what really matters.  Other Canon lines had similar, or more alarming, trends.  Other manufacturers were worse and most people were glad when Canon pulled back a bit and concentrated on improving image quality rather than just raw pixel count.  Lenses on P&S cameras aren't usually that great to begin with, so ruining the images that already suffered from low detail resolution and/or distortion with more and more noise wasn't doing anybody any good.  DSLR lenses offered plenty of inferior selection that couldn't keep up with the sensors, but eventrually more and more affordable but high quality options were released.  I also mentioned the G line in another thread.  That one at least had a larger sensor so pixel densities weren't quite as ludicrous.

Many of the consumer lines had small chips like this example but were up at 18MP before they backed off.  I was glad that Canon didn't quite follow the pack as closely as others, but that goes to show that sometimes Canon resists market trends for a reason.

robert1955 Veteran Member • Posts: 4,222
Re: New Canon on 22nd March?

THat is still only one example, and I'm not sure there is much evidence that the image quality got better from reducing the pixel count.

Note that the most important differnece between the SX30 and 40 is that Canon changed form a CCD to a BSI-CMOS at the same time. That change means IMO that you can't validate your claim on this ititaration.

howardroark wrote:

noirdesir wrote:


Again, the only example I remember was the Canon G11.


SX20, 2009, 12.1MP

SX30, 2010, 14.1MP

SX40, 2011, 12.1MP

SX50, 2012, 12.1MP

 robert1955's gear list:robert1955's gear list
Fujifilm X-E3 Fujifilm XF 16mm F1.4 R WR Fujifilm 50mm F2 R WR Samyang 12mm F2 NCS CS XF 90mm
bobn2
bobn2 Forum Pro • Posts: 60,889
Re: New Canon on 22nd March?
2

howardroark wrote:

I put the sensor size next to the number of pixels for a reason: pixel density is what really matters.

Could you provide some backing for that statement? Can you find some evidence that SNR declines for same size sensors as pixel density increases? Sure, there are one or two cases where the pixel size is pushed a little above the optimum for that pixel design - a good case is the Sony 24MP APS-C, which is a bit worse performer than the 16MP, but subsequent 24MP chips from Nikon and Toshiba showed that is not a fundamental limitation, and I'm sure that Sony will be back with a revised 24MP chip which performs fully. Unless you can give some evidence, this will look very much like another unsubstantiated opinion to many people.

Other Canon lines had similar, or more alarming, trends. Other manufacturers were worse and most people were glad when Canon pulled back a bit and concentrated on improving image quality rather than just raw pixel count.

I'm wondering here how you tested the opinion of 'most people'. Did you take a poll? How did you sample, and if you used, for instance DPReview as then medium, how did you correct for the systematic biases that would occur using the active membership here? However, it seems that DPReview reviewers were excluded from 'most people', because they noted in their review of the G11 that: 'The G10 remains the most impressive small-sensor camera we've seen, at low ISO settings. The level of detail it renders is very impressive and is enough to put some entry-level DSLRs to shame'. Remember, the G11 used exactly the same lens. The only benefit that the G10 had was its higher pixel density sensor, and that allowed it to deliver what the reviewers found to be higher image quality.

you say on an earlier post:

Many much smaller chips were shooting up to 18MP and the images were awful.

I once found a French web review of a Sony 16MP 1/2.3" P&S, comparing it with its 12MP predecessor. Pity I didn't bookmark it, because I can't find it now. Anyway, they came to a similar conclusion to you, on the basis of 100% pixel peeping. When they go some questioning comments they took the trouble to run a blind web poll, with unlabelled images from both cameras, asking which their readers thought were better. Overwhelmingly the 16MP won.

Without some evidence, the statement that 'the images were awful' will look very like unsubstantiated opinion to many people.

-- hide signature --

Bob

jonas ar Contributing Member • Posts: 606
Re: New Canon on 22nd March?

What have led you to that cocnlusion on the performance of the 24MP Sony APS-C sensor? From what I have seen they perform very similarly provided that the exposure is controlled.

Kind regards,

Jonas

bobn2 wrote:

Sure, there are one or two cases where the pixel size is pushed a little above the optimum for that pixel design - a good case is the Sony 24MP APS-C, which is a bit worse performer than the 16MP, but subsequent 24MP chips from Nikon and Toshiba showed that is not a fundamental limitation, and I'm sure that Sony will be back with a revised 24MP chip which performs fully.
Lenni Vilen Regular Member • Posts: 139
24 vs 16

bobn2 wrote:

Sure, there are one or two cases where the pixel size is pushed a little above the optimum for that pixel design - a good case is the Sony 24MP APS-C, which is a bit worse performer than the 16MP

I am not so sure if this is even case as withing the the Sony camera families the 24Mp seems to perform at least as well as the 16Mp one plus the resolution advantage. Though there is one issue is made more significant by the finer pixel pitch - with near symmetrical lenses there will be more significant magenta color cast near the edges and potential for detail smearing.

4thnebula Senior Member • Posts: 2,271
Possibly but maybe not

Depends on if they go to the rumoured 180nm image sensor process (which should be a major improvement and put them ahead of Sony/Nikon) and significantly improve liveview af (which would hlep their video as well).  If the 70D has a 3.2" articulating touchscreen LCD and liveview improves to at least near u43rds speed (with big ef lenses like the 70-200 f2.8 IS) then we will know more good things are coming.

So, we are only 2 weeks away from the answers to those questions since the new APS-C sensor will likely also be the one in the new EOS-M and 7D MKII.

 4thnebula's gear list:4thnebula's gear list
Canon EF 35mm F2 IS USM Olympus PEN E-PL5 Sony a7 III Canon EF 50mm f/1.4 USM Panasonic Lumix G 20mm F1.7 ASPH +5 more
bobn2
bobn2 Forum Pro • Posts: 60,889
Re: New Canon on 22nd March?

jonas ar wrote:

What have led you to that cocnlusion on the performance of the 24MP Sony APS-C sensor? From what I have seen they perform very similarly provided that the exposure is controlled.

The only really sensible comparison, IMO, is same exposure - otherwise you are comparing apples and oranges. Unfortunately many of the review sites do not set the same exposure (maybe that is what you mean by 'exposure controlled'). The 24MP sensor has the same per-pixel read noise as the 16MP sensor, and the same efficiency, so in the end it suffers from a little more shadow noise. That's the only place where it lacks the 16MP. The read noise scales down with the read transistor, so a fair reading of what Sony has done is use the same read transistor design in the 16 and 24MP pixels. If they had just taken the 16MP design and scaled it proportionately, it would have produced a lower per pixel read noise and higher DR overall.

-- hide signature --

Bob

bobn2
bobn2 Forum Pro • Posts: 60,889
Re: Possibly but maybe not
1

4thnebula wrote:

Depends on if they go to the rumoured 180nm image sensor process (which should be a major improvement and put them ahead of Sony/Nikon)

Hardly. Sony FF sensors are already on 180nm, Sony has fab lines going down to 65nm, and use them for their P&S and phonecam sensors. The 180nm line dates from 2010, so it would be very surprising if it hasn't been in continuous use since then, otherwise the capital investment would be just frittering away. Mostly it will be used making the new CMOS sensors Canon is using in cameras like the G15 and S100, which need the finer geometries. That is Canon's problem, if they move all production onto the new line, they both restrict their ability to make the compact camera sensors by using capacity for DSLR's, they also use the old line unutilised. Possibly Canon will go back to Sony sensors for all compacts, move to the 180nm line and just use the 500nm line for FF. But be clear, 180nm does not put Canon ahead of Sony, it doesn't even catch up.

and significantly improve liveview af (which would hlep their video as well). If the 70D has a 3.2" articulating touchscreen LCD and liveview improves to at least near u43rds speed (with big ef lenses like the 70-200 f2.8 IS) then we will know more good things are coming.

So, we are only 2 weeks away from the answers to those questions since the new APS-C sensor will likely also be the one in the new EOS-M and 7D MKII.

The 180nm process will not get them ahead, and if they haven't managed to use it to put a column ADC in the chip, they'll remain behind.

-- hide signature --

Bob

bobn2
bobn2 Forum Pro • Posts: 60,889
Re: 24 vs 16

Lenni Vilen wrote:

bobn2 wrote:

Sure, there are one or two cases where the pixel size is pushed a little above the optimum for that pixel design - a good case is the Sony 24MP APS-C, which is a bit worse performer than the 16MP

I am not so sure if this is even case as withing the the Sony camera families the 24Mp seems to perform at least as well as the 16Mp one plus the resolution advantage.

The difference is subtle, it is slightly higher shadow noise, unless you look into the deep shadows, you don't see it.

Though there is one issue is made more significant by the finer pixel pitch - with near symmetrical lenses there will be more significant magenta color cast near the edges and potential for detail smearing.

Yes, if you can't reduce the stack height in proportion to the pixel size, that is a problem - which is why the manufacturers are moving towards copper interconnect.

-- hide signature --

Bob

jonrobertp Forum Pro • Posts: 12,875
19 pt AF
1

4thnebula wrote:

Depends on if they go to the rumoured 180nm image sensor process (which should be a major improvement and put them ahead of Sony/Nikon) and significantly improve liveview af (which would hlep their video as well). If the 70D has a 3.2" articulating touchscreen LCD and liveview improves to at least near u43rds speed (with big ef lenses like the 70-200 f2.8 IS) then we will know more good things are coming.

So, we are only 2 weeks away from the answers to those questions since the new APS-C sensor will likely also be the one in the new EOS-M and 7D MKII.

Here's hoping they will go an extra step and make the price $1299 to 1499...and be worth it ..with the new 22 meg sensor, and a new or revised 19 pt cross sensor AF...6 fps is ok...we can leave the 10fps to Sept.

Many coins would roll...if the body is worth it...it's not that hard to beat the D7100...do it with AF.  They have only 15 cross sensors ...and only in the middle.   will they ?   or won't they ?  That is the question...aka Shakespear.

 jonrobertp's gear list:jonrobertp's gear list
Canon PowerShot G7 X Canon G3 X Panasonic ZS100
bobn2
bobn2 Forum Pro • Posts: 60,889
Re: 19 pt AF

jonrobertp wrote:

They have only 15 cross sensors

Line sensors are also effective, and cross sensors are less effective towards the edge of the fame.

-- hide signature --

Bob

(unknown member) Senior Member • Posts: 2,853
Re: New Canon on 22nd March?

robert1955 wrote:

THat is still only one example, and I'm not sure there is much evidence that the image quality got better from reducing the pixel count.

Note that the most important differnece between the SX30 and 40 is that Canon changed form a CCD to a BSI-CMOS at the same time. That change means IMO that you can't validate your claim on this ititaration.

I have no desire to write a dissertation on the subject.  I've paid close attention to camera technology since I bought my Sony S85 ages ago and even more closely after buying my 10D shortly after.  I watched the camera reviews, read the forums, and while I don't wish to spend my time catching you up on all of the things I, and many others, saw over that time I'm certain some digging on your part (if that's how you choose to spend your time) will support my recollection of history.

jonas ar Contributing Member • Posts: 606
Re: New Canon on 22nd March?

I meant that I have seen picutures from Sony cams using the two sensors with the same exposure (I believe posted by dpr member RicksAstro). They were very  similar. The quantitative measurements at DxOmark from a55 and a77 also seem to indicate that they would perform essentially the same and it does seem that the read noise is somewhat smaller in the a77 from your own sensorgen figures? At least in Sony's implementation of the sensor.

robert1955 Veteran Member • Posts: 4,222
Call your bluf
1

After a reply like that, I'd say your hand only has low cards

howardroark wrote:

robert1955 wrote:

THat is still only one example, and I'm not sure there is much evidence that the image quality got better from reducing the pixel count.

Note that the most important differnece between the SX30 and 40 is that Canon changed form a CCD to a BSI-CMOS at the same time. That change means IMO that you can't validate your claim on this ititaration.


I have no desire to write a dissertation on the subject. I've paid close attention to camera technology since I bought my Sony S85 ages ago and even more closely after buying my 10D shortly after. I watched the camera reviews, read the forums, and while I don't wish to spend my time catching you up on all of the things I, and many others, saw over that time I'm certain some digging on your part (if that's how you choose to spend your time) will support my recollection of history.

 robert1955's gear list:robert1955's gear list
Fujifilm X-E3 Fujifilm XF 16mm F1.4 R WR Fujifilm 50mm F2 R WR Samyang 12mm F2 NCS CS XF 90mm
bobn2
bobn2 Forum Pro • Posts: 60,889
Re: New Canon on 22nd March?

jonas ar wrote:

I meant that I have seen picutures from Sony cams using the two sensors with the same exposure (I believe posted by dpr member RicksAstro). They were very similar. The quantitative measurements at DxOmark from a55 and a77 also seem to indicate that they would perform essentially the same and it does seem that the read noise is somewhat smaller in the a77 from your own sensorgen figures? At least in Sony's implementation of the sensor.

Actually, the Sensorgen figures show that the read noise of the 24MP is lower than the 16MP (about 2.5 e- per pixel as opposed to about 3 e- per pixel), which would make the 24MP about equal integrated read noise, as you suggest. It looks like my memory is playing tricks here. So, even that exception to the rule goes.

-- hide signature --

Bob

jonrobertp Forum Pro • Posts: 12,875
Re: New Canon on 22nd March?

According to B&H...The disappointing (by many) 60D still outsold the D7000 by about 35%...1258 to 932.   Interesting.

I'm guess that ratio won't change much...unless Canon misses some important detail in the 70D.

 jonrobertp's gear list:jonrobertp's gear list
Canon PowerShot G7 X Canon G3 X Panasonic ZS100
Rick880 Contributing Member • Posts: 509
Re: New Canon on 22nd March?

jonrobertp wrote:

According to B&H...The disappointing (by many) 60D still outsold the D7000 by about 35%...1258 to 932. Interesting.

Maybe it could be from upgrades by the vast Rebel user base, I guess.

I'm guess that ratio won't change much...unless Canon misses some important detail in the 70D.

I was disappointed by 60D's lacking of vertical electronic level and was seriously tempted by D7000, but eventually decided to stay put with XSi. I hope 70D will do it right but if 6D is any indicaton, it may not.

(unknown member) Senior Member • Posts: 2,853
Re: Call your bluf

Let me understand this.  We're in some kind of a competition now?  I think I finally see what your game is, which is to say you think this is some kind of a game.

To be clear, I made a comment several levels up which included the phrase "megapixel wars" and I was told that the term was, for some reason, objectionable.  I didn't make it up, it was used at the time by many people, and I think it was an expression of frustration more than anything else.  Now I'm supposed to justify that turn of phrase to someone who interjected himself into the conversation well after the fact?

I gave you an example.  One that wasn't all the time consuming to find and demonstrated some of my point.  I didn't go after others because following Canon's model numbers in the P&S area is next to impossible and I wanted to show how a model series changed.

If we're playing poker then don't get your cards and refuse to bet until I teach you how to play the game.  If people don't agree with the phrase I used I don't care any more.  It was used even by people like me buying a 10D, 30D, and then 7D to express our desire for a little less emphasis on MP count and more on noise response and usable high ISO.  The end.  I can't condense years of forum posts/trends and several manufacturers' spec changes here....or more to the point, I won't.

Good luck, Bob.  Do you have a few accounts on here or are there just mulitple cranky old men named Bob here?  Don't answer that, it was rhetorical.  I won't be engaging in conversation any longer with people who don't want to talk but instead are trying to win at a hand of poker.

robert1955 wrote:

After a reply like that, I'd say your hand only has low cards

howardroark wrote:

robert1955 wrote:

THat is still only one example, and I'm not sure there is much evidence that the image quality got better from reducing the pixel count.

Note that the most important differnece between the SX30 and 40 is that Canon changed form a CCD to a BSI-CMOS at the same time. That change means IMO that you can't validate your claim on this ititaration.


I have no desire to write a dissertation on the subject. I've paid close attention to camera technology since I bought my Sony S85 ages ago and even more closely after buying my 10D shortly after. I watched the camera reviews, read the forums, and while I don't wish to spend my time catching you up on all of the things I, and many others, saw over that time I'm certain some digging on your part (if that's how you choose to spend your time) will support my recollection of history.

bobn2
bobn2 Forum Pro • Posts: 60,889
Re: Call your bluf
3

howardroark wrote:

Good luck, Bob. Do you have a few accounts on here or are there just mulitple cranky old men named Bob here? Don't answer that, it was rhetorical. I won't be engaging in conversation any longer with people who don't want to talk but instead are trying to win at a hand of poker.

More speculation without evidence on your part. No, that Robert is not me. Perhaps it just means that there are more people than you imagine who find your arguments holey.

-- hide signature --

Bob

qianp2k Forum Pro • Posts: 10,350
Re: New Canon on 22nd March?

bobn2 wrote:

jonas ar wrote:

I meant that I have seen picutures from Sony cams using the two sensors with the same exposure (I believe posted by dpr member RicksAstro). They were very similar. The quantitative measurements at DxOmark from a55 and a77 also seem to indicate that they would perform essentially the same and it does seem that the read noise is somewhat smaller in the a77 from your own sensorgen figures? At least in Sony's implementation of the sensor.

Actually, the Sensorgen figures show that the read noise of the 24MP is lower than the 16MP (about 2.5 e- per pixel as opposed to about 3 e- per pixel), which would make the 24MP about equal integrated read noise, as you suggest. It looks like my memory is playing tricks here. So, even that exception to the rule goes.

It only means Sony 24mp sensors are designed later with better read noise but not because it has more MP. Sony could update its 16mp APS-C sensor to have even better read noise. I personally believe 24mp is marketing gimmick - a) it doesn't gain much resolution than 16mp on diminishing return; b) its pixel level IQ is pretty bad even at base ISO that largely makes irrelevant.

-- hide signature --
Keyboard shortcuts:
FForum MMy threads