G1X's sensor size

The rumoured 1.5" sensor has one of the 2 dimensions:

4:3 => Width = 19.2mm; Height = 14.4mm
3:2 => Width = 19.7mm; Height = 13.3mm

Either case has the same crop factor of 1.8x, just in-between APS and m43. I'm not affiliated with Canon, nor w/ any NDA party related to Canon. Mitch

... or close enough! Plus, I'm beginning to buy into the G1X rumor as fact. When the rumor first came out, I believed it then changed my mind after reading the many and various interpretations and speculations. But now I accept the rumor as fact and as written. Although a strange name (G1X ?), it would appear to be a well spec'd camera and more or less similar to the G I've been "predicting" and wishing for these past couple of years, http://lightdescription.blogspot.com/2011/12/s100-review-new-g-coming.html .

migus wrote:

The rumoured 1.5" sensor has one of the 2 dimensions:

4:3 => Width = 19.2mm; Height = 14.4mm
3:2 => Width = 19.7mm; Height = 13.3mm

Either case has the same crop factor of 1.8x, just in-between APS and m43. I'm not affiliated with Canon, nor w/ any NDA party related to Canon. Mitch

I am thinking on higher end camera they are going with 3:2 to fit better with DSLRs.

Also your 4:3 calculation is perfect. But your 3:2 is off a bit, it is less than 24 mm diagonal, and not quite 3:2. Width should be closer to 20.

I am thinking on higher end camera they are going with 3:2 to fit better with DSLRs.

If i were Canon's product manager for G1X, I'd choose 4:3 as easier to defend in the exec reviews:
1) higher yields => lower cost;

2) slightly smaller lens to cover the width => compacter & simpler lens/sensor design;
3) visible differentiator from 3:2 sensors => less encroaching on dSLR's turf;
4) comparable, yet larger than all P&S, CX and m43 => apex of the 4:3 turf;

5) vertical dimension nearly equal to APS-C sensors' => exceeds m43 and competes within 95% with APS-C (100% of Foveon's X3) in photo tests and reviews.

For practical purposes, the IQ of this sensor should be virtually indistinguishable from that of an APS-C from NEX, NX or Fuji X-Pro. This is my conjecture :-).

Also your 4:3 calculation is perfect. But your 3:2 is off a bit, it is less than 24 mm diagonal, and not quite 3:2. Width should be closer to 20.

Good catch, Peter: Indeed, W = 19.97 ~ 20mm ; H = 13.31mm.

As per sensor and filters type, one can assume it's going to be a new technology, probably serving as test pilot for Canon's APS and FF lines.

The Canon G12 is naturally 4:3 and offers crops to get to 3:2 and 16:9; I bet the G1X does the same.

migus wrote:

If i were Canon's product manager for G1X, I'd choose 4:3 as easier to defend in the exec reviews:
1) higher yields => lower cost;

Why higher yield? The 4:3 part is approx 10 square mm larger. So if anything 3:2 should have higher yield/lower cost. So I think you have this one backwards.

2) slightly smaller lens to cover the width => compacter & simpler lens/sensor design;

Same diagonal, need essentially the same lens to have decent corner performance, so this one is a wash.

3) visible differentiator from 3:2 sensors => less encroaching on dSLR's turf;

More important to differentiate from compacts, with 4:3 sensors aimed at old 4:3 monitors/TVs. Also more important to match DSLR to serve as decent backup camera, so you wont have mixed ratios when you want to just carry your smaller camera.

4) comparable, yet larger than all P&S, CX and m43 => apex of the 4:3 turf;

5) vertical dimension nearly equal to APS-C sensors' => exceeds m43 and competes within 95% with APS-C (100% of Foveon's X3) in photo tests and reviews.

As far as size goes, this is still larger than m43, CX (and all compacts obv.) in 3:2.

Basically size is nitpicking. This is bigger than m43/CX and smaller than APS-C. The size is good and it will be compared.

As far as the ratio goes, I see it making more sense to go pro and match DSLR rather than stick with 4:3 that was designed to match last generation TVs/Monitors.

I think Canons answer to 4:3 is 3:2 both in ratio and size and IMO it is superior in both.

Vidicon's curse: optical format anachronism from 1947...
see http://forums.dpreview.com/forums/readflat.asp?forum=1010&thread=40237186

Joe Ogiba wrote:

1.5" =38.1mm so how is it 1.5" ?

Why higher yield? The 4:3 part is approx 10 square mm larger. So if anything 3:2 should have higher yield/lower cost. So I think you have this one backwards.

waffers being round, the smaller and squarer chip dies have higher fill-in factors; one can calculate w/ some effort (CAD tool) how many effective 4:3 vs. 3:2 chips would fit a 30"

As far as the ratio goes, I see it making more sense to go pro and match DSLR rather than stick with 4:3 that was designed to match last generation TVs/Monitors.

I think Canons answer to 4:3 is 3:2 both in ratio and size and IMO it is superior in both.

I also prefer the 3:2; however, business- and marketing-wise the mind set is entirely different. I hope you're right, though i fear that won't be the case Mitch

Joe Ogiba wrote:

1.5" =38.1mm so how is it 1.5" ?

http://en.wikipedia.org/wiki/Video_camera_tube#Size

1" = 16mm

Diagonal is really 1.5" X 16 = 24mm.

A convenient way to exaggerate sensor size, while still telling the "truth".

migus wrote:

waffers being round, the smaller and squarer chip dies have higher fill-in factors; one can calculate w/ some effort (CAD tool) how many effective 4:3 vs. 3:2 chips would fit a 30"

That might be true (to an infinitesimal degree) if they were the same area, but the 4:3 chip is 10 square mm larger and that will make it complete reverse in this case.

You can check it with an only wafer fit calculator:

http://mrhackerott.org/semiconductor-informatics/informatics/toolz/DPWCalculator/Input.html

4:3 chip will net you about 10 chips less, than 3:2 on each 300mm wafer.

I also prefer the 3:2; however, business- and marketing-wise the mind set is entirely different. I hope you're right, though i fear that won't be the case

Your business assumption about yield was faulty.

I think your marketing assumption is wrong also. 4:3 was popular with average consumers when they had 4:3 TVs and 4:3 monitors, now it is all 16:9 and I expect they would prefer wider aspect on cameras to match. I would bet in a few years consumer compacts migrate to 16:9.

yup, i was wrong about the yield of a 300mm; i should've calculated, since i had the figures. Let's wait and see Canon's decision; i hope you're right, and it's not a 4:3 as i think.

Peter G wrote:

migus wrote:

waffers being round, the smaller and squarer chip dies have higher fill-in factors; one can calculate w/ some effort (CAD tool) how many effective 4:3 vs. 3:2 chips would fit a 30"

That might be true (to an infinitesimal degree) if they were the same area, but the 4:3 chip is 10 square mm larger and that will make it complete reverse in this case.

You can check it with an only wafer fit calculator:

http://mrhackerott.org/semiconductor-informatics/informatics/toolz/DPWCalculator/Input.html

4:3 chip will net you about 10 chips less, than 3:2 on each 300mm wafer.

I also prefer the 3:2; however, business- and marketing-wise the mind set is entirely different. I hope you're right, though i fear that won't be the case

Your business assumption about yield was faulty.

I think your marketing assumption is wrong also. 4:3 was popular with average consumers when they had 4:3 TVs and 4:3 monitors, now it is all 16:9 and I expect they would prefer wider aspect on cameras to match. I would bet in a few years consumer compacts migrate to 16:9.

migus wrote:

The rumoured 1.5" sensor has one of the 2 dimensions:

4:3 => Width = 19.2mm; Height = 14.4mm
3:2 => Width = 19.7mm; Height = 13.3mm

Or it could be both 4:3 and 3:2, and maybe 16:9 and 1:1 too, if it's an 'over-sized' multi-aspect ratio sensor, like used in e.g. the Panasonic GH2.

migus wrote:

yup, i was wrong about the yield of a 300mm; i should've calculated, since i had the figures. Let's wait and see Canon's decision; i hope you're right, and it's not a 4:3 as i think.

Note that even the smaller Nikon CX format cameras (J1,V1) have 3:2 sensors. It has a crop for 16:9 but not 4:3.

4:3 is dying everywhere. I am 99% certain canon will go with 3:2 if this is a big sensor.

Peter G wrote:

migus wrote:

yup, i was wrong about the yield of a 300mm; i should've calculated, since i had the figures. Let's wait and see Canon's decision; i hope you're right, and it's not a 4:3 as i think.

Note that even the smaller Nikon CX format cameras (J1,V1) have 3:2 sensors. It has a crop for 16:9 but not 4:3.

4:3 is dying everywhere. I am 99% certain canon will go with 3:2 if this is a big sensor.

Sounds good to me. 4:3 is an shape for landscapes, and cropping constantly gets annoying.

PeterNMIF wrote:

To be a G-series camera it is possible that they went slightly bigger than G12, with a 1/1.5" sensor (that would be 2/3", the size of the new Fuji X10). But to have a sensor bigger than m43 and start at F2.5 and be equiv to 28-112mm could weigh about 1 lb. (just a guess). I'd love to see this even at that weight, but the camera/lens combo would be so much larger than the current and past G-cameras that it would probably be the start of an entirely new series of cameras not starting with the letter "G".

Well it is a new series. Hence the G1x. It won't be F2.5 throughout the range.

It will be F2.5- F5.0 or something like that.

Also bear in mind that with a built in lens they can also remove optical correction and do it digitally, thus simplifying lens design.

PeterNMIF wrote:

To be a G-series camera it is possible that they went slightly bigger than G12, with a 1/1.5" sensor (that would be 2/3", the size of the new Fuji X10). But to have a sensor bigger than m43 and start at F2.5 and be equiv to 28-112mm could weigh about 1 lb. (just a guess).

The current G12 weighs 3/4 lbs.

Given that the comparable DSLR kit weighs just under 2 lbs, 1 lbs sounds pretty decent.