[parsons]

square sensor would be very strange. it would make composition difficult to most people who are used to a normal camera. almost every image would have to be cropped to be printed on normal paper. i don't think its practical.

[gazonk]

1. A square 35x35 sensor would be very different. No other brand has that. It could be the niche that Pentax needs. (Or at least: which they say they need.)

2. It would give the missing upgrade path to every Pentaxian. Not only the ones that already have FF glass, or enough money to invest in FF glass.

But don't you understand that a square 35x35 sensor requires a larger than FF image circle, so not even FF glass would be guaranteed to work?

It would require a 50mm image circle, which is closer to 645D (55mm) than 24x36 (43mm).

[eddie1960]

Just to clarify, the 50% rule is a general business school product/marketing rule for each level of distributorship increasing the cost of delivery. Doesn't mean mail order companies have to live by that rule.

But bear in mind that the K-5's original MSRP was $1600 and street price was that. It's now down to $1K street price roughly. Seems like roughly 50% ($1K+50%=$1500) if their margins are down now as they get ready to roll out a new body...

Business school rules rarely work when it comes to consumer electronics of any type. when the camera was $1600 the dealer cost would have been higher as well as Pentax would front load the cost to recoup it earlier (which is where the idea of an early adopter premium comes from) dealers may have pulled an extra few points at the time but then again they probably sold very few at that price As time passes the first drops come out of the retailers pocket then the manufacturer drops but the dealer holds for a while, then the competition sets in again and the dealer drops and so on. probably the current cost would allow 1049 or so with the full normal points and 955 or so the dead minimum (which is the lowest price i've seen)

the Business school model does work with other big ticket industries like furniture carpeting jewelry, clothing retail when not on sale actually works to even higher initial margins but they erode quickly (with a 6 - 8 week fashion cycle) - a big chunk of my 25 years was at department store retail (Eatons) so i got a good feel for the broader spectrum of the business. Typically we worked to the largest margin possible on CE but part of the job of course is tracking the competition and adjusting for the market. when i started margins for electronics were 7-10% higher on average in the market than they are now. cutthroat price wars long term lowered the average margin permanently and still people cut their own throats on price (TV is a really good example, in the early 80's we averaged 25% GM, by the early 90's it was 20%, by 2000 it was around 16%, currently from talking with people still in the business it is 10% or less (hence all the bundle deals you see where they are trying to bring up the average GM with higher margin Audio and accessories)
Computers are even worse at least half the systems we sold at key times like back to school were at or below retail before accessory sales and warranty/services were added in. Selling an antivirus with an itme could sometimes double the total GM

Cameras aren't any different than those models. In fact since the advent of digital camera bodies have recovered GM, at one point in film days they too were less than 10% before lenses

[eddie1960]

But don't you understand that a square 35x35 sensor requires a larger than FF image circle, so not even FF glass would be guaranteed to work?

It would require a 50mm image circle, which is closer to 645D (55mm) than 24x36 (43mm).

Exactly the 35mm image circle would not be big enough (anything wider than a FA 50-77 would vignette i would guess and DA lenses would almost all vignett (except the ones we know are essentially FA lens reworks like the DA* 200)

I love the idea of a 35mm square format sensor myself for the reasons objected to above.(need for cropping) you can print square images at full res and crops at FF in either orientation,. this was on of the big advantages of 6x6 film over 645, you had far more latitude for how you printed. at worst you cropped to 645

[Clavius]

My apologies! My enthusiasm got the worst of me. I meant a 30x30 sensor of course. I think all this talk about FF and 35mm got me got me confused.

[gazonk]

My apologies! My enthusiasm got the worst of me. I meant a 30x30 sensor of course. I think all this talk about FF and 35mm got me got me confused.

Ok, but even with 30x30 you still need the same (*) image circle as for 24x36!

(*) in order to avoid nit picking: You can do with 0.84mm (diameter) smaller circle

[maxfield_photo]

square sensor would be very strange. it would make composition difficult to most people who are used to a normal camera. almost every image would have to be cropped to be printed on normal paper. i don't think its practical.

I'm not a huge proponent of the square image sensor idea, mainly because if it were to be done with the K mount, it would mean giving up about 30% of the sensor real estate off the bat, and then you'd have to crop to the desired aspect ratio from there. However, I end up cropping 90%+ of my images as it is. I work with a lot of models who need 8x10 images, and therefore my own portfolio is 8x10 as well.

So 8x10 is a 5:4 aspect ratio, my camera shoots 3:2, so if my math serves me I end up chopping about 17% off the long dimension of the image. Now I've learned to shoot wider to try and compensate, but it's an uphill battle. I was trained for so long to "get it right in camera", that sometimes I forget, or get excited, or just under estimate how much 17% is. I end up with a great shot that is totally unusable, at least without a lot of cloning in photoshop. Nothing is more frustrating. A 5:4 sensor would tickle me pink, failing that however a mini 645 (4x3) would be great. I'd only have to chop about 6% off the long dimension.

I would really like a WYSIWYG viewfinder. That would be unique. You tell the camera, "Ok, I'm going to output these shots to 8x10," and little crop masks just behind the diopter slide into the correct aspect ratio. "Ok, now I want to shoot a 16x9 landscape," and boom, the sliders show you exactly what will be in your print. "Now I want to mount this DA lens." and the VF would show you an APS-C sized image, but you'd still have the ability to specify the aspect ratio.

---

Last edited by maxfield_photo; 03-06-2012 at 02:59 PM.

[ElJamoquio]

I'm not a huge proponent of the square image sensor idea, mainly because if it were to be done with the K mount, it would mean giving up about 30% of the sensor real estate off the bat, and then you'd have to crop to the desired aspect ratio from there. However, I end up cropping 90%+ of my images as it is. I work with a lot of models who need 8x10 images, and therefore my own portfolio is 8x10 as well.

So 8x10 is a 5:4 aspect ratio, my camera shoots 3:2, so if my math serves me I end up chopping about 17% off the long dimension of the image. Now I've learned to shoot wider to try and compensate, but it's an uphill battle. I was trained for so long to "get it right in camera", that sometimes I forget, or get excited, or just under estimate how much 17% is. I end up with a great shot that is totally unusable, at least without a lot of cloning in photoshop. Nothing is more frustrating. A 5:4 sensor would tickle my pink, failing that however a mini 645 (4x3) would be great. I'd only have to chop about 6% off the long dimension.

I would really like a WYSIWYG viewfinder. That would be unique. You tell the camera, "Ok, I'm going to output these shots to 8x10," and little crop masks just behind the diopter slide into the correct aspect ratio. "Ok, now I want to shoot a 16x9 landscape," and boom, the sliders show you exactly what will be in your print. "Now I want to mount this DA lens." and the VF would show you an APS-C sized image, but you'd still have the ability to specify the aspect ratio.

I've been pushing the 4:3 for a while, myself. I guess I could learn what 17% is. Or find some focus screen that has lines marked... but a 4:3 sensor would solve those issues.

Unfortunately I think Pentax doesn't have the volume to support Sony making a new sensor like that...?

[eddie1960]

I've been pushing the 4:3 for a while, myself. I guess I could learn what 17% is. Or find some focus screen that has lines marked... but a 4:3 sensor would solve those issues.

Unfortunately I think Pentax doesn't have the volume to support Sony making a new sensor like that...?

why does it have to be Sony...they are hardly the only sensor vendor.
the problem is really the lack of volume for the sensor which would make the custom fab cost prohibitive (unless it goes in more than one body). I wonder if a 4:3 bigger than apsC can be cut without the stitching FF needs. then it could be spread across the mid and high lines in mirrorless and DSLR leaving apsc for the entry models
then it could be cost effective

[Clavius]

I wonder if a 4:3 bigger than apsC can be cut without the stitching FF needs.

That's what made the APS-H cheaper, didn't it?

[ElJamoquio]

why does it have to be Sony...they are hardly the only sensor vendor.

I forgot to put the 'et al' in there. Still it seems obvious that it doesn't have to be Sony, that the economics are similar for any chip manufacturer to produce a unique chip for 'only' Pentax, and the point is that Pentax isn't making the chip for themselves...?

[redrockcoulee]

When we owned a retail business in the 90s the mark up on Pentax binos was more like 15%. Books are generally 40% and t-shirts and cards etc were the 50%. There were a couple of very inexpensive items that were greater than 50% too.

In the 70s we bought a Pentax system from a retail store at employee discount and the second system a little later from the distributor at wholesale price as my wife was a clerk at a large camera store and got to know the rep. It was far from a 50% savings on either the body or the lenses.Also my former employer bought from one of the major companies directly and although the savings were large on a brand newly released product again the savings was far short of 50%. Closer to 25% and this was less than 7 years ago.

For the retail prices to drop there must be a drop at the wholesale level as well.

[RonHendriks1966]

That's what made the APS-H cheaper, didn't it?

And there it is again!

Well Falk told somewhere here that stitching isn't the big part in productioncost anymore. No idea about how true it is or how much it still contributes to the costs, but with a growing production in FF sensors (for D800) this does make sence.

I stated somewhere that I'm in favor for APS-H and that a Large 4:3th is one of the options.

Making a 4:3th large as possible sensor then probably the cheapest way is (when on special order) to cut it from the same wafer as where K-5 sensor is coming from. Would make a huge sensor with like 7264x5440 pixels (the same count as for 645D) and a size of 34,50 x 25,84 mm and a diagonal of 43,11mm and that will be good for FF lenses.

That would make a mini 645D with K-mount. But that would also make a need for a new 645D.

[eddie1960]

And there it is again!

Well Falk told somewhere here that stitching isn't the big part in productioncost anymore. No idea about how true it is or how much it still contributes to the costs, but with a growing production in FF sensors (for D800) this does make sence.

I stated somewhere that I'm in favor for APS-H and that a Large 4:3th is one of the options.

Making a 4:3th large as possible sensor then probably the cheapest way is (when on special order) to cut it from the same wafer as where K-5 sensor is coming from. Would make a huge sensor with like 7264x5440 pixels (the same count as for 645D) and a size of 34,50 x 25,84 mm and a diagonal of 43,11mm and that will be good for FF lenses.

That would make a mini 645D with K-mount. But that would also make a need for a new 645D.

the need for a new 645D is already there with the D800 to some degree (I know apples to oranges and all). a FF 645D would be amazing (even if the MP didn't change , but likely it would be 50MP or 60MP

A mini 645D would definitely meet the unique FF concept though

[Aristophanes]

why does it have to be Sony...they are hardly the only sensor vendor.
the problem is really the lack of volume for the sensor which would make the custom fab cost prohibitive (unless it goes in more than one body). I wonder if a 4:3 bigger than apsC can be cut without the stitching FF needs. then it could be spread across the mid and high lines in mirrorless and DSLR leaving apsc for the entry models
then it could be cost effective

Sony's Kyushu facility is the world largest sensor fab facility. It is pumping out something like 60% of the world's stock.

To meet price points, you need volume. Otherwise you are into Leica/645D territory. A custom fab for Pentax would be cost prohibitive and would take considerable time to scale up.

Also, CMOS has elements of IP that are base on licensing. If you want low-light CMOS, you need to talk to Sony right now. I am not sure Ricoh/Pentax can design a fab like this.

APS-H would give such a marginal increase in IQ compared to APS-C, at 2x the cost, that it's not worth the effort. Canon did that to get the price down for a telecentric journalists camera. An APS-H model would run smack into the D700/720 price problem.