The next Foveon camera?

Doppler9000 wrote:

tarmov wrote:

Doppler9000 wrote:

tarmov wrote:

Doppler9000 wrote:

Second, the insistence on analyzing sensors at the pixel level makes no sense. Bayer sensor designers are using smaller pixels both to increase resolution and to reduce aliasing. These advances are ignored in the “compression ratio” analysis, hopelessly skewing the results.

1:1 viewing matters on any screens, because that avoids interpolation, which is essentially weighted averaging, which is essentially smoothing.

-Tarmo

So what?

What if you crop an image?

In most cases amateurs don't do that.

Really? You ignored the cropping question.

If an image is cropped, your perfect pixel ratio disappears.

The primary way for amateurs to "consume" their images is looking at a screen.

Printed 10x15cm photo albums are ancient history (except for the older generation).

I understand that you don’t believe cell phones have had an impact on the camera market, but at 10x15cm, a modern phone would be plenty for most people.

I'm sorry, why do you think smartphone cameras keep getting better and better? Why do you think that the best smartphones (or at least the most expensive, flagship modeks) actually have three or four cameras on the back, as well as the one on the front? Why do you think there are so many smartphones with 50 MP  cameras now, rather than the same old 8MP, 12 MP, or 16 MP cameras, which could capture photos that easily fill the screen of a smartphone, with plenty of fine detail? Do you think it's because people don't want more detail in their photos? Samsung's S23 Ultra has a 104 MP camera, if I'm not mistaken. It sells for around $1,000 or more, and while I acknowledge Samsung and all the smartphone companies upgrade their phones in many ways, such as increasing RAM and processor speed, improving the screens, and adding 5G capability, the cameras do indeed help sell them. I know a professional photographer that bought a Samsung with that hundred megapixel camera, because she just wanted a phone with a better camera.

-Tarmo

Roger wrote:

Greetings Sigma Crew

As I thought more and more about this, for me all I need is a SDQ-H with a L mount and improved supper fast AF, I need speed. ISO, I'm good where I'm at.

OK, but definitely NOT in an fp-type body!

Here's an idea, use what you have, I want a L mount Sigma that automatically switches between viewfinder, and LCD , a flip up LCD not swing out. I hope it's a FF Foveon or a 1.2 crop factor with a build in flash like a Leica TL or Sigma DP2x, and both mechanical and electrical shutter or a electrical shutter like the Nikon Z9. Also weather seal.

Oh ya, a price around $1000 USD

Ok, Sigma get back to the drawing board.

Have fun

Roger j.

SigmaChrome wrote:

Roger wrote:

Greetings Sigma Crew

As I thought more and more about this, for me all I need is a SDQ-H with a L mount and improved supper fast AF, I need speed. ISO, I'm good where I'm at.

OK, but definitely NOT in an fp-type body!

I think any Foveon-style sensor will need a traditional shutter, so it can't be in an fp style body.

An sdQH with L mount instead of SA mount would be very practical. But I suppose that current fashion means it has to have a 24x36mm sensor.

Here's an idea, use what you have, I want a L mount Sigma that automatically switches between viewfinder, and LCD , a flip up LCD not swing out. I hope it's a FF Foveon or a 1.2 crop factor with a build in flash like a Leica TL or Sigma DP2x, and both mechanical and electrical shutter or a electrical shutter like the Nikon Z9. Also weather seal.

Oh ya, a price around $1000 USD

Ok, Sigma get back to the drawing board.

Have fun

Roger j.

Don

D Cox wrote:

SigmaChrome wrote:

Roger wrote:

Greetings Sigma Crew

As I thought more and more about this, for me all I need is a SDQ-H with a L mount and improved supper fast AF, I need speed. ISO, I'm good where I'm at.

OK, but definitely NOT in an fp-type body!

I think any Foveon-style sensor will need a traditional shutter, so it can't be in an fp style body.

An sdQH with L mount instead of SA mount would be very practical. But I suppose that current fashion means it has to have a 24x36mm sensor.

Here's an idea, use what you have, I want a L mount Sigma that automatically switches between viewfinder, and LCD , a flip up LCD not swing out. I hope it's a FF Foveon or a 1.2 crop factor with a build in flash like a Leica TL or Sigma DP2x, and both mechanical and electrical shutter or a electrical shutter like the Nikon Z9. Also weather seal.

Oh ya, a price around $1000 USD

Ok, Sigma get back to the drawing board.

Have fun

Roger j.

Don

Considering somebody did convertion of dp1 to mlc and fit shutter in it I would like FFF in fp body but of it is of no consecuence as I can't afford anything unless open miracle happens 🤔 But one can hope...

Doppler9000 wrote:

tarmov wrote:

Doppler9000 wrote:

tarmov wrote:

Doppler9000 wrote:

Yes, but the compression ratio measure is easier as a crude measure.

How is comparing a camera’s resolution to an arbitrary jpeg easier than just looking at the resolution?

If you are looking at a 1:1 zoom on a fixed resolution screen, then the compression ratio measures the reduced ratio of information (and noise) per output pixel.

3.7x less information than the hypothetical 24-bit maximum.

And 2.2x less information.

If that 3.7x holds as an average compression ratio for average images from that Canon camera, then one could argue that Sigma DP2m records about 1.6x more information (and noise) per output pixel.

This is a new metric to me, so please correct any errors I have made.

As a photographer, aren’t you more interested in total image resolution, or resolution per image height? Measures of per-pixel information ignore the adavantages of higher pixel counts.

Higher pixel counts are irrelevant when looked at a fixed resolution screen.

Higher pixel counts may become useful at 4x resolution or 9x resolution, because in those cases there would be exactly 2x2 or 3x3 pixel binning.

Screens have a resolution of 2,5k or 4k or 5k or 8k. Ok, there might also be 6k screen oddities.

Professionals who have to make large prints for customers may benefit from intermediate resolutions, but non-professionals are better with the Merrill 5k or with a potentially upcoming 8k Foveon sensor. Merrill APS-C upscaled to the full frame would be more or less at 8k (28-34MP), depending if the new sensor keeps the 3:2 sides ratio or goes with the 16:9 sides ratio. Such an 8k sensor would be competitive against 40-50MP Bayer sensors within Foveon's prior niche market segment.

Sigma would have to model and account for both market segments: professionals and non-professionals.

At higher ISO levels, the noise to signal ratio of the Foveon will increase relative to the Bayer sensor.

Within the Foveon niche market I would rank the relative importance of criteria the following way:

1. relative image quality (microcontrast, then noise, then colors). New Foveon cameras may continue at ISOs comparable to Merrill and Quattro, or perhaps hopefully improve that by up to one stop.

2. relative speed

3. relative battery longevity

(*) relative to how Merrill performed against its competitors and / or how Quattro performed against its competitors.

Whether Sigma manages to cram the in-camera sensor stabilisation or not, well, having it would certainly be a bonus.

The decision model composed here would inevitably be hopelessly incomplete and thus would provide very little community value besides giving false impressions.

How have you determined this? If the best case for the FFF is trivial sales volumes, it might be the case that it will have been a poor investment, your fact-free assertion otherwise notwithstanding.

All Sigma has to do is achieve comparable success as it did with Merrill and Quattro. That would guarantee the continuation of Foveon niche market. Whether such sales volumes are trivial or not, that is for Sigma to decide.

My point with the 'hopelessly incomplete decision models' was that a complete group decision model would have to get representative samples (individual decision models from each sample) from all relevant market segments - full decision models from individuals. You can't compose such a group decision model at the level of criteria, it won't work. The group decision has to be made at the level of alternatives.

Therefore such free form discussions as done in this forum are rather pointless. It only manages to map generally relevant criteria and perhaps relevant scenarios, no more.

Yes, Foveon relative image quality is important, likely the most important criterion. But that is not enough to draw conclusions on the final outcome of the group decision model.

The most plausible approach would be to treat Merrill and Quattro cameras as a relative baseline against competition. And either try to repeat that relative success or try to improve a bit, relatively.

I also doubt that the profit margins are much different from the past when digital cameras were in its infancy and production volumes were small.

Based on what?

Based on ceteris paribus.

This means “other things equal”. You don’t believe that the camera market has changed significantly over the past 25-30 years, and can’t understand how, for example, the iPhone et al have exerted significant competitive pressure on camera makers?

Merrills came out in 2012. Less than 11 years ago.

Not much has changed since then.

This reminds me of a line from Spinal Tap. In describing Boston, the manager said, “Not much of a college town”.”

Merrills did not have superior numerical resolution back then. But it did have excellent image quality at 5k resolution at ISO100 and ISO200, at least for a small subset of photographers.

All that is needed is to achieve relatively similar success. And that is doable with an 8k full frame sensor.

Based on your premise that the camera market is unchanged since 2012?

-Tarmo

Well for Sigma it likely hasn't changed much. The Quattro cameras sold, with relatively few drops in price. In fact, I believe the SD Quattro H sold out in less time than the others, even though it was the last to be introduced and the most expensive of the Quattros by hundreds of dollars. Ultimately Sigma raised the price to more than the introductory price, even though it was an "old" model by the time they were nearly sold out. The Quattros are almost all gone at this point, and there's nothing to replace them. Sigma should make something to replace them, but in L mount . . . in my opinion. In fact, I would buy one if they did, and the prices were just a little more than they used to be, even if they look and perform the same. No doubt though, Sigma woukd change components, like the processor, and the new cameras would operate faster, have a better viewfinder, and auto-focus better, and more quickly. With a faster processor and memory card slot, they could clear their buffer faster, and Sigma might decide to finally add wi-fi. If the original SD Quattro and SD Quattro H had L mounts instead of SA mounts they might have sold out in half the time, making them even more successful than they were.

Today cameras do 4K video, and most new cameras (all?) are mirrorless, unlike back when the Quattros were introduced, and both Nikon and Canon were introducing new DSLRs every year. So yeah, the camera market has definitely changed. Most, if not all new cameras introduced iver the past year have IBIS. Even medium format cameras have IBIS these dats. That was a feature unheard of in the medium format world just a few years ago.

But does Sigma need to include such features in their new cameras to sell them? As we know, the SD Quattro H sold out relatively quickly for a Sigma camera, and it didn't have many of the new features of modern cameras at that time, like tilt screens, wi-fi, and video capability. So maybe Sigma can just keep making their simple, high quality cameras, with their excellent sensors, and more and more photographers will buy them, even if they don't gave IBIS and can't shoot video - even if they can't shoot fast, focus on birds in flight, or be controlled using wi-fi. Maybe they would still sell, just because their image quality at low ISO settings is amazing.

D Cox wrote:

SigmaChrome wrote:

Roger wrote:

Greetings Sigma Crew

As I thought more and more about this, for me all I need is a SDQ-H with a L mount and improved supper fast AF, I need speed. ISO, I'm good where I'm at.

OK, but definitely NOT in an fp-type body!

I think any Foveon-style sensor will need a traditional shutter, so it can't be in an fp style body.

What makes you think a Foveon sensor can't have an electronic shutter?

An sdQH with L mount instead of SA mount would be very practical. But I suppose that current fashion means it has to have a 24x36mm sensor.

Here's an idea, use what you have, I want a L mount Sigma that automatically switches between viewfinder, and LCD , a flip up LCD not swing out. I hope it's a FF Foveon or a 1.2 crop factor with a build in flash like a Leica TL or Sigma DP2x, and both mechanical and electrical shutter or a electrical shutter like the Nikon Z9. Also weather seal.

Oh ya, a price around $1000 USD

Ok, Sigma get back to the drawing board.

Have fun

Roger j.

Don

Doppler9000 wrote:

xpatUSA wrote:

Doppler9000 wrote:

xpatUSA wrote:

Doppler9000 wrote:

xpatUSA wrote:

tarmov wrote:

Doppler9000 wrote:

Couldn’t one look at resolution directly to see if resolution was superior?

Yes, but the compression ratio measure is easier as a crude measure.

A reasonable metric and much less vague than "resolution" ...

If Canons and Sonys have a compression ratio of 3.7, while Sigma DPm has a compression ratio of 2.2 AND a person also visually prefers Sigma Foveons then I'd say there is a strong case FOR THAT PERSON to prefer Sigma Foveons.

... are we talking about the best possible out-of-camera JPEG quality, often called "fine"?

Given that “JPEG” isn’t a single standard, ...

What?

"Since its introduction in 1992, JPEG has been the most widely used image compression standard in the world,"

https://en.wikipedia.org/wiki/JPEG

how is a metric that uses the term less vague than “resolution”?

and tarmo was talking only about compression ratio

What is on the left and what is on the right if the JPEG compression ratio of a camera is 3.7?

"It all depends on what the meanin g of 'is' is" ...

Seems like a perfect performance metric for the FFF.

Agreed!

tarmov wrote:

Doppler9000 wrote:

and tarmo was talking only about compression ratio

What is on the left and what is on the right if the JPEG compression ratio of a camera is 3.7?

Compression ratio measures the amount of information (and recorded noise) per horizontal pixel that gets shown on a fixed resolution output surface such as a TV or a computer screen.

There are other definitions of what Compression does; so not everyone agrees with the above ...

Scottelly wrote:

Doppler9000 wrote:

tarmov wrote:

Doppler9000 wrote:

tarmov wrote:

Doppler9000 wrote:

Yes, but the compression ratio measure is easier as a crude measure.

How is comparing a camera’s resolution to an arbitrary jpeg easier than just looking at the resolution?

If you are looking at a 1:1 zoom on a fixed resolution screen, then the compression ratio measures the reduced ratio of information (and noise) per output pixel.

3.7x less information than the hypothetical 24-bit maximum.

And 2.2x less information.

If that 3.7x holds as an average compression ratio for average images from that Canon camera, then one could argue that Sigma DP2m records about 1.6x more information (and noise) per output pixel.

This is a new metric to me, so please correct any errors I have made.

As a photographer, aren’t you more interested in total image resolution, or resolution per image height? Measures of per-pixel information ignore the adavantages of higher pixel counts.

Higher pixel counts are irrelevant when looked at a fixed resolution screen.

Higher pixel counts may become useful at 4x resolution or 9x resolution, because in those cases there would be exactly 2x2 or 3x3 pixel binning.

Screens have a resolution of 2,5k or 4k or 5k or 8k. Ok, there might also be 6k screen oddities.

Professionals who have to make large prints for customers may benefit from intermediate resolutions, but non-professionals are better with the Merrill 5k or with a potentially upcoming 8k Foveon sensor. Merrill APS-C upscaled to the full frame would be more or less at 8k (28-34MP), depending if the new sensor keeps the 3:2 sides ratio or goes with the 16:9 sides ratio. Such an 8k sensor would be competitive against 40-50MP Bayer sensors within Foveon's prior niche market segment.

Sigma would have to model and account for both market segments: professionals and non-professionals.

At higher ISO levels, the noise to signal ratio of the Foveon will increase relative to the Bayer sensor.

Within the Foveon niche market I would rank the relative importance of criteria the following way:

1. relative image quality (microcontrast, then noise, then colors). New Foveon cameras may continue at ISOs comparable to Merrill and Quattro, or perhaps hopefully improve that by up to one stop.

2. relative speed

3. relative battery longevity

(*) relative to how Merrill performed against its competitors and / or how Quattro performed against its competitors.

Whether Sigma manages to cram the in-camera sensor stabilisation or not, well, having it would certainly be a bonus.

The decision model composed here would inevitably be hopelessly incomplete and thus would provide very little community value besides giving false impressions.

How have you determined this? If the best case for the FFF is trivial sales volumes, it might be the case that it will have been a poor investment, your fact-free assertion otherwise notwithstanding.

All Sigma has to do is achieve comparable success as it did with Merrill and Quattro. That would guarantee the continuation of Foveon niche market. Whether such sales volumes are trivial or not, that is for Sigma to decide.

My point with the 'hopelessly incomplete decision models' was that a complete group decision model would have to get representative samples (individual decision models from each sample) from all relevant market segments - full decision models from individuals. You can't compose such a group decision model at the level of criteria, it won't work. The group decision has to be made at the level of alternatives.

Therefore such free form discussions as done in this forum are rather pointless. It only manages to map generally relevant criteria and perhaps relevant scenarios, no more.

Yes, Foveon relative image quality is important, likely the most important criterion. But that is not enough to draw conclusions on the final outcome of the group decision model.

The most plausible approach would be to treat Merrill and Quattro cameras as a relative baseline against competition. And either try to repeat that relative success or try to improve a bit, relatively.

I also doubt that the profit margins are much different from the past when digital cameras were in its infancy and production volumes were small.

Based on what?

Based on ceteris paribus.

This means “other things equal”. You don’t believe that the camera market has changed significantly over the past 25-30 years, and can’t understand how, for example, the iPhone et al have exerted significant competitive pressure on camera makers?

Merrills came out in 2012. Less than 11 years ago.

Not much has changed since then.

This reminds me of a line from Spinal Tap. In describing Boston, the manager said, “Not much of a college town”.”

Merrills did not have superior numerical resolution back then. But it did have excellent image quality at 5k resolution at ISO100 and ISO200, at least for a small subset of photographers.

All that is needed is to achieve relatively similar success. And that is doable with an 8k full frame sensor.

Based on your premise that the camera market is unchanged since 2012?

-Tarmo

Well for Sigma it likely hasn't changed much. The Quattro cameras sold, with relatively few drops in price. In fact, I believe the SD Quattro H sold out in less time than the others, even though it was the last to be introduced and the most expensive of the Quattros by hundreds of dollars. Ultimately Sigma raised the price to more than the introductory price, even though it was an "old" model by the time they were nearly sold out. The Quattros are almost all gone at this point, and there's nothing to replace them. Sigma should make something to replace them, but in L mount . . . in my opinion. In fact, I would buy one if they did, and the prices were just a little more than they used to be, even if they look and perform the same. No doubt though, Sigma woukd change components, like the processor, and the new cameras would operate faster, have a better viewfinder, and auto-focus better, and more quickly. With a faster processor and memory card slot, they could clear their buffer faster, and Sigma might decide to finally add wi-fi. If the original SD Quattro and SD Quattro H had L mounts instead of SA mounts they might have sold out in half the time, making them even more successful than they were.

Today cameras do 4K video, and most new cameras (all?) are mirrorless, unlike back when the Quattros were introduced, and both Nikon and Canon were introducing new DSLRs every year. So yeah, the camera market has definitely changed. Most, if not all new cameras introduced iver the past year have IBIS. Even medium format cameras have IBIS these dats. That was a feature unheard of in the medium format world just a few years ago.

But does Sigma need to include such features in their new cameras to sell them? As we know, the SD Quattro H sold out relatively quickly for a Sigma camera, and it didn't have many of the new features of modern cameras at that time, like tilt screens, wi-fi, and video capability. So maybe Sigma can just keep making their simple, high quality cameras, with their excellent sensors, and more and more photographers will buy them, even if they don't gave IBIS and can't shoot video - even if they can't shoot fast, focus on birds in flight, or be controlled using wi-fi. Maybe they would still sell, just because their image quality at low ISO settings is amazing.

The fewer the features the better for me ... in fact a FF "SD9" would get my vote