dp0 vs fp L

Gate bois · Mar 5, 2022

Abbazz said:
Iain G Foulds said:

… Would love to see a third SD14/15 image in there. Perhaps one day…

Click to expand...

Here are two images taken (X3F raw) this morning with the same lens, the first one with my sd Quattro H and the other one with my SD14:

sd Quattro H

SD14

If you want to play with them, here are links to the raw files:

http://www.lallement.com/images/SDIM1280.X3F

http://www.lallement.com/images/SDIM8279.X3F

Cheers!

Abbazz

the rendering is very close on these two examples.

Apart from the sky and the buildings in the background where the difference is more visible.

Moreover, with the quattro, the photo gives a better feeling of depth, 3D, clarity.

On the photo of the SD14, I find that there is a slight cyan cast. The buildings in the background seem almost glued to the sky, whereas on the quattro they stand out much better from the sky, which brings more depth of relief and clarity to the photo.

Likewise for all the elements in the foreground, vegetation and buildings, the feeling of relief is more marked with the quattro. Looking at the colors in detail, the shades rendered are comparable for the two devices, except in one place where the quattro restores shades that are not visible on the SD14.

--
Paris - La Defense at 78.5Mpix:

http://www.noelbilly.eu/

Scottelly · Mar 6, 2022

paulsch said:
Tom Schum said:

I went out this morning to Great Falls with a tripod and both cameras.

The fp L has a TTartisan 21mm F1.5 manual lens on it.

Both were set to F8. Color palettes are different, but is the fp L better?

Basically I made minimal adjustments to the dp0 image and spent time working the fp L image to get it close to the color balance of the dp0 image.

fp L version, downsized to nearly match the pixel count of dp0 lo-res. 5.6mb download. Developed from DNG using SPP.

dp0 lo-res. 5.3mb download. Developed from X3F in SPP.

In my mind, the colors out of the fp L look better to me.

Click to expand...

Personally, I clearly prefer the dp0 rendering. IMHO it is the usual differentiators: the dp0 image has the more realistic appearance, more 3D appearance, particularly in the water and some parts of the rocks.

Well then, I guess I better get an SD Quattro then, rather than a $2,000 fp L or an even more expensive Sony A7r IV.

--
Scott Barton Kennelly

Big Print Photos - Photography by Scott Barton Kennelly

Landscape and nature photography, photographic prints by Scott Barton Kennelly, an Australian photographer living in south Florida.

www.bigprintphotos.com

Tom Schum · Mar 6, 2022

Scottelly said:
paulsch said:

Tom Schum said:

I went out this morning to Great Falls with a tripod and both cameras.

The fp L has a TTartisan 21mm F1.5 manual lens on it.

Both were set to F8. Color palettes are different, but is the fp L better?

Basically I made minimal adjustments to the dp0 image and spent time working the fp L image to get it close to the color balance of the dp0 image.

fp L version, downsized to nearly match the pixel count of dp0 lo-res. 5.6mb download. Developed from DNG using SPP.

dp0 lo-res. 5.3mb download. Developed from X3F in SPP.

In my mind, the colors out of the fp L look better to me.

Click to expand...

Personally, I clearly prefer the dp0 rendering. IMHO it is the usual differentiators: the dp0 image has the more realistic appearance, more 3D appearance, particularly in the water and some parts of the rocks.

Click to expand...

Well then, I guess I better get an SD Quattro then, rather than a $2,000 fp L or an even more expensive Sony A7r IV.

A number of subtle difficulties would be eliminated if one shot in black and white. All the purists do this. Perhaps this will be my next great experiment.

--
Tom Schum
"Beware of taking advice from anonymous wise men." Quote from Anon.

Iain G Foulds · Mar 6, 2022

… We really need a decent three way comparison- SD14/15, Quattro, and FP.

… Base ISO.

xpatUSA · Mar 6, 2022

Iain G Foulds said:
… We really need a decent three way comparison- SD14/15, Quattro, and FP.

We do ??

Iain G Foulds · Mar 6, 2022

… We do. Believe we would see a decline in the quality of the image with each model.

… Yes, Technical specs might be up. But, I believe that there will be a decline in the overall authenticity and quality of the image as a whole.

D Cox · Mar 7, 2022

xpatUSA said:
paulsch said:

xpatUSA said:

paulsch said:

Tom Schum said:

paulsch said:

With “color palette” of an image I wanted to say the number of different colours used in that image.

To be honest, I had not even tried that hyperlink. I think the linked site uses the term “color palette” for a somewhat arbitrary artistic collection of colour swatches which represents the overall impression of the image.

Click to expand...

Key word here is "arbitrary".

I've wondered this about the colorcheckers:

Datacolor SpyderCHECKR 24 Color Chart SCK24 B&H Photo Video (bhphotovideo.com)

Somebody picked a bunch of color swatches there, too.

There might be a history behind these things, and it might go back to colorimetry research years ago; I don't know.

Maybe a creative selection of color swatches would give Foveon colors a big advantage over Bayer. Maybe the existing selection is based on old color TV science, and that goes back to red, green, and blue phosphors on a TV picture tube. If so, it's a perfect fit to Bayer technology.

Click to expand...

I do not quite understand this.

Maybe we use the term differently. What I tried to say how I understand the term “colour palette” of an image is this:

The Foveon sensor has its data saved to three channels with 14 bits of information, i.e. a theoretical maximum color palette of 2^(3*14) or 4.398.046.511.104 different colours.

Click to expand...

Going a little deeper to avoid mis-information, Paul, all Sigma cameras before Quattro have 12-bit ADCs. And all ADC's are not perfect - for example the ones in my SD9 are about 10.5 'effective number of bits' (ENOB) as specified by Analog Devices. Then, the only pre-Quattro Foveon that saturates it's 'raw' data close to 2^12=4095 is the Merrill - all others vary quite considerably - for example the SD15 at 100 ISO saturates at around 3500 "raw" values and the SD14 saturates at around 6700 "raw" values - in other words, some different scaling goes on in each camera before writing to the card.

xpatUSA said:

And every real picture taken with the sensor contains a certain subset of those available colours.

Click to expand...

Agreed.

xpatUSA said:

And of course once the Raw data has been transformed into a different picture format (e.g. an sRGB jpeg) the theoretical maximum color palette and also the subset that is really used in the image, changes.

Click to expand...

Agreed.

xpatUSA said:

So with “colour palette” I mean the subset of the theoretical available colour values that is in use in the file.

Click to expand...

Understood.

Click to expand...

Click to expand...

Another way to look at that is from an artist's point of view where, for example, he places on his palette red, yellow and blue paints - just three colors plus black and white. Each of the five blobs of paint contains so many moles ( https://en.wikipedia.org/wiki/Mole_(unit) ). If we were to calculate the number of possible mixtures based on mole ratios then the subset of the theoretical available colour values that could be used in the painting is almost infinity!

xpatUSA said:

Thanks for the clarifying info!

Click to expand...

You are welcome!

The only painter I know of who used red, yellow and blue is Mondrian in some of his late paintings. He was under the delusion that they were The Primary Colours and by using them he would get nearer to the absolute. There's no mixing in those paintings. Gerrit Rietveld painted his famous chair in those colours, and there are no doubt other members of the De Stijl group who tried the same idea.

A better analogy might be the range of colours available in colour printing as mixtures of CMYK on white paper. This is indeed very large.

See also James Gurney on limited palettes.

Limited Palettes

Every Sunday I’ve been sharing some thoughts about color, and today I want to touch on limited palettes. When we were in grade school we all...

gurneyjourney.blogspot.com

Don

Abbazz · Mar 7, 2022

D Cox said:
The only painter I know of who used red, yellow and blue is Mondrian in some of his late paintings. He was under the delusion that they were The Primary Colours and by using them he would get nearer to the absolute. There's no mixing in those paintings. Gerrit Rietveld painted his famous chair in those colours, and there are no doubt other members of the De Stijl group who tried the same idea.

Nicolas Poussin was the leading painter of the classical French Baroque style. He is also famous because almost all of his paintings contain a red, a blue and a yellow piece of fabric:

Holy Family

Death of Germanicus

The **** of the Sabine Women

Cheers!

Abbazz

Tom Schum · Mar 7, 2022

Iain G Foulds said:
… We do. Believe we would see a decline in the quality of the image with each model.

… Yes, Technical specs might be up. But, I believe that there will be a decline in the overall authenticity and quality of the image as a whole.

I agree with you that as Foveon sensors went through several generations, the artistic quality of the images they delivered changed.

I don't mean that artistic quality has declined, so here you and I disagree.

This is why some prefer processing images from older Foveon cameras using early versions of SPP, because artistic changes occur in SPP as well. Is v3.3 one of these?

I think artistic vision is important, but it defies metrification: You can't easily nail it down. If you can't identify and quantify it, you can't reproduce it later.

It is starting to seem to me that the color standards used for these things (color checkers etc) aren't really going to allow artistic stuff to survive past its own generation.

In order to better preserve artistic ideals, perhaps a different selection of color swatches in color checkers is needed. I'm not sufficiently advanced technically to say this for sure, though.

What version of SPP are you using to process your images from earlier Foveons? As one who is not particularly interested in artistic stuff (my color vision might not be in the higher percentiles) I always use the latest version of SPP. Seems fine to me.

Scottelly · Mar 7, 2022

Iain G Foulds said:
Tom Schum said:

Definitely different color responses.

dp0 set to custom white balance using ExpoDisc.

fp L set to custom white balance using ExpoDisc, Lens optics (compensation for coloration across the frame) set for the TTartisan 21mm F1.5 at F5.6.

dp0 Lo-res, from SPP set to min noise suppression and crunchy, other sliders at default.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to taste.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to default.

Click to expand...

… Tom: Again, thanks for sharing good work.

… The simplistic colours are irrelevant. It is the sky that matters. The transition of the Quattro is much, much more natural and even. The FP is more simplistic colours, very clumsy… almost blocked out.

… A small wager that the SD14/15 would be better, more subtle and authentic than both. A camera company going backwards… just like all the rest.

Maybe that's why Sigma decided to use a 1:1:1 design for the full-frame sensor they're working on. I'm really hoping it turns out to be the best sensor ever. I just wish it would offer more resolution than my SD Quattro H. Maybe it will, even though it will only have 20 MP per layer.

Iain G Foulds said:
--
... “Photographers have been gaslighted by camera companies to obsess about every leaf on the trees, and have lost sight of the forest.” IGF

--
Scott Barton Kennelly

Big Print Photos - Photography by Scott Barton Kennelly

Landscape and nature photography, photographic prints by Scott Barton Kennelly, an Australian photographer living in south Florida.

www.bigprintphotos.com

Scottelly · Mar 7, 2022

paulsch said:
Tom Schum said:

paulsch said:

With “color palette” of an image I wanted to say the number of different colours used in that image.

To be honest, I had not even tried that hyperlink. I think the linked site uses the term “color palette” for a somewhat arbitrary artistic collection of colour swatches which represents the overall impression of the image.

Click to expand...

Key word here is "arbitrary".

I've wondered this about the colorcheckers:

Datacolor SpyderCHECKR 24 Color Chart SCK24 B&H Photo Video (bhphotovideo.com)

Somebody picked a bunch of color swatches there, too.

There might be a history behind these things, and it might go back to colorimetry research years ago; I don't know.

Maybe a creative selection of color swatches would give Foveon colors a big advantage over Bayer. Maybe the existing selection is based on old color TV science, and that goes back to red, green, and blue phosphors on a TV picture tube. If so, it's a perfect fit to Bayer technology.

Click to expand...

I do not quite understand this.

Maybe we use the term differently. What I tried to say how I understand the term “colour palette” of an image is this:

The Foveon sensor has its data saved to three channels with 14 bits of information, i.e. a theoretical maximum color palette of 2^(3*14) or 4.398.046.511.104 different colours. And every real picture taken with the sensor contains a certain subset of those available colours.

It's theoretical maximum color palette would have to be limited to a different color for every pixel, or about 19.7 million colors, right? (25.5 million for the Quattro H)

paulsch said:
And of course once the Raw data has been transformed into a different picture format (e.g. an sRGB jpeg) the theoretical maximum color palette and also the subset that is really used in the image, changes.

So with “colour palette” I mean the subset of the theoretical available colour values that is in use in the file.

Iain G Foulds · Mar 7, 2022

…Scott: Definitely hoping for the best sensor. I somewhat suspect that the sensor is taking so long because that in chasing the values of the mainstream companies, their images have become basically the same as their’s. If Sigma images do not stand clearly alone, there is just no point.

Tom: I’ve been going between 3.5 and the most recent versions of SPP. To my eyes, they seem to process the SD14 Raw images the same. Both excellent. Though, equal Auto WB issues. Suspect that the only reason Sigma is not a leading camera manufacture in the world is the inexcusable WB issues with their vintage Foveon sensor. It would dissuade the most patient photographer… spoiling so many photos.

xpatUSA · Mar 7, 2022

Iain G Foulds said:
suspect that the only reason Sigma is not a leading camera manufacture in the world is the inexcusable WB issues with their vintage Foveon sensor.

It is interesting that the ONLY reason quoted is "WB issues".

I have had no WB issues with Sigma Foveon-based cameras except the SD14.

In any case, if someone only uses Auto WB, they get what they deserve. For important shots, someone who does not select the nearest WB to lighting conditions and who does not include a neutral target in a test shot also gets what they deserve.

Indeed, if someone is not bothered by color inaccuracy, then there should be no complaints about Sigma's WB anyway.

Iain G Foulds said:

It would dissuade the most patient photographer… spoiling so many photos ...

Click to expand...

... after almost ten years of Foveon, I remain undissuaded even though I have little patience ...

--
It's all in the numbers ...

xpatUSA · Mar 7, 2022

Scottelly said:
Iain G Foulds said:

Tom Schum said:

Definitely different color responses.

dp0 set to custom white balance using ExpoDisc.

fp L set to custom white balance using ExpoDisc, Lens optics (compensation for coloration across the frame) set for the TTartisan 21mm F1.5 at F5.6.

dp0 Lo-res, from SPP set to min noise suppression and crunchy, other sliders at default.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to taste.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to default.

Click to expand...

… Tom: Again, thanks for sharing good work.

… The simplistic colours are irrelevant. It is the sky that matters. The transition of the Quattro is much, much more natural and even. The FP is more simplistic colours, very clumsy… almost blocked out.

… A small wager that the SD14/15 would be better, more subtle and authentic than both. A camera company going backwards… just like all the rest.

Click to expand...

Maybe that's why Sigma decided to use a 1:1:1 design for the full-frame sensor they're working on. I'm really hoping it turns out to be the best sensor ever. I just wish it would offer more resolution than my SD Quattro H. Maybe it will, even though it will only have 20 MP per layer.

I thought Yamaki-san already told us the resolution, Scott - which was much less than that of the Quattro H? (about 80 lp/mm versus about 116 lp/mm).

--
It's all in the numbers ...

D Cox · Mar 7, 2022

xpatUSA said:
Scottelly said:

Iain G Foulds said:

Tom Schum said:

Definitely different color responses.

dp0 set to custom white balance using ExpoDisc.

fp L set to custom white balance using ExpoDisc, Lens optics (compensation for coloration across the frame) set for the TTartisan 21mm F1.5 at F5.6.

dp0 Lo-res, from SPP set to min noise suppression and crunchy, other sliders at default.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to taste.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to default.

Click to expand...

… Tom: Again, thanks for sharing good work.

… The simplistic colours are irrelevant. It is the sky that matters. The transition of the Quattro is much, much more natural and even. The FP is more simplistic colours, very clumsy… almost blocked out.

… A small wager that the SD14/15 would be better, more subtle and authentic than both. A camera company going backwards… just like all the rest.

Click to expand...

Maybe that's why Sigma decided to use a 1:1:1 design for the full-frame sensor they're working on. I'm really hoping it turns out to be the best sensor ever. I just wish it would offer more resolution than my SD Quattro H. Maybe it will, even though it will only have 20 MP per layer.

Click to expand...

I thought Yamaki-san already told us the resolution, Scott - which was much less than that of the Quattro H? (about 80 lp/mm versus about 116 lp/mm).

But that was a long time ago. Plans may well change, especially as Sigma now sell a 61 Mpix Bayer camera.

Don

xpatUSA · Mar 7, 2022

D Cox said:
xpatUSA said:

Scottelly said:

Iain G Foulds said:

Tom Schum said:

Definitely different color responses.

dp0 set to custom white balance using ExpoDisc.

fp L set to custom white balance using ExpoDisc, Lens optics (compensation for coloration across the frame) set for the TTartisan 21mm F1.5 at F5.6.

dp0 Lo-res, from SPP set to min noise suppression and crunchy, other sliders at default.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to taste.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to default.

Click to expand...

… Tom: Again, thanks for sharing good work.

… The simplistic colours are irrelevant. It is the sky that matters. The transition of the Quattro is much, much more natural and even. The FP is more simplistic colours, very clumsy… almost blocked out.

… A small wager that the SD14/15 would be better, more subtle and authentic than both. A camera company going backwards… just like all the rest.

Click to expand...

Maybe that's why Sigma decided to use a 1:1:1 design for the full-frame sensor they're working on. I'm really hoping it turns out to be the best sensor ever. I just wish it would offer more resolution than my SD Quattro H. Maybe it will, even though it will only have 20 MP per layer.

Click to expand...

I thought Yamaki-san already told us the resolution, Scott - which was much less than that of the Quattro H? (about 80 lp/mm versus about 116 lp/mm).

Click to expand...

But that was a long time ago. Plans may well change, especially as Sigma now sell a 61 Mpix Bayer camera.

Even so, Scott is telling us that "it will only have 20 MP per layer", so he is defining the resolution in spite of his preceding 'Maybe's. Therefore, Scott could be wrong either way, eh?

I hereby speculate that the new Sigma sensor will not come anywhere close to the fp L and will probably not even match or better the H - talking in terms of sensor geometric lp/mm.

--
It's all in the numbers ...

Tom Schum · Mar 7, 2022

xpatUSA said:
I hereby speculate that the new Sigma sensor will not come anywhere close to the fp L and will probably not even match or better the H - talking in terms of sensor geometric lp/mm.

But, Foveon colors! That artistry will be back!

Scottelly · Mar 7, 2022

Iain G Foulds said:
…Scott: Definitely hoping for the best sensor. I somewhat suspect that the sensor is taking so long because that in chasing the values of the mainstream companies, their images have become basically the same as their’s. If Sigma images do not stand clearly alone, there is just no point.

Tom: I’ve been going between 3.5 and the most recent versions of SPP. To my eyes, they seem to process the SD14 Raw images the same. Both excellent. Though, equal Auto WB issues. Suspect that the only reason Sigma is not a leading camera manufacture in the world is the inexcusable WB issues with their vintage Foveon sensor. It would dissuade the most patient photographer… spoiling so many photos.

Well, I've seen comparisons between Sony and Sigma Quattro cameras, using OOC jpegs, which were shot with auto white balance. The Quattro did white balance in mixed lighting much better than the Sony, so I guess Sigma fixed that problem over the years.

Scottelly · Mar 7, 2022

xpatUSA said:
Scottelly said:

Iain G Foulds said:

Tom Schum said:

Definitely different color responses.

dp0 set to custom white balance using ExpoDisc.

fp L set to custom white balance using ExpoDisc, Lens optics (compensation for coloration across the frame) set for the TTartisan 21mm F1.5 at F5.6.

dp0 Lo-res, from SPP set to min noise suppression and crunchy, other sliders at default.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to taste.

fp L Lo-res equiv, from DNG with SPP set to min noise suppression and crunchy, other sliders to default.

Click to expand...

… Tom: Again, thanks for sharing good work.

… The simplistic colours are irrelevant. It is the sky that matters. The transition of the Quattro is much, much more natural and even. The FP is more simplistic colours, very clumsy… almost blocked out.

… A small wager that the SD14/15 would be better, more subtle and authentic than both. A camera company going backwards… just like all the rest.

Click to expand...

Maybe that's why Sigma decided to use a 1:1:1 design for the full-frame sensor they're working on. I'm really hoping it turns out to be the best sensor ever. I just wish it would offer more resolution than my SD Quattro H. Maybe it will, even though it will only have 20 MP per layer.

Click to expand...

I thought Yamaki-san already told us the resolution, Scott - which was much less than that of the Quattro H? (about 80 lp/mm versus about 116 lp/mm).

In this case Ted, I'm using SPV terminology, and not the "correct" use of the term "resolution." I'm pretty sure you got that from the way I wrote it, but just in case you or someone else who is reading this thread doesn't understand, here is what I mean:

A typical Quattro sensor (APS-C size) makes 20 MP photos (more or less). The H (that's the SD Quattro H) makes photos that are about 25% more MP. The new, full-frame sensor Sigma is working on will make photos that are about 35% more MP than the Merrills. Considering the fact that many Merrillians claim the Merrill captures as much detail as the standard Quattros, I think those same people will say the 20 MP photos from the new sensor have more detail than what the H captures.

I expect many debates, and possibly even another war, with many battles, between the Merrillians and the Quattronians, based on the arguments that will accompany almost every comparison test between the new cameras and the H. In fact, I wonder what side of the fence Noel will eventually fall on. He seems to have really embraced the Quattro.

Resolution aside, I believe the new sensor will be sort of a new beginning for the world of Foveon style sensors. It should be interesting.

xpatUSA said:
--
It's all in the numbers ...

--
Scott Barton Kennelly
https://www.bigprintphotos.com/

Scottelly · Mar 7, 2022

Tom Schum said:
In this DPR review, on page 6

Sigma SD Quattro H Review: Digital Photography Review (dpreview.com)

it seemed that the Quattro DNG and the SPP-processed X3F delivered different colors.

So, processing seems to matter.

Processing matters.

Maybe we would all be better off shooting only black and white.

I don't like how cold a woman's lips look in a B&W portrait.

Tom Schum said:
--
Tom Schum
"Beware of taking advice from anonymous wise men." Quote from Anon.

--
Scott Barton Kennelly

Big Print Photos - Photography by Scott Barton Kennelly

Landscape and nature photography, photographic prints by Scott Barton Kennelly, an Australian photographer living in south Florida.

www.bigprintphotos.com

dp0 vs fp L

Senior Member

Attachments

Forum Pro

Forum Pro

Veteran Member

Forum Pro

Veteran Member

Forum Pro

Senior Member

Forum Pro

Forum Pro

Forum Pro

Veteran Member

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Similar threads

Keyboard shortcuts