- Original pre Merrill cameras
- The Merrill generation (you're here !)
- The Quattro (working on it)
The Merrill Generation
The third important step forward on Foveon sensors was the "Merrill" with 15MP x 3 layers. After a long time, Sigma enthusiasts now have a truly high-resolution sensor.
Merrills have four flavors:
DP1M
Like the former DP1, with a 28mm (equivalent) objective. It's a DP2M with a wide angle lens.DP2M
Same as above but with a 40mm (equivalent) objective.
The DP2 image quality is outstanding.
 |
Sigma DP2 Merrill
|
DP3M
It's a DP2M with a 70mm (equivalent) macro objective. The lens quality is superb.
 |
| Sigma DP3 Merrill Jpeg from Raw without any enhancements (C) Cesar R. Henriques |
The SLR. In theory, it should give the same results of the DP series, but real-world tests tell us something very different as we will discuss later.
The image quality from these tiny brick like compacts is absolutely jaw-dropping. Honestly. They also had many improvements over the old DP1/2/3 in ergonomics. The new menu system is excellent, with a modern interface and very easy to use.
As you know, nothing is perfect. The Merrill cameras are energy eating monsters, and the tiny battery of the DPxM cameras lasts for about 70-80 pictures maximum. Yes, you read right: seventy to eighty shots per battery. But at least sigma included two batteries with the camera. Seriously, do yourself a favor and buy at least 4. I have 6 of them. They are so cheap !
Sigma said that it's so energy-hungry because the huge amount of data processing needed and the camera has a reasonably powerful processor. The X3F files from the Merrills are 48 MB in size. And it takes a LONG time to write them to the memory card. Thanks to the large buffer, it's possible to continue shooting while writing to the card. Again, do yourself a favor and buy a fast memory card.
X3F files from Merrills needs to be processed by proper software, like Sigma Photo Pro or Iridient Developer.
Many people complain about the existence of three cameras with fixed lenses instead of an interchangeable one. This was a design choice of course, but not a stupid decision by Sigma. This was made to ensure the highest possible precision between the lens and body coupling.
I complained at first but now I really thanks to Sigma for this decision.
The SD1 Merrill red channel blow up problem
The DP1/2/3 Merrill usually do a precise exposure and the images are almost always perfect, but I see that the SD1 has a tendency to blow up heavily the red channel if the exposure isn't extremely precise.
This is a real challenge especially if you like to photograph deep red/pink/orange/yellow flowers. The effect is so strong that all the details are changed by a very flat texture.
After changing some e-mails with Sigma America, Mr. Ebi told me that the SD1 is very sensitive about exposure errors and the usual approach is to underexpose the image by trial and error and then recover it on post-processing.
A not so small color shift may appear, especially on greens. To be fair, you may have a hard job to get the right colors.
Maybe the gain of the red channel is wrong. Nobody actually gave me a convincing explanation, but I'm sure if it was a firmware issue Sigma would have corrected it a long time ago.
 |
| Sigma SD1 Red Channel blowout |
It's possible to see the nasty effect of the red channel blowout when photographing reddish flowers. The actual color is something like a deep crimson/magenta. The image above was obtained using the SPP AUTO mode.
 |
| Same image, but manually adjusted |
The same image, but with some manual color tweaking.
I didn't change the saturation for a single reason: It would smear the greens. It's almost impossible to get the proper results without a separate channel saturation and curve control for each color channel. That's why SPP sucks in my opinion.
 |
| Same image, but resulting TIFF from SPP post processed with DxO |
The image above was obtained with a two-step workflow, processed with SPP to the best possible, saved as TIFF, then processed again with DxO to fix color and saturation on the red and green channels.
The image below was, again, processed in two steps, but using Apple Aperture instead of DxO.
The image below was, again, processed in two steps, but using Apple Aperture instead of DxO.
 |
| Tiff from SPP, post processed with Aperture |
I really need to do some more experiments to really figure out how to fix this. Maybe a blue filter to bring down the red part of the spectrum, like some sort of film color conversion filter.
 |
| Sigma SD1, sunlight , AWB, program mode, 1-stop overexposure |
It's easy to see the effect of the red channel saturation (clipping). The white spots are due to all channels at the saturation level (255,255,255). This is due to a mere 1 stop overexposure.
 |
| Sigma SD1, sunlight, AWB, Manual mode 1/250 F8 ISO 100 Using a color correction Wratten 81A filter |
Now the same flower, but now with the proper exposure, got by a Gossen Luna Pro Digital light meter (1/250s F8 ISO 100) and a Wratten 82A filter to push down the reds/oranges by 30-40% before they reach the sensor, to avoid the channel clipping. Exposure was compensated by the filter factor (1.2x)
 |
| The 82A transmission curve shows clearly the attenuation of longer wavelenghts |
My tests gave me the impression that if you have a bullseye exposure metering you have a good chance of no red and green clipping, but sometimes may happen.
The 82A filter makes things a bit more easy if you're photographing deep red/orange flowers, like roses.
The red
 |
| Oranges, crimsons and yellows with SPP 6.6.0 No visible blow up |
 |
| Pinks ok |
 |
| Yellows and Reds ok |
The next example is about pastel colors and diffuse daylight.
 |
| Pastel colors and diffuse light, SPP processed, AUTO mode |
Note the absolute lack of punch. The colors aren't even close to the real scene. Greens are, again, grayish. Some manual adjustments in SPP made the scene more natural. No other program was used, just some tweaking on color balance, saturation, and exposure and fill light were used.
 |
| Pastel colors and diffuse light tweaking with SPP |
Now, an example with lots of things ready to go wrong. Strangely it was pretty easy to fix with SPP. Just a minor exposure and fill light correction fixed almost all problems. Then, modified the white balance from AUTO to sunlight and bingo !
 |
| SPP Auto setting. Strong overexposure and wrong color balance. |
 |
| Same image, corrected with SPP as described on text. |
The strangest thing was in the next image. It's a very tricky one: a deep red flower against a bright blue sky. Guess what ? The image came perfect.
The camera's auto white balance is not very consistent. Usually, it tends to be a bit cooler than it should be.
The camera's auto white balance is not very consistent. Usually, it tends to be a bit cooler than it should be.
 |
| Enigmatic result. Came perfect, but because the exposure was dead on right. |
The SD1 is the trickiest camera I ever had in hands. If you have Buda's patience and a lot of self-control to not hit the camera with a sledgehammer, it's capable of absolutely stunning results. You may also try some powerful Tibetan mantras. Otherwise, go to another camera.
After your Zen training, you may be able to get some truly amazing images.
 |
| SD1 when right ... Metering was done using a Gossen lightmeter measuring incident light. Perfect ! |
 |
SD1 done right. SPP white balance set to daylight. The original image done with AWB was bluish. Even my cell phone did a better job on AWB.
|
 |
| SD1 amazing detail and stunning colors, after some SPP tweaking. White balance set to daylight and slightly more saturation. |
The Achilles Heel (or Hell): Sigma Photo Pro
I have to admit that Sigma did a huge improvement with their own X3F raw converter, the Sigma Photo Pro.
The current version (6.5.3) is way faster than the later ones and deal better with noise reduction, but there are still some serious problems and lacks some must have controls.
- No curves control
- No histogram stretching/compressing tool
- No way to set black and white points
- No separate adjustments for RGB components
- No sidecars with the applied adjustments !
I really wish someday Sigma do a decent implementation of the above-mentioned features and I honestly don't have a clue why they didn't do. It's frustrating.
Taking everything into consideration my real life experience with Sigma Foveon cameras from the early SD9 to the SD1 Merrill is:
- All DP Merrills are stunning. They can be used at higher ISOs in certain conditions or for black and white images, and up to ISO 3200 in BW if using just the blue layer of the sensor.
- The SD1 Merrill is, well... very moody and needs patience to extract all the juice, but when properly used is a true monster.
High ISO Black and White with the Merrills
This is interesting. At 6400 ISO the color image from any Sigma camera really sucks. Huge banding, chrominance and luminance noise, even in raw.
Let's take a look at a night scene shot in color at ISO 6400
Now a black and white crop, using the all the three layers
Lots of artifacts and noise
 |
| ISO 6400 mono from all layers |
Now the conversion using each channel.
 |
| ISO6400 red channel |
 |
| ISO6400 green channel |
 |
| ISO6400 blue channel |
It's clear that the conversion using just the blue layer has far less noise than the red one who has an awful noise.
The deepest sensor layer is the red one and the blue is the shallowest. The photon has to travel more through the silicon to reach the red layer and this causes more noise.
I would not recommend using the Merrills at ISO 6400 for color images, but at 1600 and even 3200 it's possible to obtain some pretty nice images in black and white.
Black and white images from the blue channel at ISO 800 are practically noiseless.
Lenses for Foveon DSLRs
When the SD14 and SD15 were launched, there were just a few lenses from Sigma on pair with the sensor detail capture.
One of those was the 70mm F2.8 EX DC Macro. This lens is superb, believe me, it's better than the Leica 90mm F4 Makro-Elmar-R and a challenge to the best macro lens ever made, the Leica 100/2.8 Apo-Makro-Elmarit R.
But about zoom lenses, things were bad. Very bad. I wasted almost $400 on the 18-50 F2.8 EX DC Macro. This lens suffers from high chromatic aberration and borders are very soft wide open at the wide angle setting (less than 30mm). Even stopping down to F7.5 it's still soft. The good thing about this lens is the extremely high central resolution for macros. It would be a great buy if the price was about $200.
The old (not Global Vision) 30mm F1.4 also had some issues with soft borders and autofocus inconsistency, especially on Canon EF mount.
No previous SA lenses from the former Sigma SA range of film cameras were good enough to be used with the SD14/15, except the 70mm macro I mentioned before.
Many people simply decided to use old manual focus lenses with those cameras and some people installed custom-made manual focusing screens for precise focusing due to the inconsistency of the AF focus lock indicator with manual lenses.
I ordered a high-quality split/microprism screen from Haoda (closed) and I was very happy using some really good old glass with the SD14/15, especially from Leica and M42 screw mount. My favorite manual focus lenses for use with the SD range were:
- The absolutely superb Leica Vario Elmar R 35-70 F4 Makro. This lens is probably one of the best zoom lenses ever made. The optical quality is outstanding.
- The Leica Elmarit-R 28mm F2.8, the best 28mm I tested up to that date.
- The Leica Apo-Telyt-R 180mm
The only problem was the lack of a decent wide angle.
Old, Manual Focus Lenses With the SD1 Merrill - The Color Cast Problem(Not happens with pre-Merrill cameras)
This lack of good lenses got critical when the SD1 was launched. The SD1 relies upon a heavy software correction to produce the final images and if you try to use any lens without the necessary electronics to tell the camera what's fitted, the resulting image will have a very weird and horrible color cast that cannot be corrected (at least in an easy way).The cast is variable with the focal lens, aperture, exposition, moon phase and who knows else! It's a very strange circular gradient from magenta to green and sometimes it's inverted!
So, manual lenses are almost always a no-no for the SD1 Merrill, or at least you may have a good headache to fix everything.
Here is a transcription of an e-mail between me and Sigma staff:
Hello !
I'm not sure if I can contact you using this e-mail address but I have some questions about the Sigma SD1.
I bought one last Friday and I'm having some sort of color cast problem when I use old manual focus lenses with this camera.
The image shows a strange green-to-magenta color cast from the borders to the center, every time I try to use one of my Leica-R lenses converted to SA mount, or even with M42 lenses. The mentioned lenses work flawlessly with all my other Sigma cameras, to know, Sigma SD9/10/14 and 15.
Do you have any advice or solution to this strange behavior ?
Thanks in advance
Antonio
From Sigma:
The SD1 and SD1 Merrill read lens information and process images.
Former cameras than this camera didn't have this specification, therefore you didn't experience this issue before.
We started to incorporate it with SD1 and SD1 Merrill camera in order to achieve further improvement of image quality.
Thanks God I was not crazy or had a broken camera =)
After changing some e-mails with Sigma support asking for a good general purpose lens that wouldn't break my budget , Mr. Shingi Yamaki from Sigma suggested me to try the 17-70 F2.8-4 DC OS HSM with the SD1 and it's indeed a very good lens.
It worked perfectly with the SD14/15 but mine had some minor back focus issues with the SD1 that were easily corrected by the in-camera micro adjustment. I also had to do the same with the 70mm macro. Now they're tack sharp and I'm very happy with them and the SD1.
He also suggested me to try the 18-200 F3.5-6.3 DC Macro OS HSM (why are their names so long ?) and I found that this lens is a cracker for the price. It has outstanding image quality for the zoom range.
If you're looking for a cheaper lens, the 18-250 DC OS HSM will surprise you. Just use it stopped down fo F8 and the image quality will come very good.
Just have in mind that long zooms have some limitations and be happy.
Sigma staff was always very helpful and kind. My special thanks to them.
The other lens I tested and I was shocked by the extremely high quality was the 18-35 F1.8 "Art". Boy, this lens is a monster. It's the highest quality zoom I've ever used. It's just insane with the SD1 Merrill.
In early 2018 I was contacted by a guy from Switzerland about the SD1 Merrill color cast. Unfortunately, I don't know his name but here is his blog.
I was curious and did some tests using a Leica Vario-Elmar 35-70 F4 Makro converted to K-mount by a Leitax adapter (it fits also to SA mount). It's a fantastic lens and I use it a lot on my Sigma SD15 and Pentax K3-II.
The process is described on his blog, but I'll condensate the most important information here:
1) Set the SD1 to aperture priority at f=1.0
2) Take the light reading at the real aperture you set the lens. Let's say the informed speed was 1/250s for F=8
3) Change to manual mode.
4) Set the camera and the lens aperture to F=8 and shutter speed to 250.
5) Take the photo.
It worked partially for me. The strong magenta cast at 70mm f=4 almost disappeared, but at 35mm F=4 the cast inverted from a strong magenta to a green one, but much weaker. I'll keep doing some tests before posting the results here. The method is very awkward.
Like I said before, the cast depends on many factors and I'm not able to figure yet about all the variables.
The following images show the problem. I know the image sucks but I did them on a run, but they are perfect for what I want.
 |
| Mode M, Real aperture = f8, camera set to f8 1/250 Slight radial gradient from neutral to green, metering from Gossen Luna Pro Lens: Leica Vario Elmar 35-70 F4 at 35mm |
 |
| Mode Av, Real aperture = f4, camera set to f1 Strong radial gradient from magenta to green Lens: Leica Vario Elmar 35-70 F4 at 35mm |
 |
| Mode Av, Real aperture = f8, camera set to f1 Strong radial gradient from magenta to green Lens: Leica Vario Elmar 35-70 F4 at 35mm |
 |
| Mode M, Real aperture = f8, camera set to f8 Strong radial gradient from neutral to green Lens: Leica Vario Elmar 35-70 F4 at 35mm |
 |
| Mode Av, Real aperture = f8, camera set to f1 Strong magenta cast Lens: Leica Vario Elmar 35-70 F4 at 70mm |
 |
| Mode M, Real aperture = f8, camera set to f8 No cast, a bit subexposed Lens: Leica Vario Elmar 35-70 F4 at 70mm |
 |
| Mode M, Real aperture = f8, camera set to f8 1/250 No cast, metering from Gossen Luna Pro Lens: Leica Vario Elmar 35-70 F4 at 70mm |
 |
| Mode Av, Real aperture = f4, camera set to f1 Strong magenta cast Lens: Leica Vario Elmar 35-70 F4 at 70mm |
 |
| Mode M, Real aperture = f4, camera set to f4 No cast, metering from Gossen Luna Pro Lens: Leica Vario Elmar 35-70 F4 at 70mm |
 |
| Mode Av, Real aperture = f4, camera set to f1 Radial gradient from magenta to green Lens: Leica Vario Elmar 35-70 F4 at 35mm |
 |
| Mode M, Real aperture = f4, camera set to f4 Radial gradient from neutral to green Lens: Leica Vario Elmar 35-70 F4 at 35mm |
All the above pictures in this series were shot under the same light and from the same point to keep consistency.
My conclusion is that the proposed method from pixelquality.blogspot.com works partially for the SD1 Merrill, depending too much on variables that aren't under control and also not very predictable.
The best results came from doing manual metering from a hand lightmeter (I used a Gossen Luna Pro Digital) that gave a reading of 1/250s F8 ISO 100. I didn't compensate for the backlighting caused by the white clouds.
I would NOT recommend to use old manual focus lenses on aperture priority or using the built-in camera light meter. You will get a color cast in 90% of the images.
If the metering is dead on, the cast is very well controlled, but not 100% guaranteed. Again, it will depend on the lens, aperture, and the light incidence angle.
All images were processed using Sigma Photo Pro 6.5.5 with all controls set to zero.
Again, just to keep things clear, the color cast with non-Sigma SA lenses will NOT happen with the SD9/10/14/15.
Final word:
- Merrills are image quality demons, but moody.
- The DP range has no issues.
- The SD1 is best used with the new Sigma Global Vision lenses.
- Sigma Photo Pro 6.6.1 still needs a lot of work
Continue reading:
No comments:
Post a Comment