Sigma DP2s Review and FAQs
Copyright © 2010-2011 by Zack Smith .
All rights reserved.
Table of contents
IntroductionThe Sigma DP2s has just been released and this is page will offer my review of this compact big-sensor camera. The price of the camera was initially $700, due to the quality of the photos it takes. During pre-Christmas 2010, the price was $599.
I presently own and am very impressed with the DP2s. As with the DP2 and DP1 before it, it lacks a few basic features that most cameras come with. But it's a camera with a very large, unique image sensor that provides excellent dynamic range and near-perfect sharpness. The result is stunning photos.
To be sure, the DP2s lacks:
But for my needs it is excellent. It is what it is: A small camera with a big sensor, a high-quality fixed lens, an excellent manual mode and the option to add filters.
My unboxing videoThis is down for the moment.
Written ReviewThe DP2s is a very nice camera in most respects.
So far I can say:
Photographic tests: ISO aperture priorityThese ISO tests involved my using several 40 watt bulbs, situated about 9 feet (3 meters) from the subject, progressively turned on. In the case of 0 bulbs, there is a nightlight nearby on the floor, whose light did not directly hit the subject.
These are a redo of my A tests to ensure that the exposure is always +0.0.
White balance in camera and in SPP4: Incandescent
Photographic tests: ISO manual modeFor these shots, everything was the same as the aperture tests except that I set the shutter speed myself in order to match the brightness of what I saw.
Photographic tests: ISO comparison to point and shoot camerasFirst, two photos from two cameras, same subject and light, same f-stop (f3.2), same aperture (ISO 1600), and same shutter speed (1/15 second). The Sigma DP2s performs dramatically better, of course.
The Canon is perhaps not the best point and shoot to compare to. Here is a shot that I took with a cheaper and smaller FujiFilm J30 12 MP camera that nevertheless performed better than the Canon. One light bulb was on, ISO 1600, 1/20 sec, 12 megapixels shrunken using GIMP to 2640x1980: JPEG.
Photographic tests: Barrel distortionTo determine how much barrel distortion the DP2s introduces, I created a test pattern and placed it on a basically flat surface. I took the following shot from 17 inches above it (43.2 cm), demonstrating that there is some barrel distortion, but not much.
Photographic tests: Close auto-focusThe manual focus dial indicates that the minimum focus distance (MFD) is 0.28 meters, which is 11 inches. MFD is by definition measured from the sensor, not from the end of the lens. To determine more precisely what the minimum distance to the subject is to achieve adequate sharpness, I created a test pattern page, set the camera to auto-focus and took some shots.
In my distance calculations, I'm assuming that the sensor is about 2.5 inches behind the end of the extended lens.
All the following shots were at ISO 100, f7.1 and 1/20 second. Both the DP2s and Sigma Photo Pro 4 were set to the default sharpness value of 0.0. No macro lens was used of course.
On my pattern page, the thick lines are ¼ inch apart, the thin lines that are closest together are 80 per inch (31.5 per cm), and the other thin lines are 40 per inch. This is printed with a laser printer.
Photographic tests: Using generic macro lensRather than buy the +10 diopter Sigma macro lens, I got the generic +8 diopter 46mm lens from Deal Extreme for $8. The DP2s auto-focus appears to be able to work with this lens. Here are some test shots.
Photographic tests: Using two generic macro lensesI bought a second Massa +8 diopter lens, which as far as I know is mainly available direct from China, and used that on top of my first macro of the same type, achieving results like this:
Photographic tests: Infrared photographyThe DP2s and its Foveon sensor, when combined with an infrared filter like an R72 (720 nm) on the end of the HA-11 hood adapter, is very useful for infrared photography. The infrared photos that I took before, with a DP1, were astonishing. They do require a sunny day for best effect.
Photographic tests: AstronomyLate night photos of stars are an interesting use of the DP2s. The following shots are all 15 second exposures, f2.8, daylight. No telephoto or telescope used.
Photographic tests: Dark scene & lightsThis requires a large aperture because
Image qualityThe RAW photo quality provided by the Foveon sensor is astonishing in its sharpness and respectable in its dynamic range.
As was revealed in the DPReview Sigma DP2 tests, the dynamic range does not cover the dark end well, but it does cover brights better than most sensors.
I am finding that I need to do less post-processing in Sigma Photo Pro 4 for the DP2s's photos than I did for the DP1's in SPP 3.5.
The DP2s's JPEG photos are said to be good and better than the DP1's by far, but they're not the purpose of the camera. You buy this camera for the RAW photos, just as you buy a sports car for the fast driving, not for delivering newspapers. Personally I rarely used the DP1 for JPEGs so I can't compare.
Speed tests: Shots per second and SD writing rate
Class 10 Sandisk Extreme SDHCTo determine the maximum performance, I used this Class 10 card, ISO 200, f2.8, 1/500 second.
Class 4 Kingston MicroSDHCTo test using a more modest card, I put this MicroSD into an SD adapter and used continuous shooting in Manual mode, with ample light, full 4.7 megapixels resolution, sharp manual focus, at ISO 800. In each test, I shot for 60 seconds and writing to SD ended at about 65 seconds.
With a Class 4 Kingston MicroSDHC card in a Kingston SD adapter, the results were as follows.
I also redid the Fine, Normal, and Basic shots above in low resolution, and got the same results even though the files were considerably smaller.
Class 2 Sandisk MicroSDHCWith my Class 2 Sandisk MicroSDHC card in a Sandisk SD adapter, shooting full resolution only:
If the shots-per-second were limited by the maximum speed that one can write to SD, one would expect that many more of the smaller Basic and Normal JPEG files would be written per minute than of the Fine JPEG files. But this is not the case.
The SD card is under-used when writing JPEGs, with a maximum of 31 shots per second using the fastest card.
But what is the maximum write speed? RAWs take up the most space, so they represent a better test of the SD interface's performance. Again the fastest card indicates the best performance, which was 13.6 MB/second in my tests.
Is it worth spending extra for the super-fast card? In my opinion, a cheap Class 2 card should suffice for most people.
Speed: USB transferI realized that writing to SD while taking photos could be slowed down by the camera's processor being busy, or by the maximum read speed out of the Foveon sensor. So here I determined the bandwidth of writing from the computer through USB onto the SD that's in the DP2s, and the reverse.
WritingFor this test, I connected the DP2s to a laptop's USB2 port, and then took a very large file (639 MB) and copied it onto the DP2s. The card in the DP2s was my Class 10 Sandisk SD Extreme, whose maximum speed is only 5.48 MB/second.
time ( cp -v bigfile f:/ ; sync )
Writing a 2 GB file took 381 and 384 seconds in two attempts, averaging 382.5.
I also tried the same test with my Class 2 card and running the test on a different laptop, and it wrote at 5.3 MB/second.
ReadingTo test USB read speed I used the same basic approach, but this case requires a little more care: If the operating system has ever read or written a file, there is a chance that some or all of it will be cached in RAM. So for this test I copied bigfile to the flash card, renamed it, turned off the camera and then on again. This ensured that Vista believed it was unfamiliar with the file.
cd f:/ time ( cp -v bigfile c:/ ; sync )
Testing my Class 10 card using a 2 GB file, the time was 288 seconds for a rate of 7.26 MB/second.
Testing my Class 2 and Class 4 MicroSD card in the SD adapter and running the test from a different laptop, I got a read rate of about 6.6 MB/second.
Tests of SD cards without cameraTo put the USB transfer speed into context, I tested my flash cards in the SD slots of my laptops.
Using BASH, I ran this command under Windows, similar on the Mac.
time ( cat files* f:/ ; sync )
The Class 10 Sandisk card, inserted into my Macbook Pro, performed as expected: he write speed was 14.7 MB/second and the read speed was 25.9 MB/second.
The Sandisk Class 2 and Kingston Class 4 cards, inserted into my Windows laptop, performed identically to each other in two computers. Each had a sustained a write speed of 5.2 MB/second.
I also tested my Transcend 16 GB Class 6 card, but it fared worse at only 4.7 MB/second when writing.
This confirms that the SD cards are not the limiting factor when doing USB transfers.
ConclusionThe transfer rates for USB 2 are less than stellar. USB 2 itself is rated as being able to transfer up to 60 MB/second, however this camera maxed out at 7.26 MB/second while reading, and 5.48 MB/second while writing.
It's clear from the shots-per-second tests that this camera can write to SD quickly, up to 13.6 MB/second in the case of the Class 10 Sandisk.
Focus speedThe focus of the DP2s is quite fast. The era of slow-focusing Sigma DP cameras appears to be over. The DP2s, and the DP2 with firmware update, fixes the slow focus problem with previous models that everyone complained about. The DP2s focus feels instantaneous.
DurabilityThe metallic housing is tough and impressive and reminds me of pre-digital cameras. Mind you, it's not tough in the sense that you can drop it -- it's a camera after all.
To avoid dropping the DP2s, you should definitely get a wrist strap. It comes with only a neck strap.
The screen could definitely use protection. I use a Delkin shade to protect my DP2's screen, however its screen protector film is glossy, thus in bright sunlight it reflects and this will sometimes defeat the purpose of having a shade.
TelephotoWhile in some cases it's no problem to move the camera closer to the subject, in others like astronomy this is not an option.
You can put a telephoto lens on the end of the HA-21 hood adapter. Luckily the DP2 and DP2s, because of their focal length, will not cut off the four corners of the image like the DP1 does.
I bought a generic telephoto lens intended for camcorders on Ebay and put it on the end of the DP2s, as you can see here:
Video modeEveryone says the video capability of the DP-series cameras is a joke and they are basically correct. The resolution is 320 by 212 with a black bar 28 pixels high at the bottom, to make it 320x240.
Note! According to a Foveon brochure from a few years ago, their X3 sensor can produce 640x480 video. For some reason the Sigma cameras don't take advantage of that.
Some people have pointed out that the clarity of the 320x240 video is good enough that you can increase it to 640x480 resolution and still have it look as good as the video of cheaper compact cameras.
Timelapse moviesThe fact that the Sigma DP2s produces such amazing images may increase your temptation to use it to create time-lapse and stop-motion films. You can do this in three ways, but methods that would be better aren't possible.
Here is my compilation of some of my DP2s time-lapse movies.
Three segments are visible light and one is infrared.
Here is one of my night timelapse movies:
How to make a time lapse movieTo convert a series of frames to a movie under Windows, use AviSynth using the method that I describe on my AviSynth page. It's very simple.
To remove jitters from a timelapse movie, use VirtualDub in combination with Deshaker plugin as I describe on my page.
Stop-motion moviesTraditionally stop-motion movies involve moving some item a small amount, taking a shot, then repeating that. Stop motion movies usually require pressing the shutter because the time between shots is not typically fixed.
You can then string the shots into a movie using AviSynth or FFMPEG, and stabilize the video using VirtualDub and Deshaker, as I describe on my AviSynth page.
Bright redsThe Sigma DP2s is, like the DP2 and DP1 before it, very sensitive to bright red colors. In bright sunlight a red flower will be extremely red and because of this it's necessary in Photo Pro to use the X3F mode rather than Auto. The Auto mode will overemphasize the red.
Yellow blotchesSigma's Photo Pro software performs some chromatic noise reduction that unfortunately can actually add noise to some low-light RAW photos, in the form of yellow blotches. I've found that using non-Sigma software to process RAW photos can alleviate this.
Battery lifeAs a first test, I set the camera to intervalometer mode, 30-second interval, enabled the power-saving Eco mode, left the LCD sleep time at the default of 1 minute, left the LCD brightness at the default level, and started it shooting. It stopped after 220 RAW shots. However it seemed to me that the screen was on a lot of the time, rather than turned off to save power. After the 220 shots, the battery gauge showed two thirds remaining, but soon after the camera reported it was empty.
For a second try, I turned down the LCD brightness and got 223 RAW shots. However the firmware crashed in the middle and I had to cycle the power.
As a third test, I set the LCD sleep to 30 seconds and the LCD brightness to the lowest level and starting it shooting. The LCD remained on.
Minor gripesI personally think the DP2s is a fantastic camera, but I also recognize that its purpose is relentlessly myopic: To provide excellent image quality in a small package. I'd rather own a DP2s, which achieves this goal than a generalist camera that produces fuzzy photos.
But it is probably this specialized nature of the DP series that has resulted in a curious phenomenon, the tendency of the Sigma DP series to attract a great many minor gripes. It's almost a tradition at this point. And so, without further delay, here are my own minor gripes.
Software hacksI'm in the process of writing a computer program that dissects X3F files. One interesting tidbit that I've learned along the way is that each X3F file is not just a RAW file: It's a RAW file that additionally contains two JPEGs.
The first JPEG is 2640x1760, and the other is a 221x147 thumbnail. The larger one is a JPEG that contains a TIFF, oddly enough. The TIFF contains some not-so-useful characteristics information as well as an EXIF section that I haven't bothered to decode. My program extracts both of the JPEGs, of course.
The X3F contains a bunch of data about the shooting parameters, namely:
AEMODE, A AFMODE, MF AP_DESC, 4.5 APERTURE, 4.50000 BRACKET, BURST, 0 CAMMANUF, SIGMA CAMMODEL, SIGMA DP2S CAMNAME, CAMSERIAL, 1001105 DRIVE, 2S EXPCOMP, +0.00 EXPNET, +0.00 EXPTIME, 10000000 FIRMVERS, 1.00.0.008 FLASH, OFF FLENGTH, 24.2 FLEQ35MM, 41 FOCUS, M IMAGEBOARDID, IMAGERTEMP, 40 ISO, 200 LENSARANGE, 2.8 to 14 LENSFRANGE, 24.2 LENSMODEL, PMODE, M RESOLUTION, HI SENSORID, SH_DESC, 10 SHUTTER, 10.00000 TIME, 1272734689 WB_DESC, Daylight CM_DESC, Standard
The actual RAW image from the DP2s is 2688 by 1792, but the extra pixels beyond the central 2640x1760 are often blank/garbage. Actually I should be more precise:
Thus, by using the extra RAW pixels produced by my DP2s that my program is able to provide, I got an image that's 2652*1767 = 4.686084 megapixels, as opposed to the 2640*1760 = 4.6464 MP image that the marketing material describes.
My experimental program, which is a work in progress, is here. It is still not completely clear to anyone but Sigma how they encode the raw pixels and how they should be processed on the computer. The output of the program is consequently imperfect. This code is based largely on observations by Keith M.
Frequently Asked Questions (FAQs)
ConclusionThe DP2s is an affordable, pocketable, big-sensor compact camera. I have found that it's a very nice camera and the shots are quite good, better than those of many DSLRs.
The DP2s is definitely an improvement over the DP1 that I owned before it. It's all around faster. It can write to SD while you take the next shot and it seems to handle bright reds better. Low-light color distortion in the form of yellow blotches is still a problem.
The DP series lacks major features that are expected by many consumers such as a zoom lens and a decent movie mode, but that isn't a big issue for me most of the time. For instance, if I want to do video on the go, I'll buy a $100 Canon with a CCD sensor.
Compared to the Olympus PEN series, the DP2s offers much more realistic looking shots. The PEN cameras produce overly contrasty photos in my opinion.