Tag Archives: Foveon

Sigma SD1 breakdown

Posted on by
9

Sigma SD1 breakdown as reported from Japan.  The camera is developed there and prototypes have been previewed and used by some there so I’ll cut and paste some specs from that japanese blog post below.

I’ve seen the full sized fashion shoot photos done with the 46mp SD1, and the details are stunning.  One of the top master printers in the USA who did the printing for Sigma has stated that the SD1 out resolves every single pro DSLR available today (think Nikon D3x or Canon 1Ds Mark III), but he said it falls shy of medium format.

Before you poo-poo the camera, consider this… the target price for body only is $1600-$1700 US.

  • X3F RAW file sizes range from 45-50MB
  • The JPEG in RAW+JPEG mode are around 7MB
  • There is little to no difference between write times for RAW+JPEG and plain RAW
  • One RAW file takes about 10 seconds to write to card
  • At present the camera is verified to utilize UDMA Mode 6 cards
  • They (Sigma?) is currently verifying UDMA Mode 7 cards and these should be faster than Mode 6 cards.
  • A new version of SPP will be needed to process SD1 RAW files. (You get an error if you try to open SD1 files in the current version of SPP.)
  • The RAW processing algorithms are completely different.
  • SPP for the SD1 is currently under development (It will support RAW files from current cameras.)
  • [ed: hmm...] An additional “Adjustment RAW mode” will be added due to the increased complexity of the RAW data coming from the SD1
  • In “Adjustment Mode” you won’t be able to use the loupe or magnify the image, however, processing speed will be increased ten-fold.
  • You will be able to use this mode with current RAW files, but there won’t be an equivalent speed increase
  • The workflow will be where you modify a RAW image in “Adjustment mode” then you convert it to a Full RAW file.
  • Sigma is looking into a Kelvin-based white-balance adjustment, but they are finding it difficult to achieve. This is mainly because with Bayer sensors you deal with a two-axis adjustment, but with Foveon it is a three-axis adjustment.
  • High sensitivity is an algorithm with a goal to keep as much color information as possible. So with the current Foveon sensors (in current cameras) produce a large amount of noise as sensitivity increases. With the new Foveon sensor there is barely any noise.
  • ISO sensitivities from 100 to 6400
  • ISO800 produces almost no noise and ISO1600 is very usable
  • ISO6400 is available in “Extended Mode”
  • The SD1s shown at CP+ had a Firmware version of 0.5. It looks like there are quite a number of incomplete functions and a number of bugs remain.
  • The back-side AF button was on the camera, but it currently doesn’t function.
  • Autofocus on the demo units was clearly faster than the SD15 and will only improve over time.
  • All 11 AF points are twin-cross sensors and every sensor is equally sensitive.
  • The AF sensor was announced to be sensitive from -EV2 to 18, but the final unit will be sensitive from -1EV. [ed: there was perhaps a typo here. I couldn't tell if the announced sensitivity was -2EV or just 2EV...]
  • AF Micro-Adjustments will be available on a per-lens basis, but it’s not clear how many lenses worth of storage the final unit will support.
  • The top LCD was removed from the SD1, but each button will display information on the LCD screen when pressed.
  • The SD1 specs show the rear LCD to be the same specs as the unit on the SD15, but the LCD on the units at the show seem to be different and much better.
  • All of the SD cameras have been easy to hold, but the SD1 is even more so.
  • There is a mysterious menu item selection that seems to indicate that you will be able to change the lens information (this is unknown and may not even end up in the final product)
  • Even though the camera might record the lens information, Sigma does not and will not apply and lens corrections to the files in-camera.
  • There is a “Cleaning Mode” in the menu system (The feature was in previous SD cameras, but it was never easy to remember how to do it.)
  • The shutter sound is even quieter in the SD1
  • Mirror-shock is smaller (even though it was small enough even in the SD15)
  • The viewfinder looks even brighter now that the crop-factor is 1.5x instead of the 1.7x in other SD cameras.

Pentax K-x initial test

Posted on by
4

The Pentax K-x arrived today.  I don’t want to ramble on about all the features and such yet because I have not played with it enough to pass judgement on it yet.  I just want to post this three up image of shots I snapped just before a partly cloudy sunset ended here today.  It shows (top to bottom) HDR Capture Mode “Strong”, then HDR Capture Mode “Standard”, then no HDR Mode and just one simple frame.

The image is of one of the public beach access points here.   It has not only rained at this location recently, but one of the natural dune lakes that sits against the beach (100 yards north of the gulf) broke through one of its natural walls (sand) some dark brackish waters spilled onto the beach.  Hence the wet and dirty strip of sand you’ll see.  It is a naturally occurring thing here where some of these lakes sit.  Only Australia has similar dune lakes naturally occurring like we have here along the Emerald Coast of North West Florida.  The lakes are large and beautiful and allow you to canoe vast areas just hundreds of yards from the gulf of mexico.  They run right up against many beach points.

When it rains here, certain dune lakes will sometimes overflow and spill into the gulf itself and salt water and lake waters will flow back and forth and then seal back up again when weather calms down.  Sometimes a lake wall just oozes a little lake water onto the beach area as it did this time.  It happens a few times a year and within a few weeks the powdery white sand covers the darker sand back up again as if it never happened.  This generally occurs in the fall/spring outside of tourist season.  However, we have had unusually horrible weather here the last two weeks with virtually no sun and many rainy days.  The dune lakes have filled up and this particular one spilled over onto this beach.

I say all this because I don’t want you to think those dark areas are an artifact from the HDR mode versions captured (below).  The HDR mode and tonemapping did NOT have to do with this strip of dark sand.  It was the lake overflow.

The images below were simply resized down.  All three are JPEGs right out of camera.  The HDR mode fires off a 3 frame AEB and the final image is a tonemapped JPEG.  I purposely chose a shot that is not something any camera sensor can handle in one shot.  The image on the bottom is one simple frame with no HDR mode.  As you can see, the sun despite being shielded slightly by clouds, still blows out badly and much of the water reflection is also blown out.  The middle image is using the camera’s HDR Mode set to “Standard” and the top image is using the camera’s HDR mode set to “Strong”.  No photoshop or post tweaks.  The saturation differences are a result of the merging (tonemapping) process and I can’t control that.  Although I did have the camera setting set to “standard” for color.  I’ll experiement with “Vibrant” and the like in the future to see variations.

I was in a rush and by mistake I had the camera set to auto white balance (ugh!!!).  So color differences are probably even less scientific between shots given that idiotic oversight.  BTW, this was not a planned shoot… I just rushed down when the cloud cover broke and snapped a few shots.  I shot in Aperture priority mode for each.  The clouds moved of course… the HDR Capture Mode takes 3 shots bracketed and then it takes about 10+ seconds for the camera to merge the shots before you can retake another shot, so there is no way to show you a true comparison with the sky moving as briskly as it was.  The 3 shot burst itself is quick.  The camera sports a 4.7 fps speed, but the merging process takes time.

So what you see below is not bracketed shots taken into post, merged and then tweaked.  These are right out of camera JPEGs, but there is obviously some tone mapping operator doing some work in camera.  The “Strong” is definitely too intense for my commercial work and “Standard” while less intense is borderline workable for what I do… but I’ll wait until I’ve put it through a bunch of tests and interior architectural shoots.

K-x quick test of HDR Capture Mode

K-x quick test of HDR Capture Mode

Weather is supposed to stay crappy here for almost another week so it may be some time before I get around to getting this camera pushed through a true test run and then posting results.  But I will ASAP.

For instant gratification, the K-x rocks.  No long merge to HDR time in post.  Just shoot and in 10-12 seconds you can see the JPEG and change your settings to reshoot.  Oh, and that reminds me to mention that you can of course use the exposure compensation….

… the Pentax K-x allows for +/- 3.0EV so you can easily control where the base exposure will begin.  And of course I’ll be experimenting with taking multiple HDR captures taken at various +/-EV settings and then use those JPEGs to merge in post to see if anything interesting comes from it or if it is not beneficial.  We’ll see.  Merging tonemapped jpegs?  Can you say “artifacts”?  Probably.  But I just have to do it for the hell of it.

Oh #2… the AEB function allows for a traditional 3 AEB burst at +/-3.0EV.  You can bracket as little as .3 and it goes up by steps of .3 right up to a full +/-3.0 EV like a Canon 7D or Sigma SD14 does (both of them shoot a 3 AEB which can max out at +/-3.0EV).

More to come when weather breaks for the better.

HDR Cameras – 2009 All-Stars

Posted on by
Reply

Jack Howard has put together a great list of top class HDR capable cameras over at Adorama’s Learning Center.  Check out the list here:   2009 HDR All-Stars

——————-

For you iPhone addicts, check out the very interesting app that allows you to take two images and they get tonemapped in the iPhone.  The app is called TrueHDR =  http://pictional.com/TrueHDR/Videos.html

——————-

Here is a recent shoot I did for an existing client. On this shoot I used a D3 exclusively and many of the shots required a range of -5 to +5 EV in total.
Builder Shoot (Gallery Link)

——————-

Here is a link to a High Dynamic Range Image sequence.  It is a “tilt” sequence I shot at a local garden near a state park.  This series was shot with a D3 and each frame is actually 7 images spaced 1EV apart and then tonemapped before being placed one after the other in post.

The amount of quality lost when exporting to a web video format is horrifying, but I have no choice.  The full 1080p version has no loss of color, detail/clarity or color shifts.  That said, here is the link to the web clip:
HDR Video (link to landing page)


Autobracketing for HDR – Camera Specs

Posted on by
Reply

Here is a chart showing a few dozen cameras.  From full frame to 4/3rds.  The last column of numbers is the basis for this chart.  The benchmark.  This chart is ranking from top to bottom the cameras that can cover the largest amount of dynamic range using the camera’s default AEB settings.

The final column is displaying the total EV distance from the first shot to the last shot taken in an AEB sequence.   I’m calling it a spread, but it is the distance in EVs I’m measuring.  It is NOT a number that shows the totality of EVs captured using any other measurements.  Just the AEB distance where shots can be taken in EV jumps using the default number of frames in the AEB sequence and maximum  EV jump between shots.

To be redundant… I am ONLY calculating the distance covered in total for the EV jumps.  It is to show  the widest possible spread each camera is capable of landing exposures/shots at.

Obviously each image itself has it’s own dynamic range that it captures at that exposure. That is not being taken into account on this chart.

Now you can see why I am such a canon basher about the crippling 3 AEB of all but their most expensive offerings.  Look at how wide the 1Ds/1D series can bracket.  Canon clearly has the capability to add a higher AEB function into their cameras.  The 1Ds/1D series AEB settings are the benchmarks in the industry for default camera settings. However, canon opts to cripple every single  camera body under $5000 with only 3 AEB.  So the chart is a barbell.  Canon rules the top tier, but owns the bottom of the barrel as well.

If the 7D had been given the same AEB as the 1Ds/1D series, then it would be one of  the best HDR Capture cameras to date.  It would have been “KING HDR” for APS-C sensors.  Because at that price point with 8fps for continuous capture, High ISO performance, 18 megapixels, etc…) it would probably have been my number one recommended camera for HDR capture.  But 3 AEB at +/- 2.0EV is extremely limiting for professional HDR capture.  95% of my work falls outside of canon’s 3 AEB range.

If the order below seems a bit strange, it is because in the few situations where cameras have the same default AEB settings, I put the camera which has a sensor with a higher dynamic range just above a camera  with a sensor that has a lower dynamic range reading (see past few blog posts if this doesn’t make sense).
[ EXAMPLE: Fujifilm S5 Pro sits above the other Nikons with same AEB settings because the S5 Pro's sensor has a higher Dynamic Range reading according to DXOMark.com. ]

The columns are NAME, then number of maximum shots that can be taken in an AEB sequence, then the maximum EV step that can be taken between each frame and finally the total EV spread from the base exposure to the final exposure (this number is the distance in EV ‘JUMPS‘ the AEB sequence can cover)


EDIT – The CANON 7D in this chart shows +/-2.0EV jumps, but I recently found out that the Canon 7D is actually capable of +/-3.0EV.


Camera's Default AEB and maximum EV Spread