Erik's web

  • Increase font size
  • Default font size
  • Decrease font size

A closer look at Pentax 645D image quality

E-mail Print PDF

After long development the  Pentax 645D was released this year. The new camera offers MFDB sensor chip at DSLR prices and with DSLR handling. DSLR handling with exception to LiveView, that is. More about that later.


So, how good is the new camera? The best one is to get one to test, the next best way to find out is to analyze what others have tested. In this case Imaging Resource has a lot of images available for download and a few conclusion can be made from those images.


Please note that these image are not taken by me but were downloaded from Imaging Resource.


Note: The essential finding here is that the Pentax 645D delivers as promised. Lloyd Chambers is testing the Pentax 645D right now. His general impression is very favorable, but that doesn't seem to apply to the lenses which seem to be sharp on axis but less so off axis. I would suggest anyone considering buying the P645D to subscribe to DAP the TOC probably gives a good idea about all the stuff on that site.

Comparing prints

According to the Michael Reichmann of Luminous Landscape fame "The proof is in the print". Looking at images on the web is not very demanding, what is sometimes know as "Full HD" is about two Megapixels so seeing an image on the screen tells little about the detail in the image. Looking at the image at actual pixels is excessive magnification on the other hand. The only way to really judge images is to compare prints.
I downloaded two "raw images" from Imaging Resource. These images I presume to be reasonably well made comparison shots. The images are from Nikon D3X, what I would consider to be the leading full frame DSLR and the new Pentax 645D. The images were then processed in Lightroom. The processing in Lightroom was essentially:
  1. White balance on Color Checker second brightest grey patch.
  2. Match Total Exposure
  3. Sharpen using : Sharpening - Narrow Edges (Scenic)
For the Nikon D3X and additional I made a virtual copy using a sharpening I often use for with my Sonya Alpha 900 with settings that use deconvolution. This was essentially intended to compensate for the OLP filter on the D3X.
Prints were made from crops corresponding to A2 and A0 size on an Epson printer, probably at 480 PPI directly from Lightroom. Standard output sharpening for glossy was use. The prints were than scanned on a simple Canon LID 25 flatbed scanner at 300 PPI. The images are show below. Click on each image for actual pixels view. Nikon on the left and Pentax is on the right.


Center of the image. Click in image for actual 300 PPI scan of printed image.

Fine text detail near the center of the image. Here we can see some jagged lines on the Nikon images while the Pentax image has smoother edges. It's obvious that the Pentax image has more detail even if the deconvolved image from the D3X may appear sharper.

Click in image for actual 300 PPI scan of printed image.

The brush is probably out of focus on the Pentax 645D it is much softer than on the Nikon D3X images.

Click in image for actual 300 PPI scan of printed image.

And in A0

Click in image for actual 300 PPI scan of printed image.

Here we can observe that the deconvolved D3X image has far more artifacts than the more conservatively sharpened image. Little artifacts are seen on the Pentax 645D image. It also seems that the Penatx 645D image may need some additional sharpening.

Click in image for actual 300 PPI scan of printed image.



A more theoretical view

The comparison gives some qualitative results comparing the two different systems. We can see the advantage the Pentax 645D has over the Nikon D3X but it doesn't really say how much that advantage is. MTF comes to rescue and an excellent program developed by Norman Koren called Imatest allows us to measure it. MTF essentially quantifies how much contrast would be transferred for small details. A good discussion of MTF is given here:
The MTF plot for both system is shown below:


The MTF50 figure is suggested as a good indication of visual sharpness. The figures above indicate that the Pentax 645 would resolve around 4600 lw vertically with an MTF of 50% while the Nikon D3X would achieve around 3050 lw for the same MTF. This really indicates that would be able to linearly enlarge the Pentax image about 50% more for the same visual acuity. This presumes that viewing distance was kept constant.


Another observation is that the Pentax image has pretty high MTF the Nyquist limit. MTF beyond Nyquist cannot be utilized but may fold back as lower frequency aliases and produce false detail. One form of aliasing is the well known Moiré pattern. The sample image with the bootless has some bad Moiré pattern.

Subjective Quality Factor

The perception of sharpness has been subject to significant research. One method of measuring percepted image quality was suggested by Ed Granger of Kodak Eastman and is called SQF which has been popularized by Modern Photography. SQF takes viewing distance into account. Imatest can also calculate SQF values. The SQF plots are below. Please note that the vertical scale the two plots is different.


The SQF takes the viewing distance into account. It is expected that a large image will be viewed at a longer distance, but the viewing distance is less than proportional to the image size. If we would make a print which is 100 cm on the short side this image would give an SQF of around 97 on Pentax 645D and around 93 on the Nikon D3X. Viewing distance would be around 95 cm.


I tried looking at my "A0-size" crops at 95 cm and could not really tell the Pentax 645D and the convolution sharpened D3X image apart. I could easily tell the "Convolution sharpened" image from the one having the "Narrow Edge - Scenic" sharpening, however.

Visual examination of prints

I asked a colleague having considerable experience of working at a high end professional lab for his opinion and he clearly found that the P45D images were sharper, although he pointed out that tonal differences could falsify the comparison.


The Pentax 645D delivers on promise. At least under ideal lab condition it will achieve better sharpness than the Nikon D3X. Lets not forget that are far cheaper alternative to the Nikon D3X around. Any comparison of cost must also include the glass in from of the camera, however.




Last Updated on Monday, 20 December 2010 06:20  


OS : Linux u
PHP : 5.2.9
MySQL : 5.0.67
Time : 06:33
Caching : Enabled
GZIP : Disabled
Members : 1510
Content : 72
Web Links : 1
Content View Hits : 516353