Review: Niek Haak Reviews the Sony DSC-H9 - Part 2

Niek Haak contributes a continuation of his expert review of the new Sony DSC-H9 digital camera to the IR Buzz. Here he provides an emphasis on it's implementation of the Sony Nightshot feature. In this article, Niek builds on his excellent discussion of the issues, features, and pros and cons the H9 owner who wants to shoot IR will encounter. Be sure that you don't miss Part Three, do not miss any of this must-have information about the H9.

Guest Author -- Niek Haak

Part Two:

In part One we looked at the DSC-H9 Nightshot mode, options for infrared photography and for attaching filters. In this part we will check out image quality for infrared photography. With visible light the H9 can take excellent pictures. If conditions are good pictures are usually better (more details, less noise) than those from my DSC-F717. But by now most of you have probably heard about image quality problems with the H9. The major issues are unpredictable noise/smearing artifacts when conditions are less favourable (especially in low contrast/low detail areas), bad corner sharpness in wide angle mode, and above average chromatic abberation. These issues are present both with normal and infrared photography (CA usually translates into more corner softness for IR). I didn't notice any problems with autofocus or exposure metering with infrared light, it simply works, so no need for further comments (it is much easier than using a DSLR for infrared!).

Getting the exposure right. In Nightshot mode the H9 uses maximum aperture and a fixed 1/30 sec shutter speed. This is even more restrictive than with previous Cybershot models and the variable aperture of the lens (f/2.7-f/4.5) makes it more difficult to get the exposure right by adding ND filters. Too much light gives blown highlights. With insufficient light image quality suffers, probably because the H9 boosts the ISO value then. The H9 does not report ISO in the EXIF in Nightshot mode, so it is difficult to check what is really going on. In my first tests I used a Cokin filter holder with 89B + ND4 + ND8 filters, a combination that works well with other Cybershot cameras in Nightshot mode (the 89B is very similar to Hoya R72). My first results were very disappointing: lots of noise with smearing artifacts, bad contrast and totally unacceptable corner sharpness. Even on a 4x6 inch print the problems were clearly visible. In hindsight, using this filter stack was not a good start. The H9 lens has a smaller aperture compared to the DSC-F7x7 cameras, so exposure was probably insufficient. Using just the 89B and ND8 filters did not work, because pictures were overexposed. As I mentioned in Part One, there is no perfect solution for stacking IR+ND filters on the H9, so getting the exposure right is difficult. But with the small H9 sensor and its potential noise problems, getting it right is very imporant!

IR filter choices. I nearly gave up on the H9 for infrared, but before giving up I decided to check vignetting issues by holding a 58 mm infrared filter in front of the lens in Nightshot mode. To my surprise, using just a Hoya RM90 (a relatively strong IR filter), I got a very nice infrared image on the LCD display! The histogram shows that the RM90 gives almost perfect exposure in normal sunlight. No ND filters are necessary, so it is possible to hold just the Hoya RM90 filter in front of the lens and shoot in infrared. The noise problems are strongly reduced this way and there is no vignetting. So I continued testing mostly with the RM90. For more shooting comfort the Pemaraal Galileo II tube (not available yet) seems like it will be an excellent choice. There are probably easier ways to mount a 52 or 58 mm IR filter in front of the lens, if changing filters (or adding a converter) is not necessary. The bright 3 inch LCD of the H9 makes infrared shooting a joy. It is easier than ever to see the scene in IR. Without sunlight the RM90 would be too dark (increasing noise), but most IR shooting happens with sunlight anyway. A B+W 403 dual band filter (similar to XDP filter sold in the US) works OK too but usually the results with the RM90 had more punch. I also tested some cheap IR filters that are available on Ebay: the '950 nm filter' (similar in transmission curve to RM90) was too dark for the H9. It transmits 2-3 stops less infrared light than the RM90. The '850 nm filter' (similar to 87c filter) gives overexposure in sunny conditions (it needs an additional ND2-ND4), but would probably work fine when there is no direct sunlight. Adding one ND2/4 filter on top of the 850 nm filter may cause some vignetting at maximum wide angle (I didn't check this).

Lens quality. With the exposure/filter problem solved, we can continue testing and check lens quality for IR photography. A pleasant surprise is that the H9 (at least with the RM90 or similar filters) has no 'hotspot' like most other cameras in infrared, making processing very easy. The Nightshot images are nearly monochrome, and just 'auto levels' is enough to get a pleasing infrared image (check the examples on my website). The lens has relatively strong light falloff towards the corners in wide angle mode but this can be easily corrected in postprocessing if required. The soft corners problem at wide angle remains though. Any zoom setting below about 50 mm equivalent has problems with sharpness outside the center of the image, and with IR photography it is even worse than with visible light. Of course it doesn't help that the camera uses full aperture in Nightshot mode. For subjects where corner sharpness is critical the wide angle setting should be avoided. Check the examples on my website to judge for yourself!. Lens quality for visual and IR photography improves in the normal range and from about 2x zoom it is very good (better than my DSC-F717). Adding the benefit of the H9 image stabilizer, it is clear that for infrared photography at longer zoom lengths, the H9 is often a better choice than other digicams. After testing, my camera went back to Sony because of uneven sharpness across the frame, an issue that some other users have experienced as well. I will wait with my final verdict on H9 image quality until I have tested another H9 camera.

In Part Three, I will look more closely at options for pseudo-color infrared photography, and compare infrared results from the H9 with those from other digicams. Is the H9 a good choice for IR shooters, or is it better to buy a used Cybershot on Ebay? We will also check if H9 IR results can be improved by 'IR-modding': removing the blocking filter to enable IR photography in normal camera modes.

Here's a sample H9 IR image from the Test Gallery.

IR Test Gallery

Copyright Statement: All images and written materials published in this Guest Article Contribution are copyrighted and are the exclusive property of the author. Images and material may not be reproduced or used in any way without the written consent of the author.

Part Three

Home