Professional Photographer Yamada Kumio has just posted a page (in Japanese) giving a brief report of his experience with a beta EOS-D30 along with several sample images and some head-to-head samples comparing the Canon EOS-D30 with Nikon's D1 (the results may interest you). DPR has exclusive permission from Yamada to bring you this article translated in its entirety.

Rough translation of Yamada Kumio's Canon EOS-D30 report:
(Apologies for the pigeon English but a the same time as translating the report I'm also trying to reflect Yamada's comments in Japanese). As we are currently awaiting Yamada's permission to reproduce the thumbnails used in his review here you may wish to open the above URL and put the window side-by-side with this translation. My comments are highlighted in green.

Canon EOS-D30 Beta version image quality / sample photography report

The photographs taken from real life (image samples) open to the public for the first time here and possibly in the world. The samples here are taken with a beta version of the EOS-D30, although this is the very latest version available (to EOS-D30 testers). Wanting to publish sample images for a long time, this is the latest version state which is very close to production. I wanted to explore image quality which is an unknown area at this time, although the "D30" has already been exhibited at various events and reports. This report has only been made available to the public after Noon on August 5 (Japan Time).

Photographs taken from life by the latest version

[ 4 thumbnails of samples ]

At last it was possible to to take samples for the evening yesterday, about one hour of photography with the latest version. To be honest I had a chance to take samples with the "version under development" during the first ten days of July. As for the version used this time the image quality has improved considerably since that first early sample. The quality of this latest version is felt to be good enough to allow publication of sample images to the public (even though my time with the camera is limited).

Even though the picture quality matters, to be able to produce a good image quality in a digital SLR at this price is impressive enough. It was a surprise and dynamic range is very large and image is very clean with no noise, in fact better than forecast. Pictures seem very "natural" without digital peculiarity, some may find it soft.

The photographs taken is similar to the EOS D2000 in that images although they may seem soft are improved (to my initial intention) by post processing. This can be seen as very different than the approach seen on FujiFilm's S1 Pro which produces a more processed image. This difference is similar to the difference betwen colour negative film and reversal films. Moreover peoples worry over the narrowness of dynamic range with a CMOS sensor was not seen by me. Noise, the worry of old CMOS sensors was supressed to a level at which it cannot be seen unless of course at the extreme exposures (Phil: High ISO / Long Exposures). A sharp image can be achieved by appling a unsharpen mask before the time of printing.

Phil: There's an awful lot of detail in the D30 images but it appears Canon aren't applying any in-camera sharpening (which will please the purists), get one of those samples into a photo package and apply a light unsharpen mask (75%, radius 0.6 pixels, threshold 2 levels) and the detail will jump out...

[First set of Canon EOS-D30 vs. Nikon D1 samples at ISO's 100 - 1600,
scene of a Japanese street]

[200% Crops of above samples]

Both cameras settings for these samples: 35mm F2.0 lens, JPEG FINE mode. Program AE, White Balance Daylight. D1 colour space conversion: none.

The expectation of CMOS was for high noise, and the relationship between sensitivity and image quality was always in question when the machine was first announced, however here you can see in comparison with the machine of 'maximum rivalry', the Nikon D1 (which is my personal property). No colour space conversion has been made for the D1 samples. To me a very beautiful image can be seen at ISO 400 and even at ISO 1600 noise is unexpectedly low.

Phil: First impressions of the EOS-D30 sample compared to the D1 is that it appears softer (though that can be rectified with a very light unsharpen mask), that colours are in a better "colour space" than the D1 (less magenta) and that visible noise at high ISO's is no worse (or at least it's of a similar level) to the D1. Consider the price difference and that could indeed be interesting news for Nikon.

[Second set of Canon EOS-D30 vs. Nikon D1 samples, comparing
dynamic range at different sensitivities, two smiling children]

Both cameras settings for these samples: 35mm F2.0 lens, JPEG FINE mode. Program AE, White Balance Daylight. D1 colour space conversion: none.

[Third set of Canon EOS-D30 vs. Nikon D1 samples, again comparing dynamic range,
shot upwards of buildings with a bright blue sky behind]

Both cameras settings for these samples: 35mm F2.0 lens, JPEG FINE mode. Manual Exposure, White Balance Daylight. D1 colour space conversion: none.

The traditional understanding of CMOS was a narrow dynamic range, although it seems from these samples that the dynamic range is great and (at least with these few samples) is not inferior to the Nikon D1. As for saturation it seems natural and close to original colour.

Phil: Dynamic range looks good, the worries over dynamic range related to a CMOS sensor don't seem to have caused a problem in the reality of the final CMOS sensor used in the D30.

[Nikon D1 samples for comparison with sample images at the beginning of this article]

Marvellous Noise Reduction Function

[Three sample images taken with 8, 15 and 30 seconds exposures]

The noise reduction function is a custom function which can be set on the D30. This function corrects for noise generated from very long exposures. Noise on long exposures can be seen as a big fault of digital cameras, using noise reduction the image and the noise are distinguished and as such noise is removed from the final image.

Phil: Impressive, although there are still a few "hot pixels" in the 30 second exposure overall things look good with the D30 if you love to shoot night scenes.

This report centered on the image quality as a first impression. And the model used for these samples although not production is very close to final production model, though some tuning may still be done before the D30 goes on sale.

All text (c) 2000 Yamada Kumio, All Sample Images (c) 2000 Yamada Kumio, accept no responsibility for inaccuracies incurred during the translation process, original text available here in Japanese.