Sony A350: Full-Time Live View at 14.2MP
by Wesley Fink on April 3, 2008 3:00 AM EST- Posted in
- Digital Camera
Resolution and Sensitivity Tests - Sony A350 vs. Pentax K20D
All crops represent a view of 230x300 actual pixels cropped from the larger 14.2, 14.6, or 12.2MP images. The crop area on the 1.5X multiplier Sony A350 and Pentax K20D are represented by the red rectangle on the full image above. Since the Canon 5D is a full-size sensor the coverage of the 50mm is greater on the 5D than the two 1.5X multiplier sensors. Therefore two sets of crops and full images are presented for the 5D. One set is taken from the same location and the 50mm lens provides a greater field of view on the 5D than on the 1.5X multiplier cameras. The second set of 5D images were shot with the same camera and 50mm lens moved closer to the image to try to maintain the same field of view. Despite the different fields of view, all Canon 5D cropped images are still maintained at 230x300 pixels. The comparison to the Canon 5D sensor is presented on the next page.
Links to the full JPEG images are also available on each camera sensitivity crop. These files are huge, but they can be downloaded for those who wish to view the actual images or explore EXIF data embedded in each image. A shareware EXIF viewer, OPanda IEXIP 2, is available at for download at Opanda.
| ISO Comparison - Sony vs. Pentax | ||
| ISO | Sony A350 | Pentax K20D |
| 100 |  |
 |
| 200 |  |
 |
| 400 |  |
 |
| 800 |  |
 |
| 1600 |  |
 |
| 3200 |  |
 |
| 6400 |  |
|
Click on any of the above image crops for the full image. Note: Full size images are between 3.5MB and 11.4MB! |
||
The Sony A350 exhibits very accurate color with the tungsten setting at all ISO settings. Noise is well controlled through ISO 800, with little increase in noise. By ISO 1600 we begin to see increased noise, but considering the crops are larger on screen than a 20x16 enlargement we supsect the images would still be very useful at ISO 1600 even for enlargements. By ISO 3200, noise has become an issue, but color is still very accurate at the Tungsten preset under tungsten light. ISO 3200 images are usable for small prints but are much too noisy for enlargements. The K20D noise is well controlled through ISO 3200. Noise at ISO 6400 is comparable to the Sony A350 at ISO 3200, making images at ISO 6400 useful for small prints only. Color on the Tungsten preset remains warm in some colors with the K20D, much like the Canon 5D crops shown on the next page, but whites are very clean. Custom white balance would probably do a better job than the Tungsten preset on the K20D.
Facebook
Twitter
RSS
113 Comments
View All Comments
strikeback03 - Monday, March 31, 2008 - link
I didn't think that was possible. Sounds like it falls in the same category as the optical viewfinders some compact cameras have, designed for emergency use only. I can't wait until some owner, blindly going by what is recommended as "best" on the forums, is trying to shoot with a 70-200/2.8 on an A350 held out at arms length.Wesley Fink - Monday, March 31, 2008 - link
The Olympus E-510 and E-410/E-420 actually have a viewfinder magnification of .92X, but the smaller sensor with the 2X crop factor makes the viewfinder look smaller than competing 1.5x/1.6x APS-C cameras. As mentioned in the review Oly got around this on the E-3 by going to a 1.15X multiplier. The E-3 has a great viewfinder that still needs a bit more eye relief to bbe nearly perfect. By comparison the A350 has a viewfinder magnification of .74x on a 1.5X crop camera.Deadtrees - Monday, March 31, 2008 - link
As many others have pointed out, the test is just wrong.It's so wrong that it even makes me wonder about credibility of computer hardware tests done here in Anandtech. Well, no, I do trust hardware reviews but not this one.
1.How can you possibly claim that it's the resolution test when Sony camera had its aperture set as f/2 and other set at f/1.4? On 50mm f/1.4 lens, the difference it makes is quite huge. It'd just be retarded to claim that this camera has better or worse resolution when the apeture value is full 1 stop different on a 50mm f/1.4 lens.
(I checked some pictures' EXIF data and it seems like they have all mixed apeture values. That's just so wrong)
- Use fixed aperture value.
2. Do you know that the actual ISO level can be quite often different than the others? For instance, 5D has higer ISO level than the other cameras: ISO 3200 on 5D can be ISO 4000 on the other cameras. Dpreview mentioned it as well.
- Show at least basic EXIF information regarding exposure setting so that readers can at least think about it.
3. Why the claimed Sony A350 ISO 1600 shows exif ISO level of 800?
Also, A350 was set to use manual WB whereas 5D was set to use preset tungsten WB. Based on this, how can you claim "Canon 5D is somewhat warm at the Tungsten preset compared to the cooler and more accurate rendition of the Sony A350."
Well, I gotta go now so I'll finish here for now.
Basilisk - Monday, March 31, 2008 - link
I understand you're still working out your aperture/f-stop/ISO parameters for testing. When you've decided on those, you might want to increase your available illumination considerably to get shorter exposure times: 1/6 second seemed unnecessarily long, but that is indeed what the EXIF indicated. Or, perhaps you're setup is immune to vibration?Unlike an earlier poster, I'm not so sure camera reviews are a natural extension to your other hardware reviews. A few other sites have been doing those pretty well for some time....
Wesley Fink - Monday, March 31, 2008 - link
We have already reshot the K20D and 5D tests using a fixed aperture and remote shutter release. It is not possible to reshoot the A350 since it has already been shipped to another user. Otherwise we would reshoot the entire sequence using f4.0 on the f1.4 leneses. We have talked about the changes that will be made in future reviews and we have definitely listened to comments and suggestions.The EXIF information is available (and has always been available) in the full images that can be downloaded by right-clicking any of the crops. You can see those for yourself in the review. Our conclusions were based on analysis of the full images and not the crops. The crops were designed to show details on the web and we agree they don't communicate noise well in theri current form - that's why we're updating procedures. The EXIF for the 5D on H, BTW, reports 3200 ISO. We did check all the EXIF values and we used the latest 1.1.1 BIOS just released a few days ago by Canon.
haplo602 - Tuesday, April 1, 2008 - link
hmm I still get f-stop 1.4 on the pentax 100iso shot (I did not bother to check the others). truth is, I am using the builtin opera image properties option, but it looks like exif data to me ...Deadtrees - Monday, March 31, 2008 - link
You haven't replied to some of my crucial questions.1. What's the point of ISO 1600 comparisons when A350 was set at ISO 800?
2. A350 had Custom WB setting and 5D had Tungsten WB:
Given that, how could you possibly claim that "Canon 5D is somewhat warm at the Tungsten preset compared to the cooler and more accurate rendition of the Sony A350"
3. What's the point of this review when you said "It is not possible to reshoot the A350 since it has already been shipped to another user?" Should readers just suck it up this false review because you can't redo the review? I think it's just better to take it down as this review just makes fool of Anandtech. Oh well, after all, it's the April Fool's day, isn't it?
BTW, your answer on the second paragraph didn't make any sense. Not only you did not understand what I was saying, you came out with some weird answers for what I haven't asked. Why are you giving me the instruction of how to view EXIF info? Why are you telling me the point of having cropped images and that the conclusion was based on full sized images? I was not talking about those and I didn't say that the review is false because of that.
Sorry if I'm being harsh but as you and Anandtech Staffs know, this is Anandtech. Though it's not a site about cameras, the credibility you guys gained over the decades does have a deep impact over readers.
I really think camera reviews should, at least, meet half of hardware reviews.
Wesley Fink - Monday, March 31, 2008 - link
1. The ISO 1600 being an 800 ISO image is an honest mistake. We uploaded over 127MB of full images for the 3 camera comparisons to our servers and that file name was incorrect. I will replace the 1600 with the correct file as soon as this comment is posted.2. The A350 ALSO used Tungsten white balance. I just checked the A350 files with OPanda EXIF 2 and the files are indicated as a Tungsten light source and "Custom White Balance".
The Canon is almost famous for its poor performance on Auto WB in Tungsten lighting. On AWB it was extremely orange in cast. Even on the Tungsten setting it is warmer as we were comparing Tungsten WB to Tungsten WB. A custom light temperature would probably work better on all 3 cameras.
3. The review is definitely NOT false, and we stand by our conclusions as we looked at many more images than those published here in reaching our conclusions. Our procedures can definitely be improved. We've committed to doing that in future reviews and we have listened carefully to all suggestions. Cameras are not given to us by manufacturers to hold indefinitely, and we generally have to beg manufacturers for a few days of use to review new cameras. That will change in the future, but short of buying the cameras we evaluate we have them for a limited time.
strikeback03 - Monday, March 31, 2008 - link
It's not an EXIF issue, it's that the actual sensitivity of the camera is not what is indicated (when compared to a good light meter).http://www.dpreview.com/reviews/canoneos5d/page21....">http://www.dpreview.com/reviews/canoneos5d/page21....
qbfx - Monday, March 31, 2008 - link
Hmm, I thought it was in the same price and performance class as the Canon 450D and the Nikon D300, why are they comparing it to the 5D ?? I still think the 400D and the 450D are better