Olympus E-400 noise compared to E-330 and Sony Alpha A100

[Ian]

I wasn't able to get over to the test studio this afternoon after all, so instead I did a quick and simple noise comparison test using the Olympus E-400 (with the new 14-42 kit lens) which arrived today, an Olympus E-330 (with 14-54 lens) and a Sony Alpha A100 (18-70 kit lens).

The pictures were taken in a 12 foot square room lit with three ceiling-mounted 60W household (tungsten) lights with the reflectors upwards facing. All the cameras were tripod mounted, set to aperture priority auto exposure mode, with the aperture set to f/7.1 and segmented metering mode. At ISO 100, the two Olympus cameras agreed the shutter speed should be 1 second and the Sony reckoned on 0.8 seconds. Auto white balance was used and JPEG recording was set to best quality. The Sony has a wider aspect ratio frame but all three were aligned to the same horizontal points. The E-330 is only 7.5 megapixels compared to the 10 megapixels of the other two.

Below you can compare the results of a 250x250 pixel crop from a near-central area of the frame. What you see is a section of a Gretag Macbeth ColorChecker SG colour test target perched on a mantlepiece in the room.

The Olympus E-330 (left), up to now, has represented the best Olympus has been able to offer concerning resolution versus noise, being decidedly superior to the 8MP E-500. I wasn't optimistic that the one third increase in pixel count would do the noise characteristics any good, but I have to say I'm pleasantly surprised with what I see here. The crops look clean and noise is, if anything, a good match for the E-330 at the higher sensitivities, with the benefit of superior resolving ability. The E-400 also compares favourably with the larger-sensored Sony Alpha A100. There is obvious fringeing in the Sony crops, something neither Olympus lens suffers from.

Comments here on the forum most welcome!

Ian

[Ian]

Re: Here are the full frame views from each camera...

Here you can see the full frame view from each camera compared:

[tarzieboy]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

no expert, but e400 looking good to me Ian,,,john

[Stephen]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

no expert, but e400 looking good to me Ian,,,john

Well I'm not so sure

[Ian]

Re: Here are the full frame views from each camera...

Here you can see the full frame view from each camera compared:

By the way, the walls in this room are pink, which is rather a tough challenge for the white balance systems on these cameras

Ian

[Stephen]

Re: Here are the full frame views from each camera...

By the way, the walls in this room are pink, which is rather a tough challenge for the white balance systems on these cameras

Ian

I certainly would not consider the WB an issue in this case

[Goatsmilkuk]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

E330 seems to be the best from what I can see, the noise looks less in the first tests as the iso gets higher. Plus the room shot looks the most natural.

Stu

[Ian]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

Put your cards in the table, Stephen

Ian

[Ian]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

What I see when I test cameras at higher ISOs is obliteration of the detail as the noise reduction systems ramp-up. What impresses me about the E-400 is that the detail (in the marked number scale in the samples, for example) has been preserved well at the higher ISOs. There is less noise in the E-330 - and so there should be as its pixels are less densely packed - but the extra pixels of the E-400 are retaining more detail at the higher ISOs. I see the Sony's noise reduction being more evident - there is a bit less clarity in the Sony image.

I'm absolutely sure the 10MP E-400 is doing a significantly better job at balancing noise reduction and detail retention than the 8MP E-500, so it shows that Olympus is beginning to catch up on the image processing technology, areas in which Canon and Nikon have enjoyed a healthy lead. It will be very interesting to see how the new Canon EOS-400D compares here.

Ian

[Stephen]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

Put your cards in the table, Stephen

Ian

Well if you insist
Firstly, I'm not a great fan of these tests where a section of an image is enlarged to extremes, but I can see perhaps why some like them.

Secondly the point of this particular exercise was to check on noise in the two cameras. So to even mention WB is pretty irrelavant as its self evident it was all done under far from ideal conditions. I would go so far as to say that all cameras would struggle under these conditions and surely the longish shutterspeeds won't help in any tests of noise etc.

So all things considered I can accept the results of this quick test as being no more than an indication, and that real world conditions would be much better, before people start shouting about this or that and drawing conclusions

Personally I feel that the E400 comes out worst here in the noise test, and the levels of noise at 800 and 1600 is not brilliant looking to say the least. Yet neither is it in the others. On balance perhaps the Sony has the edge.

[Ian]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

Well if you insist
Firstly, I'm not a great fan of these tests where a section of an image is enlarged to extremes, but I can see perhaps why some like them.

Granted - I knocked this up in double-quick time this afternoon and I don't claim it to be definitive, but it is interesting to some I did try at least to be consistent and I think, on the whole I have achieved this, so the results can be compared usefully among the three cameras tested.

Secondly the point of this particular exercise was to check on noise in the two cameras. So to even mention WB is pretty irrelavant as its self evident it was all done under far from ideal conditions. I would go so far as to say that all cameras would struggle under these conditions and surely the longish shutterspeeds won't help in any tests of noise etc.

A long shutter speed is a good test of noise. The room is representative of the lighting conditions that you might want to shoot in, without flash, at a higher ISO speed. I felt I should mention that the cameras were set to AWB as a matter of basic fact and that the walls were actually pink, rather than all three cameras getting the walls badly wrong

So all things considered I can accept the results of this quick test as being no more than an indication, and that real world conditions would be much better, before people start shouting about this or that and drawing conclusions

Yes, it was just a quick test, but it tells me a lot as I can relate the results to other cameras as the E-330 and A100 are known reference points. The E-330 is better than the E-500 by about a stop. I'm sure the E-500 would look pretty bad in this company, which reflects rather well on Olympus progress in getting on top of noise, which is regarded as a bit of an Olympus Achilles Heel.

Personally I feel that the E400 comes out worst here in the noise test, and the levels of noise at 800 and 1600 is not brilliant looking to say the least. Yet neither is it in the others. On balance perhaps the Sony has the edge.

Well, there I have to disagree. The A100's noise reduction is working harder than the E-400's and though there may be less measurable (or visible noise - but I reckon only marginally) I think it's at the expense of the detail. In case anyone is wondering, the Super Steady Shot stabilisation system was switched off as the A100 was tripod-mounted. On paper, the A100 should have won this test by a clear margin as it has about the same sensor pixel density as the E-330 but more pixels to record finer detail. In the end it was a close run thing between the A100 and the E-400, which in theory had the least ideal spec.

I'm sure you will agree that a much better test is to take a selection of shots under different conditions and then to print them and, indeed, that is what I aim to do with all cameras I test.

Ian

[Stephen]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

Well, there I have to disagree. The A100's noise reduction is working harder than the E-400's and though there may be less measurable (or visible noise - but I reckon only marginally) I think it's at the expense of the detail.

I'm sure you will agree that a much better test is to take a selection of shots under different conditions and then to print them and, indeed, that is what I aim to do with all cameras I test.

Ian

I'm not about to argue the outcome, I accept the point of the test and if we accept your reasoning about the Sony, then fair enough. My comment was based on my impression from the squares on the test card as it seemed to me. I'll certainly be interested in further tests.

[Stephen]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

I've just been looking at someof the samples from Japan from this camera and spotted a shot taken at ISO800. Seemed perfectly adequate so far as I could see,without any prominent or worrying noise. A real world example without having to pixel peep LOL

[Ian]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

Right! I see what you mean - I think! Of course in a recent editorial I basically questioned the sense in 'pixel peeping' - especially with the latest very high resolution cameras. And I stand by that - I firmly believe that to really appreciate the quality of your own camera you should make prints, or at least appreciate the whole image.

But that's not what I was doing yesterday. The sole purpose of the exercise was to compare the noise characteristics of three cameras throughout the ISO range and the only method I had at my disposal to was to present the results on the Web by creating a matrix of small 100% crops.

I fully accept that this doesn't evaluate the overall image quality - far from it and I didn't intend to do that at all. All I was interested in was the noise and how it affected resolving power and the driving force behind this was the relatively poor reputation that Olympus has gained in its DSLR image noise.

Noise is something that people do obsess about. In many cases it's not that important to overall image quality, but it can be argued that Olympus need to up their game in terms of noise levels and on paper the increased pixel count of the E-400 was a retrograde step. I've always felt that Olympus DSLR noise was not a fundamental issue to do with the slightly smaller Four Thirds sensor size, but more to do with the way the sensor signal is processed. To date, Nikon and Canon have done a much better job in this respect, but if my hunch is correct, Olympus could achieve improvements. Olympus said of the E-400 that they had made useful improvements and the results actually back these claims up.

Ian

[Rob Schroeder]

Re: Olympus E-400 noise compared to E-330 and Sony Alpha A100

Hello!

Having talked a bit about E-400 noise already in several German forums and being one who sometimes at least tries to produce such comparisons himself, I do find your results quite interesting, making me join this forum for a couple of words I'd like to add as a comment. My own interest in high-ISO, by the way, comes from my own amateurish "ambitions" in available light photography as well as hand-held telephotography, which I currently do mainly with an E-330 (after having used all earlier E-System models for some time and still keeping an E-500 for backup purposes, for the time being), and from the limits I frequently do reach there notwithstanding the availability of some fairly fast glass. First thing, I too find it good to see that the E-400 seems to do slightly better than the Alpha, and that it seems to go along well enough compared to the E-330.

That said, I recently got the strong impression that the Alpha is one of the worst-performing DSLRs I've ever seen, in terms of high-ISO capabilities. A couple of weeks ago I had the dubious pleasure to compare ISO 1600 shots of the same scene, one from the 6 megapixel Dynax 7D, the other from the Alpha, both shots made under matching settings and conditions. To my surprise, it was clear to see that the ten megapixel Alpha, while reducing noise itself to an acceptable amount, was not able at all to deliver any more detail than the D7D, while the D7D even produced a nicer, more contrasty overall impression with better saturated, naturally looking dark shades, where the A100 only showed speckled patches of gray. And I've found the E-400 samples which have been published so far at least to hint in the same direction - while your comparison cannot completely dispel such doubts.

When I recapitulate my own earlier attempts at high-ISO comparisons, I now find them lacking in several aspects, especially since I've started to use RAW on a more regular basis. Comparing JPEGs (which I gather is what you've been doing here) primarily shows what the JPEG engine's noise reduction algorithms do, not what the camera (or even the sensor) is able to deliver, in terms of detail resolution on one hand, and in terms of noisiness or lack thereof on the other. Which is why I've decided for myself to do future comparisons primarily from RAW, developed with closely matching parameters. Even though I'm aware of the fact that this still might not be enough to achieve something like "perfect comparability", which is probably not achievable at all. A stronger antialiasing filter on one camera, for example, would make stronger sharpening necessary to achieve results matching those of another camera with a weaker AA filter - but stronger sharpening would produce stronger noise as well, in spite of exactly the same sensor being used.

Back to your comparison. I notice that your E-330 samples do show a more pronounced noise, especially chroma noise, than I've become used to with my own camera. Did you have high ISO NR activated or deactivated? And how was the E-400 set in this regard? More criticism: While the crops are well enough suited to show the amount of noise in larger greenish-brownish areas, I find them suboptimal when it comes to judging resolution and sharpness of fine detail. Something I really would like to know is how the E-400 might compare to the E-330 in that latter aspect. After reducing noise to a comparable amount, or without reducing noise at all, is there really a visible advantage in achievable detail resolution, when looking at fine details? The detail your crops show seems to be not fine enough to judge.

Anyhow, thanks for the comparison, which positively adds something to the first-hand information which is available on the E-400 so far.

Cheers,
Robert