Sorry Ray, I hadn't realized you were making fractions.  

Imperial measurements in machining are commonly done to 0.001 inches or in thousandths.  So 25 thousandths would generally be interpreted as 0.025", not 1/25 000.

Now, most measurements are done in metric, even here in the US.  But most of the old school people still 'think' in imperial units.  It's easy to remember 1 mm=0.040" and do a quick conversion.

Just me, putting my personal biases and preconceptions into the discussion.

I had expected the Sony 828 (8mp) images to be bigger than those from the 10D (6mp) when comparing 100% crops. But they look exactly the same size. Can Michael explain the methodology he used to find comparable focal lengths ?

Ray,

I'm allergic to shooting stitched scenes.

I'll probably do it occasionally with the 828, perhaps 2 or 3 shots, but I'm just not someone who can visualize without the image showing up on the live LCD.

Absolutely! Lovely rocks! I wondered how long it would take you to jump in and defend the reputation of your 35-350mm  :D .

Bob,

I've made 13x17 prints with my 828 and it blows the 5MP cameras away as far as resolution goes.  This camera eats resolution for lunch.

I like using the 100% coverage live LCD for composition -- this is the biggest reason I chose the 828 instead of a drebel.  I also prefer 4:3 instead of 3:2.  The resolution is better than the drebel in the short dimension, it's a wash in the long dimension.  But since I typically print 4:3 prints or 5:4, the 828 has a bit of a resolution.

However, there are some drawbacks to the camera.  The biggest one for me is the CA, although I found it only shows up in a minority of shots I take (20% or so) and even there (unlike Uwe) I was always able to clean it up easily with a photoshop adjustment layer (saved to a function key action, no less).  I didn't try Uwe's method (his method never worked with me on any camera) instead desaturating the exact color of the CA and darkening it a bit in a saturation adjustment layer.  If you did have an image with violet you would have to use the adjustment layer mask, which would be more painful.

I sent my 828 back in hopes of scoring a unit with less CA.  It should show up next week -- I'll let all know whether it is any better.

I'd like to thank Michael for his nice review of the F828, I just bought this camera and I share Michael's enthousiasm and concerns.

One thing I've found is that there is quite a discrepancy between the ISO sensitivity of my F828 and my Canon 10D, the latter being twice as sensitive with the same ISO rating. This translates the F828's ISO 100 to something like an ISO 50 for the 10D, making the F828 pretty useless in low light situations, despite of its fast lens.

I owned a Canon G3 previous to my 10D and I remember being able of taking some very usable indoor shots with it, with my new F828 I can't get anything useful indoors without using a flash...

To be fair, the 828 matrix meter underexposed that scene a bit.  You can push the histogram hard to the right without blowing out the whites on this camera, and get a lot less noise than this.

Ray,

I suppose when you are talking about enlargements past 16x20 or so I am a "die hard" 6x7 MF and LF fan.  Personally I am not going to use a view camera right now -- it doesn't fit my working style nor budget.  But for big enlargements it definitely makes even the 1Ds look poor from a details standpoint.

Now with prints up to 13x17, the 1Ds is very impressive, and even 16x24 prints look good.  But the detail on a 4x5 inch plate of provia is a force to be reckoned with.

I think the 4/3" cameras will be around for some time to come, if only because you can make a very nice, high resolution camera with two rather smallish, very high quality lenses cover 28-400mm.  If they get IS (or the rumored IS sensor or teleconvertor) then the system will be very appealing indeed for bird photographers.

Karel -

I'm in the same boat as you.  I don't have extremely deep pockets and have looked for the best fit camera for my budget.  I chose the A1, not saying that it's the camera that would best fit your needs.

I was willing to spend up to $1kUS for a new camera, didn't really want to spend that much.  I knew the conditions under which I normally shoot, how large (and how little) I actually print.

I could have gotten a dRebel/300D for my money.  But I would have had a marginal lens.  So I would have ended up spending even more for a good quality IS lens.

So I've got an image stabilized camera with a 28-200 lens and it only weighs a pound.  That means a lot more fun to me, less weight and more money for travel.

Just be careful with the F828, be convinced that it has the extra pixels and that they aren't wiped out in processing.

And I'm with BJL, dSLRs win in terms of exchangeable lenses, higher ISOs, and the hope of shallower DOFs.  The big complaint I have with fixed lens digitals is the lack of shallow DOF for isolating subjects.

Just thought I would add that the 828 can certainly use teleconversion lenses, and they work just fine.  I'd recommend sticking to the TCON 17 or B300 instead of using the much heavier HGD-1758 or TCON 14B.  Most of the other afocal teleconvertors used on digicams are complete crap and should be avoided like the plague (I owned 3 before ponying up for the B300).

Fabio,

I'm surprised you find the G3 more usable indoors than the 828.  Can you elaborate on this?  The G3 doesn't have an ISO 800 mode unless I am remembering incorrectly.

Michael, thanks for the review of the 828.  I'm very pleased that your unit did not display atrocious Chromatic aberration problems at the wide and telephoto ends of the lens.  As soon as I can get another unit I'm swapping mine out.

I'm happy that your problems with the camera were with the RAW mode and buffer issue and the lack of "shooting priority" with the top control knob.

PS I've heard rumors that the beta version of photoshop CS will open Sony RAW files.  Thomas Knoll was asking about SRF files a couple months ago and he mentioned the encryption problem, so I guess he managed to crack the encryption (or my rumor source is lying).

To my eyes also. Check out fig. 18 (crops of the Royal York). The noise is so bad on the F828 shot, the shadow side of the skyscraper on the right has lost all its windows. ???

By the way, Scott, 4x10-5 inches is one 25 thousandth of an inch (1/25,000). What are you getting at?  :)

I haven't mentioned it anywhere before, but I received a Fuji 7000 for review when it first came out a couple of months ago.

I never reviewed it because it was so terrible in almost every respect. I simply sent it back to Fuji with polite thanks.

It wasn't a patch on the Sony.

Michael

Absolutely! The F828 sounds like terrific value, but let's not kid ourselves that the extra noise, which is very apparent in all the sample shots in Michael's review, is not an issue. Noise is always an issue, which is why we have programs like Neat Image. But such programs, in my experience, always sacrifice at least some detail.

I find this review of Michael's inconsistent on the noise issue, especially considering the panning he gave the Kodak 14n. I don't recall any conciliatory note on this issue, that the increased noise at ISO 400, compared with the 1Ds, would not show up on prints, and was therefore virtually nonexistent.

Scott,

Let me try alleviate your confusion.

First of all, I already own a comprehensive Canon outfit. I don't own an Olympus E-1 system. And I wouldn't, because as I wrote in my review -- although it's a fine camera I feel that the format is a dead end, and that's why I wouldn't buy one.

Secondly, I didn't do a "comparison" when I did the E-1 review because frankly one would assume that the difference between it and the other mainstream DSLRs in that price range are subtile to non-existant in terms of final image quality. That was not the assumption that I initially made vis-a-vis the Sony and DSLRs, for all of the obvious reasons. I believed that readers, just as I did, would be interested in how it stacked up against a camera like the Canon Rebel, though I only had a 10D to compare it with. (Image quality between the 10D and the Rebel is essentially identical, so that's why I felt the comparison to be fair).

Finally, I wish people would get off the "noisy F828 kick". On 13X19" prints at ISO 64-100 there is usually no more visible noise than from Canon 10D images. Really! Why would I say this if it weren't true, since pretty soon tens of thousands of photographers will be able to see this for themselves?

Higher ISOs? Sure, noisier. But I have done some stunning prints made from  F828 shots at ISO 400, once it's run through Noise Ninja (a 3 minute task) that again are comparable to Canon 10D output. (It's the final print that counts -- right?)

Here's the final point. The Sony F828, complete with its truely excellent 28-200mm f/2.8 lens and 8 Megapixel resolution, is positively tiny and feather weight compared to the equvalent combination of Canon 10D, 24-70mm and 70-200mm f/2.8 Canon lenses. It can fit in the pocket of my rain poncho. I could rock climb all day, or walk the street of a foreign city inconspiciously, and not even know it was with me. If It gets broken, dropped or stolen in some rugged or dangerous place it's almost cheap enough to be considered a "throw-away".

Yet, it can produce Super A3 sized prints that are in most ways are as good as those from a 6MP DSLR.

I hope that this answers the question. Now I'm offline till mid-January.

Michael

Your confidence in the 4/3 system is admirable, and it seems that everyone who supports the system has hedged their bets on all sorts of future developments for this system.

However, the situation right *now* is that Canon (and Nikon) have the fast primes, the IS/VR, the huge selection of high-quality lenses at various price points (i.e. one can choose between 16-35/2.8 or 17-40/4, 70-200/2.8 IS or 70-200/4, 300/2.8 or 300/4, etc), broad third-party lens support to further supplement lens selection, a wide variety of well-established bodies, and a very large user base (with its concomitant benefits like software support and strong used lens market).  It seems that the most prudent choice for the moment is still to stay with an established system unless you really *need* things like an 28-108 equiv. single-lens solution or smaller form factor, since presently all the above advantages I've listed are still on the side of the established systems.  Keep in mind that it's a risky move to be an early adopter to any system based only on what "will" (or will not) come.

And it's an expensive risk at that--who knows if the other companies won't suddenly change the rules, like Canon did with the DRebel?  Nikon's new 12-24 does knock some of the wind out of the 4/3 system's sails since it shows that Nikon is willing to use their lenses to play the small-sensor game while keeping their bodies backward compatible with existing 35mm lenses.  With even more products coming out, who knows how things will look after this year's PMA, or even next year's?

Having said that, the 4/3 system does sound great in theory and I do hope the 4/3 system turns out well, because competition and choices are good for everyone!  I found myself rooting for the D2h before the reviews came out because it meant that Canon will have to work that much harder on a 1D replacement that I might be able to afford myself.  Eventually.  Some day...

Ray,
  perhaps I can meet you part way, acknowledging the desirability of focal lengths up to 400mm for a certain niche of amateur 35mm SLR users, and hence of 100-400 (Canon) or 80-400 (Nikon) lenses for 35mm format. On the other hand, (a) this is a statement about 35mm format, and since Canon and Nikon offer nothing beyond 400mm for under US$5,000, that is clearly a rarely entered domain; and ( in 35mm format, wide angle coverage to 24mm and beyond is significantly more popular (and affordable) than telephoto to 400mm.

  At the telephoto end, the counterpart of 100-400 in 35mm format is about 50-200 in 4/3 format, and since the Olympus 50-200 is smaller, cheaper and faster than 35mm zooms reaching 400mm, we should be able to agree that this is a desirable lens for that same niche of amateurs.

  At the wide end, Olympus has an 11-22mm f/2.8-3.5 zoom coming soon, at a rumoured US$700, giving wide coverage comparable to between 20mm and 22mm in 35mm format, and they have also more vaguely announced a subsequent ultra-wide zoom for later this year (7-14mm, so reaching the equivalent of 13 or 14mm?).
  In contrast, for Canon's amateur level DSLR bodies, the widest Canon zoom lens (16-35) reaches only 25mm equivalent, and costs about $1300; nothing wider has been announced or even hinted at. This seems related to Canon's intention to target "APS" DSLR's at low to mid-level amateurs, while expecting more serious photographers to pay the for now far higher price of 35mm format DSLR's, so that they are not doing much to accomodate some needs of more serious amateurs who use their smaller format DSLR's.

Fabio,

I think we need to do a bit more testing before settling on it as gospel that the 828 meters 2 stops slower than the G3.  It could be a fault with one or the other of the cameras you tested.

In my testing the Canon cameras are close to a full stop more sensitive than other cameras in their metering / ISO ratings, but not 2 stops.

I'll see if I can run a couple of tests tonight if I get a chance to run by the local electronics stores.

Intended use versus fit for use....

Being very interested in the F828 or Minolta A1 or Digital Rebel, I cannot agree with Michael that the level of noise and CA is not important in the F828.

To me the F828 seems to be intended (of course I have no idea for which audience/use Sony designed the camera) for someone who wants a relative compact (thus highly portable) camera offering all kinds of settings and deliver high quality images (the price...). A camera you can take with you most of the time to most places. Photograph landscapes, city/street, ... any use. This is entirely different from a situation where you can use flash power (studio) or tripod (landscape) and where it does not matter that you must choose low iso. For a camera like the F828 I would like to be able to use as high an ISO setting as possible in some situations (the Minolta A1 anti-shake might help also).

May I suggest you take a look at a sample image (over 3 MB!) at steves-digicams:
http://www.steves-digicams.com/2003_review...es/DSC00606.JPG
look at the ceiling in the upper left corner, then look at the hand of the guy behind the keyboard and tell me if you are happy with what you see. For reference, check the Digital Rebel/300D sample images on the same site (e.g. the one of the couple at a party with fill in flash, look at the people outside the flash range).

For portability (carry everywhere, anytime) I am very happy with my 300v/analogue Rebel with the 50mm/1.8 so I think I will go with the 300D or whatever is available after the Photokina this september. It should be a "hot" year with the PMA and the Photokina  :-) although I would have loved 28-200/2.0-2.8....

Bob,

I already sent my 828 back to get one with less CA.  So obviously I don't just look at the camera with rose-colored glasses.

There are already reviews all over the web, including at A Digital Eye where you can compare images shot at the same time between the 828, A1, and other cameras.  The resolution difference is easy enough to see in the samples.

But then again you have the A1, so I guess you too "have a dog in this fight" too.

If I had to recommend one of the compact digital cameras at this point to the average photographer, I'm not sure I would recommend the 828.  The CA in enough of a problem on at least many of the units that a casual photographer who didn't know his or her way around Photoshop would be suffering nasty purple fringing on a fair number of shots.  My hope is that some of the units have a lot less CA and that Sony can tighten up their QA process.

The A1 has a lot to recommend it -- I'd probably tell someone getting a compact camera to look at it before the 828, unless they needed the higher resolution or better video of the 828 and were able to handle the fringing issue.

We've been investigating over at DPReview STF and it looks like there are two classes of digital camera ISO ratings.

On the one side are Canon and Minolta's digicams (and Canon dSLRs) whose ratings are more sensitive.  Sony, Nikon, Fuji and Olympus' digicams are in the less sensitive group, and are equivalent to film ISO ratings.

Generally ISO 200 on the less sensitive group is, more or less, as sensitive as ISO 100 on the more sensitive group.

However, if we are talking about landscape, I think we need to consider how the cameras perform in their lowest ISO, as that is where the overwhelming majority of landscape images (particularly with digicams, whose smallest aperture is typically f/8 or f/11 instead of f/22 - f/48 for dSLR lenses).

And I would have to say that the 828 just destroys any 5MP digicam as far as resolution is concerned.

The good news is it looks like Nikon and Minolta will probably be releasing 8MP digicams in the near future.  So you'll be able to have other options than the Sony 828 with (hopefully) equivalent image quality.

Michael,

The histogram (and exposure preview) are inaccurate with exposures longer than 1/2 second.  Not sure why that is, but it works that way with the 717 as well.  Live Histogram gets progressively darker as does the live preview, even though after you take the shot the exposure is right and the post-shot histogram shows the correct values.

There does definitely seem to be a metering discrepancy vs. the Canon dSLRs -- although in my tests I'd estimated it to be 2/3 stop instead of a full stop.  That would put ISO 200 on the Sony about equal to ISO 130 on the dSLRs.  Perhaps if you look at this you can do more rigorous testing -- I just was playing around in a store.

I agree.
Michael states that at 64 ISO the noise is "essentially non-existant", but to my eyes the images posted are very noisy, similar to 10D's 800 ISO !  ???

The lens seems very sharp so kudos to Zeiss for this little gem of optics (7x fast zoom which has to work on such a tiny sensor), but this new RGBE CCD didn't impress me...

And now I read that the ISO sensitivity is even less than the nominal ?
Are you saying that the samples posted at 64 ISO should be compared with something like 32 or 40 ISO ?!?
Well... these have always been the correct settings for Velvia 50, so maybe we are back in film times  

Ciao  

Marco

Well, here's another angle.  I, for one, was sort of surprised by the qualitative difference between Michael's reviews of the Nikon 828 and the Olympus E-1.

The 828 review, while "harsh" in well deserved areas, was overall a polemic in defense of this camera, warts and all.

The Olympus E-1 review, on the other hand.... seemed to grudgingly admire certain aspects of the system (and also being critical where appropriate) while dismissing it as a serious image making machine, the 4/3 system and its potential, and Olympus, more or less "out of hand."  

I thought the images Michael captured with the E-1 were stunning, compared to the noisy stuff shot with the 828.  Also, conspicuously absent in the E-1 reivew are the kind of comparisons found in the 828 review, vis the 10D.  It seems to speak of "not caring to bother" that supports my paragraph above... dismissed out of hand.

I, for one . . . honestly don't get it.  This is not a rant, or bash, or any such thing.  I just honestly don't get it.

Seems to me like a trip to Africa where one could take a camera w/ two lenses and a 1.4x extender and cover 28-560mm+ equivalent focal length, at less than half the cost and and significantly less weight than a 1Ds with equivalent range... well, "I" think it would be a great thing and a wonderful test of a "whole new system concept."  Instead we have a noisy 8mp digicam off to do "serious work."  As I said... I don't get it.

Cheers,
Scott

Found the review and the discussion interesting. One question though. There seem to be one camera that has fairly similar features to the 828 (maybe except for build quality), and that's the Fuji S7000.
Although it has "only" 6+ megapixels it's the closest competitor to the 828. It's also equiped with a EVF, a similar zoom (in range) and raw support.
It would be interesting to see a direct comparision with the 828.

Ray,

The type of noise displayed between the F828 and the 14n are of completely different types.

On the Sony (with the very clear exception of ISO 800) the noise takes on the appearance of  "grain", similar to that of film. Using this as an analogy, on the Sony at ISO 64 it's almost invisible, maybe like Panatomic-X. At ISO 100 it looks like Plus-X. At ISO 400 it's like Tri-X developed in Rodinal.

The Kodak 14n on the other hand displayed blotchyness similar to the Sony at ISO 800, even at ISO 400.

So, while the fine "grain" of the Sony is either invisible to subdued with images from the Sony, the 14n's noise characteristics (at least from when I tested it) was much coarser, and therefore more visible and annoying.

Michael

Hope this helps.

Ray,
  about the Canon 100-400/4.5-5.6 IS; given that it costs almost US$1400 (over Au$2000?), your idea of "relatively popular" must be very different than mine. I stick to my story that the overwhelming majority of 35mm SLR photographers never use lenses going outside the range of about 20mm to 300mm, but beyond the "one lens entry level crowd", they typically aspire to at least 28-200 total, and to about 28-105 (or maybe 24-85) in a single "standard" lens.
  On your shift to discussing lenses for the 1Ds; its cost puts it completely out of the range of about 99.9% of SLR users, so it is irrelevant to considerations of the mainstream DSLR market. I think I am very typical of would-be DSLR buyers in that for my camera buying purposes, I restrict my considerations to DSLR systems costing under about US$3000 including one or two lenses and other essential accessories. It also strikes me that your proposed comparison of two camera systems differing in price by a factor of about four would not be very enlightening (unless the cheaper system were to do better!) And bringing it up takes us even further away from the 828, so let us return to that topic.

  About image quality comparisons between the E-1 and 828, particularly in respect of noise, you are joking, right? After the criticism you have passed on about the high ISO performance of the E-1 (only a significant issue at ISO800 and above, especially now that you have put noise reduction post-processing on the table), surely you must have noticed that the 828 is far worse in this respect (with problems starting maybe as low as ISO100 and certainly by ISO200). The difference is far more than can be compensated for by the 828's half to one stop advantage in minimum aperture ratios. Not to mention the frequent complaints about colour fringing (CA?) with the 828.
  I also find it strange that after you recent enthusiasm for extreme telephoto reach, you suddenly shift to discounting the difference between the 828's limit of 200mm equiv FOV with no TC available and the E system's option of zooming to 400mm equiv. FOV (before TC). Not to mention the current limitations of EVF's as in the 828; by contrast, I recently got the chance to look through the E-1's viewfinder, and found its image size, brightness and clarity very nice for my purposes.

  On the other hand, for the many photographers who are satisfied with a 28-200 equiv. FOV range, using an EVF instead of a TTL OVF, somewhat limited high speed/low capability, no option of ever using any prime lenses or of being able to upgrade the body without throwing away the lens too, the Sony 828 or Minolta A1 might be quite attractive options. So as a variant on your 828 to E-1 comparison, how do you think the 828's superfically impressive advantages in pixel count (and probably test pattern resolution) and aperture ratios, plus its far wider zoom range, compactness, more solid build, and all the cute digicam advantages like movie mode, swivelling live LCD VF and live histogram match it up against the comparably priced Canon 300D + 18-44/3.5-5.6 lens kit? I suspect your claimed enthusiasm for the 828 will wane at this point.

BJL,
If I could have got the Canon 100-400 for a mere A$2000, it would have been a no-brainer. I'd have bought it a lot sooner than I did. Expensive lenses do not sell like hot cakes, whatever their usefulness or focal length range. But those that are useful and seem good value, despite their high price tag, tend to sell quite well, and I think the Canon 100-400 IS comes within that category.

But rather than tackle you on each point, I'll just try to describe where I'm coming from since we seem to be coming from different directions, so to speak.

Trying to assess the desirability of a camera or camera system is no easy task for those who are serious about their photography and for whom photography is likely to be a growing interest or passion.

Each person has different aspirations and means when it comes to buying photographic equipment. I've listed below the issues that are important to me in the form of 6 guidelines which I hope might be helpful to others. These are my guidlines, distilled from my own experience. They cover my concerns. They may not cover yours. They are not in any order of importance.

(1) Weight, bulk and general handling.

(2) Build quality and other features of a more professional nature, such as MLU, high maximum shutter speed, waterproof seals etc etc.

(3) Range and quality of lenses and accessories, which of course will include both expensive lenses and cheap lenses, affordable lenses and not affordable lenses, popular lenses and 'unusual' lenses.

(4) Maximum system image quality achievable. Notice I used the phrase 'system image quality'. This allows for the possibility of increasing resolution by simple attaching a different body to the lens, as you can do with the Canon system. Of course, maximum quality only comes at a high price, but the point is ... do you have that option in the system you're considering?

(5) Company image and reputation for innovation. Ie. buying into a camera system which happens to meet present needs at the right price, might be shortsighted. Many of us know how easy it is to accumulate lenses and accessories over the years and then feel trapped in the system. We can't afford to switch.

(6) Perhaps the most important consideration of all, price and general value of the product.

I think anyone using these guidelines, and perhaps awarding points out of 10 in each of the 6 broad categories and then buying the camera/system that has the highest total, would not go far wrong.

I guess I'd rather have a few excellent zooms with good coverage than more bigger more expensive zooms to do the same coverage in a package that can't be carried long distances as easily.

Comparing a small sensor F828/A1 type camera with a dSLR, even the 300D is a compact pickup vs. full-sized, V8, 4x4 monster.  They're just not operating in the same categories.  

Smaller sensors have smaller photosites which gather less light per unit time.  The ambient electronic noise in the system is greater in relation to the signal produced by that smaller capture.  One can't amplify that signal (use ISOs) in the same way that one can amplify the signal from a larger photosite.

If you want a smaller, more compact, less expensive camera and choose a F828/A1 type camera it's probably best to make the purchase that you are buying the equivalent of film camera that won't accept high ISO film.  Be prepared to shoot at low ISO.

That said... If you're truly a landscape photographer, that is if you generally take studied shots, take your shots from tripod, etc. being limited to ISO 100 shouldn't be a big handicap.  You probably shot film in this range in order to get fine grain.

If the F828 does actually produce the resolution that one would expect from an 8 meg camera (the jury's out) then it should be a budget option for landscape photographers.

For my photography I need/want a camera that carries well.  I bought an A1 knowing that I probably won't ever shoot above ISO 100.  In-camera stabilization will help in low light conditions since I can't crank up the ISO.

I also drive a compact Toyota pickup.  I don't need to haul cubic yards of gravel very often.  I can make a couple of trips when I do.  ;o)

Well, Matthew, no offense, but I wouldn't consider you an objective reviewer.  You've spent actual money on a F828.  You've got a dog in this fight.

Me?  I'm just an interested bystander.  I didn't even seriously consider the F828 because I heard that it wouldn't be out until the spring and I wanted a camera now.  (OK a few weeks ago.)

But I am an  interested bystander.  I'm looking forward to someone who has to be as objective as possible shooting controlled test shots of the F828 and comparing it to 5 and 6 meg cameras.  

If it out-resolves them, Great!  Next time out I might take Sony more seriously, and people who really can't afford a dSLR have a less expensive option for good resolution.

I'm not sure why anyone would expect a 5 meg small sensor to resolve as well as a the 8 meg Sony.  

I think my question is whether you get the full 8 megs worth of resolution, whether you get 1/3rd more resolution than you would get with a 6 meg camera.

Hi,

I have just read Michael's review of the Sony F828, and I am very surprised: What I see is not what I read... ???

For example: Fig17 (comparison F828@64ISO vs 10D@100ISO).
From the comment on the lens ("almost identical") and the previous comment on noise ("Essentially non-existant"), it should be very close...
But when I look at those pictures, I see Chromatic Aberration (white edges of windows) and strong Noise at 64 ISO (dark area of windows - and everywhere...).

Is there something more to take into account ?
Did I miss something ?

Thank you for your answers,

Olivier

Noise can be an elusive beast.  I use an Olympus E-20, a camera notorius across the world wide web as being too noisy.

If I get the exposure right, then visible noise is not an issue, it is tough to spot on screen at 100%, and won't show up at all in a print.

Underexposures can be a problem, and shadow areas can be as well.  It depends on how much correcting I have to do, and how much sharpening I apply to the image.  I can correct a certain amount by blurring a copied layer, and setting it to colour, but that doesn't always help.

Even then the final print may look ok.  It depends on the size of the print, and the method used to create the print.  I've found that noise will show up on a Fuji Frontier print that does not show up on an inkjet print.  I've read that this is because the ink on an inkjet print tends to run together and blur out some of the noise.  Looking at the screen, I can guess by the amount of noise whether or not it will show up in a print, but I always proof to make sure.

In general, I would say the noise issue on the E-20 is blown out of proportion.  I do occasionally throw out an image because of noise, but it is usually because I messed up the exposure.

The 828 is interesting to me.  I'm not sure I would buy one, but I'd probably consider it.  To me, the limitation is that you aren't working on a camera system like you would be with a dslr.

Don't forget the lens that comes attached to the 828.  Even putting aside lens quality, if you buy a 300D you will have to buy more glass to cover the focal length of the lens attached to the 828.  That will add a lot to the dollar part of the equation.  

To me, this could easily be a deciding factor.  I could put up with a lot of limitations in a camera (and am, come to think of it) simply because I can afford it, and it will be good enough.  A few hundred dollars may not mean much to some people, but it does to a lot of people.

Michael,

Thanks for the response.  In that context it makes perfect sense.  I not only "know about" your extensive Canon kit, I envy you on a daily basis!

I for one, hope the 4/3 system doesn't turn out to be the "Beta Max" of digital photography, ergo the dead end that you predict.  On the other hand, Olympus did abandon the OM system despite its success.  Their track record isn't good in the long term support-development department.  Only time will tell.  It does appear as though they are selling as many E-1's as they can produce, but I suspect this is true of all new-release digital cameras.

I also appreciate, and agree with, your comment on the difference between visible noise "on screen" versus in prints.  

I'd still like to see you get an E-1 system on loan from Olympus some time, and take it with you on one of your trips -- work the heck out of it, abuse it, make a bunch of prints, give us a thorough field report in an issue of the Video Journal.

I'm one of those people who doesn't have an investment in either of the "big two's" glass, and find the 4/3 size, weight, focal range to date ( and scheduled for release soon ) very interesting / appealing.

Again, thanks for the feedback.

Scott

Matthew,
I really hate to hit someone when they're just getting up, being such a gentleman, an' all that , but there are too many things about this 4/3rds system that don't stack up, so I'm going to compromise my principles.

I can only look at things from my own biased perspective of course, but when I consider the broad areas of performance that are important to me in a camera system, the Olympus E-1 doesn't have enough going for it.

The jewel in the crown is without doubt the 300mm F2.8. This would seem ideal for birds and wildlife, especially with the 1.4x converter. The 35mm equivalent then becomes 840mm F4, which is something to drool over.

The trouble is, that Zuiko 300mm lens is very expensive. The lowest price I could find on the net is US$7000, which is about the same as the lowest price for the Canon 600mm F4 IS. Sure that Zuiko 300mm lens is lighter, but not as light as you might think (3.3Kg as opposed to 5.3Kgs for the Canon).

As fine as the Zuiko lens might be, don't tell me it can compare with a 960mm F4 which is what the Canon 600mm lens becomes attached to a 10D. Okay, that extra 2Kgs can be a pain. Someone who's frail and wealthy might be willing to sacrifice a bit of quality (and/or focal length whichever way you look at it).

But the point I would make is, with a Canon system no sacrifices are required, not in terms of price, not in terms of quality, and not always in terms of weight.

If the Canon 600mm lens is too heavy, then go for the 500mm F4 IS. This lens is definitely cheaper than the Zuiko 300mm, no heavier, surprisingly, and I would suggest still gives better quality images attached to a 10D or 300D (800mm + 6MP as opposed to 600mm + 5 MP).

However, if you really want a lightweight system with perhaps only a slight sacrifice in quality, then a 300D attached to a Canon 400mm F5.6 prime would be an excellent choice. At a 640mm 35mm equivalent, this lens weighs only 1.25Kgs as opposed to the Zuiko 3.3Kgs. And best of all, it's a lot, lot cheaper. Of course, the Zuiko lens is 2 stops faster and that's a big advantage. But it's also much, much more expensive and heavier. So where does that leave us? Two steps forward and two steps backward  ??? .

I find it difficult to make a case for the E-1 system.

Well if the smaller form factor is that important to you and you're willing to take the risk, by all means enjoy the Olympus if and when you decide to take the plunge.  Maybe Olympus will even top everyone with a 14-100 (28-200 equiv) lens that'll be exactly what you were looking for, who knows.

I don't currently own any lens longer than 70mm on my DRebel, so the whole "lens too heavy to carry" phenomenon is lost upon me.  No bird or squirrel shooting for me.

Welcome back to the fray, BJL!  

"... rarely used extremes of telephoto reach." ??? I think you'll find (if you take the trouble to do the research   ) the Canon 100-400 IS zoom is a relatively popular lens and probably a minimum requirement for anyone interested in photographing wildlife. This relatively cheap lens takes you beyond what is currently available within the E-1 system at an affordable price.  

The notion of a single, high quality zoom that covers all your requirements must be the Holy Grail of all zoom lens designers. There will always be a compromise in quality, as anyone who's bought a 10x zoom has found out. (I believe the Canon 35-350 zoom, despite being an L lens, is a bit soft at the 350mm and 35mm ends. But I doubt very much that it would be worse than the Zuiko 50-200mm with 1.4x extender. Surely it's widely known by now that all teleconverters seriously degrade lens performance. I can only assume that Matthew is just trying to be entertaining   .)

It's true the Olympus 14-54mm fills a gap in the Canon range as it applies to the 22x15mm sensor, but not as it applies to full frame 35mm. I'd love to see a comparison between the Zuiko 14-54 and the Canon 28-135 IS zoom attached to a 1Ds ...... at full aperture and widest angle.  

Since this thread is primarily about the Sony F828, we should at least give it the occasional mention. For all those with 'average' requirements who are searching for a light, compact camera with a single lens that covers most situations, the F828 might be just the ticket. It's certainly cheaper, lighter and more compact than the Olympus E-1 yet produces about the same quality of image, perhaps even slightly better with a bit of help from Neat Image   .

Matthew,
   thanks for the correction (one review said otherwise, but probably out of ignorance.)

  One grievance I have had with telephoto supplementary lenses is that they typically allow little or no zooming; they can only be used at or close to maximum focal length: is that true with the 828?

Here is one shootoff

http://www.a-digital-eye.com/Shootoff4.html

They have other ones at shootoff1, 2, 3 etc.

You can also buy a real good prime for not a lot of money, use a fast enough shutter speed to eliminate mirror slap motion.  The extra ISO settings of a dSLR may give you the room to cut shutter speed.  

But until there are some really good comparisons of the *actual* resolution of the F28 vs. 6 meg dSLRs one might want to keep their credit card number off the web.

I haven't seen any real tight studies (maybe they are out there and I haven't seen them) but I have seen a couple of examples that makes me wonder if the noise suppression routines of the F828 might be wiping out some detail in the final product.

But...  It may yet be a good budget landscape digital.  Just like the role that the 14n plays for full frame.

Quite so! But the larger pixel usually has the advantage of superior DR and less noise. The F828 appears to be the first digicam that has significantly boosted pixel count without significantly degrading the image in terms of DR and noise.

The comparison between the F828 at 200mm, f5.6 and ISO 64 and the 10D at 100 ISO, same everything else, shows the difference.

The Sony F828 is inevitably suffering from 'small pixel syndrome'.

But not quite as small as Michael has stated. A micron is not a millionth of an inch, by a long shot, but a thousandth of a millimetre, or a 25 thousandth of an inch.

Sorry to be a nitpicking pain in the arse.  :D

Matthew,
You're beginning to sound like a diehard film buff who thinks digital images look artificial, or an analogue hi fi buff who thinks the crackles and pops add ambience. Are you sure you're not losing objectivity  ???  Many owners of even the D30 would not consider 12x18" an over enlargement. However, I agree that 24x36" for a D60 image is stretching things a bit. This size really shows up the deficiencies of the lens. But one should bear in mind that such prints are generally not viewed from a distance of 15" as an A4 size print is.

The point that the F828 noise might not be noticeable in a Super B3 size print has a certain validity - but which print are we talking about? All A3/B3 prints?

Noise is related to dynamic range. I often expose for the highlights with my D60, which sometimes results in large areas of the image being dark and underexposed. Fortunately, because the D60 has low noise at ISO 100, I can bring out the detail in the shadows in post processing without too much noise becoming evident.

I wonder if I could do the same with the F828.  ???

In a actual print, there are 2 factors in play: the number of pixels, and the quality of those pixels. Obviously, the Sony is noisier on a per-pixel level than the DSLR. But the fact that it has 33% more pixels than the DSLR helps counteract the higher per-pixel noise level. Add in the apparently weaker anti-aliasing filter of the Sony compared to the Rebel (on a per-pixel level) and it seems to get even more interesting. If I was in the market for a second camera, I'm not sure which one I would get, frankly. I already bought into the Canon DSLR system (I have a 1Ds), so I have lenses and flashes that would fit the Rebel. But if I really needed something compact, light, and unobtrusive, I would seriously consider the Sony.

Ray,

The 300/2.8 is overpriced no doubt -- but I suspect you will be able to get it cheaper in a couple years.

The other lenses are not overpriced, and quality is very high.  With the 14-54 and 50-200 and a TC you have a small, portable kit capable of shooting from 28-580 with very high quality lenses.

I think a lot of people want a smaller dSLR kit with high quality.  When I go for a dSLR I'm certainly strongly considering the 4/3 system.

If a couple other vendors (think Sony) decide to pick up 4/3, I think it could slowly dominate over time as the lenses can be smaller, cheaper, faster, and just as sharp as those designed for a 35mm sensor.

Well I'm very interested in facts but so much bothered about majority opinions. The fact is I have a budget large enough to buy any midrange DSLR, the basic accessories, and a couple of lenses. I want it to be lighter and close to as versatile as my Canon T90 FD system. Believe me I've looked at everything. The Canons don't quite click with me. I'm very appreciative of all Canon have done to raise the standards, but the midrange DSLR bodies all have various flaws (in my opinion, so I'll keep it to myself) and the 1Ds, even if I could afford it, is too big and heavy for my kind of use. Actually no Canon EOS body has ever really screamed "buy me". Closest is the EOS 3, but even then, all those focus points ?  The lenses are another matter, there are some jewels, and the T/S range are almost enough to win the case on their own. But Canon EOS lenses are designed for 35mm, up to now, so fitting them to a non-full frame body is inevitably going to lead to some compromises. A Canon 10D or 300D with a "35mm" lens reminds me of a duck-billed platypus... not quite one thing nor the other.
Obviously the same goes for Nikon, although they've at least said they're committed to "APS" format and have started to flesh out a range of lenses. But if Olympus (say) were as bad at meeting delivery targets as Nikon are with their APS lenses, can you imagine the howls of derision ? Actually I love the Fuji S2 - in theory - but I'd rather see Fuji develop their own body and adopt their awesome Fujinon lens technology to digital. I simply cannot imagine why they're happy to be a Nikon second string.

I think you can do some fantastic work with the E-1, and I also see a lot of sense in going for a system where the lens range is perfectly tailored to the camera. A completely new system is also quite exciting, and I like the 4/3 format.  I could choose a 10D and then dither for ages over which lenses to choose - and I know that whatever I choose will be non-optimal.  I also know personally that the spash proofing of the E-1 would get a serious test in my hands. I doubt the 10D (or 300D) would survive very long...

It is certainly true, as Ray points out, that competition will not stop, but this helps everybody. Ok, so Canon will bring out a mega-machine in one year - but so what - everybody who has bought a 10D or whatever isn't going to get a discount on it. It doesn't really matter. There's also no logical reason at all in getting hung up on "full frame" sensors - except if you're tied in to a 35mm legacy system. If pixel size really is the dominating factor (and I'm not saying it isn't) then "full frame 35mm" has limitations too. Basically if your ambitions are limited to printing up to Super A3, as mine are, for now, then whilst 5Mp is a bit weak, 10Mp (say) would be overkill...depending on the frame geometry of course.

Just my thoughts, really. I'm not trying to convince, or hoping to be convinced.  I really don't care what other people are using, beyond a general interest. I've never really understood all this Canon / Nikon stuff either. I've been using Canon gear for about 20 years but I'm not emotionally attached to it...