Ad
Ad
Ad
Pages: « 1 2 3 [4] 5 »   Bottom of Page
Print
Author Topic: Is It Just Me or ...  (Read 16466 times)
Ray
Sr. Member
****
Offline Offline

Posts: 8812


« Reply #60 on: November 04, 2008, 09:09:09 PM »
ReplyReply

Quote from: DarkPenguin
I don't think the glass mattered that much.  From what I've been able to see the problem is in the analysis.  The Jay Turberville stuff (in the dp forums) seems to indicate that the samples provided by dpreview show nearly the resolution increase one would expect from the increased pixel count.  (Like around a 20% actual increase in resolution.  22% projected.)  EDIT: This isn't what was reported, however.  So the question becomes how did they judge resolution?  By eyeballing it?

There seems to me to be great confusion as to the significance of the claimed higher resolution of expensive, high quality glass. The best and sharpest lens that Canon ever made, the 200/1.8, that's now discontinued and is very heavy and used to cost several thousands of dollar, is only marginally sharper than the humble 50/1.4 standard lens that costs just a few hundred dollars.

The reason such lenses (as the 200/1.Cool are so expensive is that they are sharp at wide apertures. That's where the expense comes in. It's very difficult to make a lens that is as sharp at F2.8 as it is at F8. The 200/1.8 is sharpest at F4, but only marginally sharper at F4 than it is at F8, and only marginally sharper at F8 than the 50/1.4 at F8.

If we look at the potential increased resolving power of the 50D sensor (compared with the 40D) we find that it should be about 22%. (Twice the pixel count equates to a 1.4x increase in resolving power, and a 1.5x increase in pixel count equates to a 1.22x increase in resolving power, in terms of lp/mm.)

However, (and it's a big however), this potential 22% increase in resolution is only achievable if the lens used also has a 22% increase in resolving power. Resolution in the image is always a product of lens resolution and sensor resolution.

We don't seem to have much MTF information on lenses nowadays, but there is no lens that was ever tested by Photodo that had even nearly 22% more resolving power than a basic and inexpensive prime such as the Canon 50/1.4, comparing lenses at their sharpest aperture, whatever that may be. We might be looking at a 5% increase at best and typically a 2 or 3% increase.

The real advantage of a very sharp, expensive prime is that it allows you to use a faster shutter speed and get a shallow DoF without sacrificing the sharpness that can always be obtained at F5.6 and F8. The 200/1.8 is significantly sharper at F1.8 than the 50/1.4 at F1.8. That's what you pay for. Faster shutter speeds usually result in sharper images, and a shallow DoF tends to enhance the appearance of sharpness at the plane of focus, by contrast with what's out of focus.

There would be little point in buying expensive glass to use with the 50D unless you need or want the faster shutter speeds and shallow DoF that such lenses offer. Such lenses would be useful with any camera.
« Last Edit: November 04, 2008, 09:12:40 PM by Ray » Logged
JohnKoerner
Guest
« Reply #61 on: November 04, 2008, 11:08:39 PM »
ReplyReply

Quote from: Ray
There would be little point in buying expensive glass to use with the 50D unless you need or want the faster shutter speeds and shallow DoF that such lenses offer. Such lenses would be useful with any camera.

Huh??? This is exactly bass-ackwards of every other recommendation.

So what you're essentially saying is ............. when you buy the very latest Canon cameras, don't bother getting "the best lenses" ... just buy low-end?

You're saying just go ahead and buy the EF-S 18-200mm f/3.5-5.6, and there will be "no difference" between this and the EF 70-200mm f/2.8L IS USM?

Surely you jest? Or surely you just had a long day  

Jack
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8812


« Reply #62 on: November 05, 2008, 08:52:26 AM »
ReplyReply

Quote from: JohnKoerner
Huh??? This is exactly bass-ackwards of every other recommendation.

So what you're essentially saying is ............. when you buy the very latest Canon cameras, don't bother getting "the best lenses" ... just buy low-end?

I never mentioned low-end. Some low-end zooms are sharpest only at F16. What I'm saying is that many reasonably priced lenses, such as the Canon 50/1.8 and 50/1.4, at their sharpest apertures of F5.6 and F8 are only marginally less sharp than the most expensive lenses at the same apertures of F5.6 and F8. The expensive lenses tend to be significantly sharper (than the 50/1.4 etc) only at wide apertures, therefore their primary benefit with a 50D would be in circumstances where you want a very shallow DoF without compromising resolution, just as with any other camera.

The 70-200/2.8 at its sharpest aperture and at it's sharpest focal length will not be as sharp as the 50/1.4 at it's sharpest aperture, and it's doubtful that even  the very expensive 85/1.2 at its sharpest apertures of F5.6 and F8 will be more than very slightly sharper than the 50/1.4 at F5.6 and F8.

If you want to get the best out of a camera to test its sharpness, any good prime will do if it is used at its sharpest aperture, which is usually between F5.6 and F8.
Logged
DarkPenguin
Guest
« Reply #63 on: November 05, 2008, 10:02:19 AM »
ReplyReply

Thanks, Ray.
Logged
Slough
Guest
« Reply #64 on: November 05, 2008, 11:13:30 AM »
ReplyReply

Quote from: Ray
The reason such lenses (as the 200/1.Cool are so expensive is that they are sharp at wide apertures. That's where the expense comes in. It's very difficult to make a lens that is as sharp at F2.8 as it is at F8. The 200/1.8 is sharpest at F4, but only marginally sharper at F4 than it is at F8, and only marginally sharper at F8 than the 50/1.4 at F8.

Strictly speaking they are more expensive because the fast aperture requires larger glass elements, meaning more work to fabricate. In some cases it also means exotic elements - aspherics and low dispersion glasses (where glass is used in a loose manner) - in order to achieve the desired image quality, which further increases the cost. Many of the Nikon non pro grade lenses (e.g. non F2.8 zooms) are pretty sharp wide open, or stopped down a little.
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8812


« Reply #65 on: November 05, 2008, 12:27:46 PM »
ReplyReply

Quote from: DarkPenguin
Thanks, Ray.

DarkPenguin,
I don't recall ever being thanked before on this website. A new experience. Thanks for that    .

Logged
Plekto
Sr. Member
****
Offline Offline

Posts: 551


« Reply #66 on: November 05, 2008, 01:22:12 PM »
ReplyReply

Quote from: Slough
Strictly speaking they are more expensive because the fast aperture requires larger glass elements, meaning more work to fabricate. In some cases it also means exotic elements - aspherics and low dispersion glasses (where glass is used in a loose manner) - in order to achieve the desired image quality, which further increases the cost. Many of the Nikon non pro grade lenses (e.g. non F2.8 zooms) are pretty sharp wide open, or stopped down a little.

True, but IME, the best lenses still are pretty much useless at the last stop or two.  IME, this seems to hold true for most all lenses, which is why their specs should be taken with a large chunk of salt.    Still, a F1.2 lens that's actually usable at say, F2.0 is a big gain over a typical F2.8 that maybe is okay at F4.0 or so.

Too bad most review sites don't compare lenses at their optimum settings but instead just drop any old lens on the camera and start taking pictures.
Logged
DonWeston
Full Member
***
Offline Offline

Posts: 148


« Reply #67 on: November 05, 2008, 02:34:52 PM »
ReplyReply

Just some general comments on reviews of all sorts, camera gear or other, could easily apply to cars, audio gear as well.  The best reviews are not gospel,  but are valuable to give some general insight in to what the model can do. There will be sample variation between bodies, lenses, etc. It may not be with a lens you particularly have or use, this is not a fault. Most photographers who have been in photography for  more then 5 min. hobby or professional, have used or own a 50mm prime lens. It is a known quanitity, and as such serves as a guide, again not gospel, to the results of the review or test. It helps to know how research and testing is done. One can pick apart any review, both in gear or methodology. This too is irrelevant, as an intelligent person, one should look only for some insight into what the gear is capable of. In the end, only you can be the final judge as to whether this particular gear would suit your needs. I find I can only get the general idea from even the best reviews, whether they be done by Michael, Thom Hogan, or Sean Reid or Bjorn. Too much sample variation exists, let alone use variables to use any of these for more. No one should condemn a piece of gear for a review or the review itself. Even with gear that is generally accepted as excellent and well built, this is NO guarantee that if you buy the same gear, it will perform as reported. Caveat emptor.......

This seemed never to be such an issue with my old Hassy gear and 4x5 kit, but as soon as electronics have reared its massive head, and the gear is sold to many more users, many non professionals, equipment failure became a fact of life and "problems" have arisen. Both poorer quality in manufacture and less knowledgeable end users contribute to this ongoing situation....
Logged
JohnKoerner
Guest
« Reply #68 on: November 07, 2008, 05:28:31 PM »
ReplyReply

Quote from: Ray
There seems to me to be great confusion as to the significance of the claimed higher resolution of expensive, high quality glass. The best and sharpest lens that Canon ever made, the 200/1.8, that's now discontinued and is very heavy and used to cost several thousands of dollar, is only marginally sharper than the humble 50/1.4 standard lens that costs just a few hundred dollars.
The reason such lenses (as the 200/1.Cool are so expensive is that they are sharp at wide apertures. That's where the expense comes in. It's very difficult to make a lens that is as sharp at F2.8 as it is at F8. The 200/1.8 is sharpest at F4, but only marginally sharper at F4 than it is at F8, and only marginally sharper at F8 than the 50/1.4 at F8.
If we look at the potential increased resolving power of the 50D sensor (compared with the 40D) we find that it should be about 22%. (Twice the pixel count equates to a 1.4x increase in resolving power, and a 1.5x increase in pixel count equates to a 1.22x increase in resolving power, in terms of lp/mm.)
However, (and it's a big however), this potential 22% increase in resolution is only achievable if the lens used also has a 22% increase in resolving power. Resolution in the image is always a product of lens resolution and sensor resolution.
We don't seem to have much MTF information on lenses nowadays, but there is no lens that was ever tested by Photodo that had even nearly 22% more resolving power than a basic and inexpensive prime such as the Canon 50/1.4, comparing lenses at their sharpest aperture, whatever that may be. We might be looking at a 5% increase at best and typically a 2 or 3% increase.
The real advantage of a very sharp, expensive prime is that it allows you to use a faster shutter speed and get a shallow DoF without sacrificing the sharpness that can always be obtained at F5.6 and F8. The 200/1.8 is significantly sharper at F1.8 than the 50/1.4 at F1.8. That's what you pay for. Faster shutter speeds usually result in sharper images, and a shallow DoF tends to enhance the appearance of sharpness at the plane of focus, by contrast with what's out of focus.
There would be little point in buying expensive glass to use with the 50D unless you need or want the faster shutter speeds and shallow DoF that such lenses offer. Such lenses would be useful with any camera.


Thanks Ray.

Question though: are "sharpness" and "resolution" the same thing? I honestly don't know.

The 50mm prime you speak of is considered "sharp" on the 40D but not so on the 50D ... and my (admittedly limited) understanding is because of resolution. Not sure what that means exactly, but that is the complaint I keep reading. Lenses like to 100 macro are "sharp" too, but provide better resolution, which the 50D maximizes. At least that is my layman (not really sure) interpretation of what I am reading.

On the flipside, though, the positive news I really do understand is that the 50D has already gone down in price by $200 within the last week, thanks to the bad review, so maybe I should thank DP Review rather than be angry  

Jack
« Last Edit: November 07, 2008, 05:29:14 PM by JohnKoerner » Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8812


« Reply #69 on: November 07, 2008, 06:54:24 PM »
ReplyReply

Quote from: JohnKoerner
Question though: are "sharpness" and "resolution" the same thing? I honestly don't know.



John,
Not quite, but they are closely related. There are entire books devoted to sharpness and sharpening routines, but I won't pretend I've read them. I use the sharpening controls that are included in ACR; Smart Sharpen in Photoshop and/or a plug-in called Focus Magic. I like to keep it simple.

My basic understanding is that all lenses will cause a loss of contrast of the detail in a scene that is transmitted through the glass. The sharper the lens, the lower that loss of contrast. The sensor or film inevitably adds to that loss of contrast, so in a sense you have a double whammy.

A loss of contrast causes an appearance of softness in an image. Sharpening routines are an attempt to recover that loss of contrast. However, if the detail is particularly faint or fine in the real world scene being photographed, the combined loss of contrast from both the lens and the recording processes of the sensor, taking into account also the resolving limitation of the sensor in terms of the fineness of detail it can capture period, will result in such faint detail not being captured at all.

In such circumstances, no amount of sharpening can recover detail that simply hasn't been recorded.

My impression is, at any given aperture that is used with both cameras (including F16, but forget F32) the 50D will record slightly more detail than the 40D, provided there is such detail in the scene to be recorded. However, it's also true that the differences between the 40D and 50D will be more noticeable when the lens is used at its sharpest (least-loss-of-contrast) aperture. However, as I've previously mentioned, that least-loss-of-contrast aperture in 35mm lenses does not get much better in expensive lenses than it is already at F5.6-F8 in all 'good' lenses. The benefit of the expensive, top-notch lens tends to be, you get F5.6-F8 sharpness at F1.8 to F4, and sometimes a bit more, but not much more.
Logged
BruceHouston
Sr. Member
****
Offline Offline

Posts: 308



« Reply #70 on: November 07, 2008, 10:35:40 PM »
ReplyReply

Quote from: Ray
John,
Not quite, but they are closely related. There are entire books devoted to sharpness and sharpening routines, but I won't pretend I've read them. I use the sharpening controls that are included in ACR; Smart Sharpen in Photoshop and/or a plug-in called Focus Magic. I like to keep it simple.

My basic understanding is that all lenses will cause a loss of contrast of the detail in a scene that is transmitted through the glass. The sharper the lens, the lower that loss of contrast. The sensor or film inevitably adds to that loss of contrast, so in a sense you have a double whammy.

A loss of contrast causes an appearance of softness in an image. Sharpening routines are an attempt to recover that loss of contrast. However, if the detail is particularly faint or fine in the real world scene being photographed, the combined loss of contrast from both the lens and the recording processes of the sensor, taking into account also the resolving limitation of the sensor in terms of the fineness of detail it can capture period, will result in such faint detail not being captured at all.

In such circumstances, no amount of sharpening can recover detail that simply hasn't been recorded.

My impression is, at any given aperture that is used with both cameras (including F16, but forget F32) the 50D will record slightly more detail than the 40D, provided there is such detail in the scene to be recorded. However, it's also true that the differences between the 40D and 50D will be more noticeable when the lens is used at its sharpest (least-loss-of-contrast) aperture. However, as I've previously mentioned, that least-loss-of-contrast aperture in 35mm lenses does not get much better in expensive lenses than it is already at F5.6-F8 in all 'good' lenses. The benefit of the expensive, top-notch lens tends to be, you get F5.6-F8 sharpness at F1.8 to F4, and sometimes a bit more, but not much more.

Over the past year during which I have eased back into amateur photography from a 30-year plus hiatus, I have noticed that the resolution of my Canon lenses keeps changing for the better.  These magic lenses get ever sharper as my focusing and aperature selection techniques improve.  
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 6910


WWW
« Reply #71 on: November 08, 2008, 01:04:03 AM »
ReplyReply

Hi,

I would say sharpness is a visual thing. The eye/brain complex perceives something as sharp or not. Sharpness has probably more to do with acutance than with resolution. In a sense resolution and sharpness go hand in hand. I have seen examples of possible scenarios of lenses having low resolution but high sharpness, but they have all been contrieved.

Now, we need to keep in mind that the eye has a limited resolution itself, it is dependent mainly on the distance between the cones in the foeva and is normally considered to be about one minute of arc (1/60 degree). If we look at a picture at a certain distance the eye can resolve things which are certain distance apart, let's say 1/10 of a millimeter at 25 cm. Any resolution beyond that is wasted, the visual impression is dependent on how much of the contrast is transferred at that resolution. If you make a larger enlargement or change viewing distance the playing field is redefined.

The perceived contrast can be enhanced using different methods of sharpening. Sharpening does not necessarily increase resolution but increases the contrast on the edges. Because of all the factors involved like enlargement size, viewing distance and equipment, the needed amount of sharpening is subjective. There are a lot of other factors affecting image quality, flare and ghosting come to my mind.

Note that resolution can actually be improved in a sense using deconvolution. Deconvolution is only possible if the "point spread function" is known, that is if we know how a single point of light is converted to an image. This is normally not known, but if we assume that the primary cause of unsharpness is lack of focus we can predict a pretty good "PSF" for the circle of confusion. The effect of the antialiasing filter may also been taken into account.

Tools that use deconvolution known to me are:

FocusMagic (doesn't work with CS3 on Intel Macs unless under Rosetta)
FocusFixer
Smart Sharpening in Photoshop CS
RawDeveloper has Lucy-Richardson deconvolution

One observation I have made is that an image can look sharp event if it lacks detail, especially at a distance. I have made A2 enlargements from 6MP images which look good at some distance, but if I'm pixel peeping I can see that edge are poorly resolved. The essence is that because the edges have high contrast the brain thinks that "the details must be there". Similarly I have made an 70x100 cm enlargement from a 10 MPixel image, with "mucho sharpening". It's very nice if viewed from longer than about 80 cm distance.

You may check:
http://www.bobatkins.com/photography/technical/mtf/mtf1.html

A much deper presentation is here:
http://www.normankoren.com/Tutorials/MTF.html



Quote from: Ray
John,
Not quite, but they are closely related. There are entire books devoted to sharpness and sharpening routines, but I won't pretend I've read them. I use the sharpening controls that are included in ACR; Smart Sharpen in Photoshop and/or a plug-in called Focus Magic. I like to keep it simple.

My basic understanding is that all lenses will cause a loss of contrast of the detail in a scene that is transmitted through the glass. The sharper the lens, the lower that loss of contrast. The sensor or film inevitably adds to that loss of contrast, so in a sense you have a double whammy.

A loss of contrast causes an appearance of softness in an image. Sharpening routines are an attempt to recover that loss of contrast. However, if the detail is particularly faint or fine in the real world scene being photographed, the combined loss of contrast from both the lens and the recording processes of the sensor, taking into account also the resolving limitation of the sensor in terms of the fineness of detail it can capture period, will result in such faint detail not being captured at all.

In such circumstances, no amount of sharpening can recover detail that simply hasn't been recorded.

My impression is, at any given aperture that is used with both cameras (including F16, but forget F32) the 50D will record slightly more detail than the 40D, provided there is such detail in the scene to be recorded. However, it's also true that the differences between the 40D and 50D will be more noticeable when the lens is used at its sharpest (least-loss-of-contrast) aperture. However, as I've previously mentioned, that least-loss-of-contrast aperture in 35mm lenses does not get much better in expensive lenses than it is already at F5.6-F8 in all 'good' lenses. The benefit of the expensive, top-notch lens tends to be, you get F5.6-F8 sharpness at F1.8 to F4, and sometimes a bit more, but not much more.
« Last Edit: November 08, 2008, 01:05:50 AM by ErikKaffehr » Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 6910


WWW
« Reply #72 on: November 08, 2008, 01:35:22 AM »
ReplyReply

Hi John,

I think that DPReview is slightly opportunistic. They probably decided that Canon 50D is going over the limit and are a bit positive regarding data showing into that direction. Keep in mind that a 50% increase in Megapixels is just 22% increase in resolution (because resolution is per inch and Mpixels per square inch). Most lenses were designed for film, very few lenses are as sharp as they could be, which is called "diffraction limited". Almost any lens will be diffraction limited if you stop down enough. The loss of sharpness is gradual.

Increasing "megapixels" gives diminishing returns and there is a price to pay for that, because DR (Dynamic Range) and noise get reduced when pixel size is reduced. I think that DPReview may make to much noise about noise, even if I think they may have a point.

If you are shooting low ISO megapixels are probably your friends. You can make larger enlargements with less interpolation and artifacts. The other side of the coin is that Canon had the 40D at 10 MPixels while the competition had 12.5 (Nikon/Sony) or even 15. It may have been the case that the 40D was rushed to the market. Canon may also have a small problem of not upsetting their price/megapixels relation. They cannot have a cheap 450D and an expensive 50D having 5 MPixels less resolution!

I think that DPReview actually has good reviews but they are quite subjective and also very much focused on JPEGs (which myself never shoot)

You may check the pages below which clearly indicate the advantage of 50D over 350D:

http://www.photozone.de/canon-eos/404-cano...6is_50d?start=1

http://www.photozone.de/canon-eos/181-cano...-review?start=1



Best regards
Erik

Quote from: JohnKoerner
Thanks Ray.

Question though: are "sharpness" and "resolution" the same thing? I honestly don't know.

The 50mm prime you speak of is considered "sharp" on the 40D but not so on the 50D ... and my (admittedly limited) understanding is because of resolution. Not sure what that means exactly, but that is the complaint I keep reading. Lenses like to 100 macro are "sharp" too, but provide better resolution, which the 50D maximizes. At least that is my layman (not really sure) interpretation of what I am reading.

On the flipside, though, the positive news I really do understand is that the 50D has already gone down in price by $200 within the last week, thanks to the bad review, so maybe I should thank DP Review rather than be angry  

Jack
Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 6910


WWW
« Reply #73 on: November 08, 2008, 08:08:17 AM »
ReplyReply

Ray,

It's more than just Mpixels. I was shooting with my old friend Pierre on Iceland, he was toting a Canon 20D and I Konica Minolta D7. In retrospect I can see that both mine and his images are quite soft on a pixel by pixel basis. Shortly after the Iceland trip I upgraded to the Sony Alpha 100 at 10 MPixel. The pixel by pixel image quality was much better on the Alpha 100, at least at low ISO. That said, it was not easy to tell apart A3+ (49 x 33 cm) prints I made from both. On the other hand it was actually easy to tell them apart once I have learned where to look. Now I have a Sony Alpha 700 and I feel quite comfortable doing A2 prints (42 by 59 cm). I can make A2s from the Iceland trip but they lack good detail.

I'd think that the 40D is similarly much better than the 20D, I would probably guess that this may be due to a weaker antialiasing filter.

William Castleman seems to have done a lot of good tests and he has some nice data on the 20D vs. 40D on this page:

http://www.wlcastleman.com/equip/reviews/40D/index.htm

Erik


Quote from: Ray
Don't know what all the fuss is about. The incremental pixel count increase in the 50D is in line with previous upgrades in the APS-C cropped format that Canon have given us. The only exception was the very first upgrade from the 3mp D30 to the 6mp D60, which was a 100% increase. The 10D to the 20D was a 33% increase; the 20D to the 40D a 25% increase; the 40D to the 450D a 20% increase and the 450D to the 50D a 25% increase.

I don't recall getting excited about any resolution increase when I upgraded from the D60 to the 20D. It seemed fairly irrelevant to me. The exciting feature of the 20D was its remarkably low noise at high ISO. ISO 1600 on the 20D was at least as good as ISO 400 on the D60.

I've always been of the view that anything less than a 50% increase in pixel count was irrelevant. The doubling in pixel count from the 3mp of the D30 to the 6mp of the D60 caused Michael to comment in his review that the D60 would be good for A3+ size prints whereas the D30 only good for A3, by the same critical viewing standards.

The 50D represents a 50% increase in pixel count over the 40D, so one might conclude that such an increase gets into the category of being significant. However, when the increase is from such a high base, it becomes less relevant. An increase from 3mp to 4.5mp, or an increase from 6mp to 9mp would be worthwhile... just. But an increase from 10mp to 15mp is hardly worthwhile. The best I can say is that it's better than no increase at all, provided there are no downsides such as increased noise.

For me, the 50D should have been my upgrade from the 20D. But someone in Bangkok made me an irresistable offer when I was out shopping for the EF-S 17-55/2.8, and my resolve weakened and i bought the 40D. I knew already I could expect no increase in resolution or reduction in noise that would be significant. The fast frame rates and the LiveView feature were the main drawcard for me, just as with the 50D the main drawcard was the 'autofocus micro adjustment' and the high resolution LiveView screen which facilitates amazingly accurate manual focussing. The additional pixel count I consider as a small bonus.
« Last Edit: November 08, 2008, 08:09:52 AM by ErikKaffehr » Logged

JohnKoerner
Guest
« Reply #74 on: November 08, 2008, 10:09:20 PM »
ReplyReply

Thanks Ray and Erik.

Ray, I gather your impression (as an owner of both) is that the 50D is ultimately better than the 40D.

I have seen favorable comparisons of the 50D on the 100-400, and the 100 macro, repeatedly obtain.

Have you personally tried a comparison of the two on a MP-E 65mm macro?

If so, what were your observations?

Thank you again.
Logged
John Camp
Sr. Member
****
Offline Offline

Posts: 1252


« Reply #75 on: November 08, 2008, 11:31:37 PM »
ReplyReply

One of the key things about camera comparisons is that not only are they NOT standardized, there's no way to standardize them. The best comparisons are done by experienced photographers who look at both cameras and then come back with a menu of differences, that the potential buyer chooses from. You may get better resolution from a 1DsIII than from a D3 (actually, you will) but there may be complicated arguments about color, about ISOs, about ergonomics, and about end-use, that make more difference to a specific photographer than ultimate resolution.

Sharpness and resolution are different, though tightly entangled, especially in landscape work. If you take a photograph of a mass of millions of leaves in a tree, the resolution is whatever it is; but post-processing sharpening will find the edges of those leaves, and emphasize them essentially by darkening that line. With detail that's large enough to handle the sharpening, but small enough to really need it, good sharpening will fool the eye into believing there is more resolution than in a non-sharpened photo...but there isn't. It's just that the edges have been artificially sharpened. Think of one of those old paintings by somebody like Fragonard, where he actually drew thousands of leaves -- you don't actually see a tree that way, from any distance at all (you see a dark mass of leaves), so what Fragonard's drawing each of each leaf does, actually, is artificial sharpening... 8-)  

You also have to consider that DPR does its tests with .JPGs, and quite a large number of photographers have said that Nikon has made a deliberate series of design decisions that require its photographs to be sharpened to taste in post-processing, to get the best out of them. Whether or not that's BS, I couldn't say, but might explain why a tester might think that it's good to sharpen a D300 but not another camera, where the JPG is made through a different process. In other words, he may be trying to balance the different problems, by comparing best-to-best, rather than whatever comes out of the camera, to whatever comes out of the other camera, figuring that most people who buy that level of camera will be sophisticated enough to understand the problem...

JC
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 6910


WWW
« Reply #76 on: November 09, 2008, 01:55:52 AM »
ReplyReply

Hi,

Anything but very careful sharpening would introduce artifacts in the image. For that reason some or most photographic agencies are quite sensitive about sharpening. Cameras intended for professional use require more sharpening (on JPEGs) than amateur cameras probably for that reason. If you shoot raw you are always on your own.

I would suggest that many aspects are measurable but many are not. I also would suggest that cameras may work better or worse in certain situations. Look at Canon, they have the 1DIII and the 1DsIII. 1DIII has few but big pixels and high ISO, high speed and crop factor of 1.3 and is intended for photojournalists the 1DsIII is a studio/landcape camera. Smaller pixels, full frame, lower ISO.

Erik


Quote from: John Camp
One of the key things about camera comparisons is that not only are they NOT standardized, there's no way to standardize them. The best comparisons are done by experienced photographers who look at both cameras and then come back with a menu of differences, that the potential buyer chooses from. You may get better resolution from a 1DsIII than from a D3 (actually, you will) but there may be complicated arguments about color, about ISOs, about ergonomics, and about end-use, that make more difference to a specific photographer than ultimate resolution.

Sharpness and resolution are different, though tightly entangled, especially in landscape work. If you take a photograph of a mass of millions of leaves in a tree, the resolution is whatever it is; but post-processing sharpening will find the edges of those leaves, and emphasize them essentially by darkening that line. With detail that's large enough to handle the sharpening, but small enough to really need it, good sharpening will fool the eye into believing there is more resolution than in a non-sharpened photo...but there isn't. It's just that the edges have been artificially sharpened. Think of one of those old paintings by somebody like Fragonard, where he actually drew thousands of leaves -- you don't actually see a tree that way, from any distance at all (you see a dark mass of leaves), so what Fragonard's drawing each of each leaf does, actually, is artificial sharpening... 8-)  

You also have to consider that DPR does its tests with .JPGs, and quite a large number of photographers have said that Nikon has made a deliberate series of design decisions that require its photographs to be sharpened to taste in post-processing, to get the best out of them. Whether or not that's BS, I couldn't say, but might explain why a tester might think that it's good to sharpen a D300 but not another camera, where the JPG is made through a different process. In other words, he may be trying to balance the different problems, by comparing best-to-best, rather than whatever comes out of the camera, to whatever comes out of the other camera, figuring that most people who buy that level of camera will be sophisticated enough to understand the problem...

JC
Logged

Slough
Guest
« Reply #77 on: November 09, 2008, 02:35:49 AM »
ReplyReply

Quote from: JohnKoerner
Question though: are "sharpness" and "resolution" the same thing? I honestly don't know.

To add to the replies, they are related. I agree with the person who suggested that sharpness is what we perceive, and of course resolution is the measured value. A lens with high contrast may produce prints that appear sharper than one with low contrast though the resolution may be similar. Higher contrast produces a snap to the image which we (or at least I) perceive as sharper. And of course sharpening in software increases contrast at edges thus increasing perceived sharpness without actually adding any more resolution.
Logged
Dan Wells
Sr. Member
****
Offline Offline

Posts: 329


WWW
« Reply #78 on: November 09, 2008, 10:27:22 PM »
ReplyReply

Another question that comes up here is "How large is your printer?" The differences above 12 MP really don't show up on less than a 17 inch printer, and can't possibly show up on any monitor or projector while looking at the whole image. 17 inch printers are big, heavy beasts that are a major commitment to live with - they sit on a table, but barely. 24 inch printers are worse (on the order of figuring out where to put a small upright piano!). If you have a printer that size and print to its capability, the highest resolution cameras (including Hasselblads, Phase backs, etc...) might make sense. If it's all going to be printed 8x10 or shown on the web anyway, any camera works (with the exception of dynamic range where the 8x10 print (on a really good printer) is concerned - it IS possible to play "spot the Hasselblad" on an 8x10 print of certain images, but it'll be the 12 stops of dynamic range that gives it away, not the resolution). If you're doing anything exclusively for electronic display, whether monitors, projectors or whatever, they're all not only under four megapixels (and 4 MP is a 30 inch monitor, a more than slightly exotic beast), but also have only about five or six stops of dynamic range - a really good compact is plenty for even a 30 inch monitor.


                      -Dan
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8812


« Reply #79 on: November 10, 2008, 05:46:49 AM »
ReplyReply

Quote from: JohnKoerner
Thanks Ray and Erik.

Ray, I gather your impression (as an owner of both) is that the 50D is ultimately better than the 40D.

I have seen favorable comparisons of the 50D on the 100-400, and the 100 macro, repeatedly obtain.

Have you personally tried a comparison of the two on a MP-E 65mm macro?

If so, what were your observations?

Thank you again.

John,
I would not have bought the 50D purely for the increased number of pixels. The micro adjustment of autofocussing was paramount in my mind, and the high resolution LCD screen is just delicious. 10x magnification in LiveView mode with a 400mm lens is awesome, when you half press the shutter button to activate IS.

I don't have the MP-E 65mm macro, but shortly after receiving delivery of my 50D and checking it out, I bought the EF-S 60mm macro from the same internet company because their prices were so good and they were so responsive.

I've checked out the lens to make sure it was not a dud, but haven't done any critical comparisons comparing it with my 50/1.4. Macro lenses sometimes  shine only at very close distances to subject. But some are excellent to infinity.

Such a comparison will have to wait till I've completed installation of my new Panasonic 50" Plasma which, interestingly, has an SD card slot from which it will display jpeg imagesat 1920x1080p. I don't have any SD cards, but already I'm planning getting one or two big ones and transferring some processed images in jpeg format, and having a slide show of 2mp images.  
« Last Edit: November 10, 2008, 06:23:06 AM by Ray » Logged
Pages: « 1 2 3 [4] 5 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad