Ad
Ad
Ad
Pages: « 1 ... 5 6 [7] 8 »   Bottom of Page
Print
Author Topic: Mini Medium Format Shootout  (Read 28766 times)
Vivec
Newbie
*
Offline Offline

Posts: 23


WWW
« Reply #120 on: November 22, 2010, 04:57:42 PM »
ReplyReply

Glad you're back Mark. I was biting my tongue the entire week that you were away and while the kids had
Michael

First of all, i think it was a very nice and informative review -- thanks Nick, Mark, and Michael.

However, irrespective of whether joseph is a 'newbie' or not, he did seem to have a valid point that there is something wrong with the 100% crops. And after reading the responses it seems to me that the authors have not addressed this - right?  it is fine if a mistake was made, but it would be good to confirm this. Anyway, I am quite interested in the results -- not everyone has the opportunity to work with such amazing cameras :-)

Thanks again for taking time to do such review.
 
« Last Edit: November 22, 2010, 10:13:22 PM by Vivec » Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8877


« Reply #121 on: November 22, 2010, 06:17:57 PM »
ReplyReply

But my expectations of you were better than this. You are supposed to be a seasoned professional photographer who can see objectively without personalizing issues using phoney psyc-101 non-insights. Obviously, despite the content and context of this article you remain completely clueless about the mindsets brought to bear on its research and preparation. Comments of this ilk don't belong on this website because they contribute NOTHING to learning and understanding, and this site and forum are meant to be about that.

I see that I have upset you Mark. Sorry! But I did use smileys.

On a serious note, I do genuinely get the impression there's an element of self-deception amongst many photographers extolling the benefits of MFDB systems.

Take the issue of the AA filter for example. I remember well, a few years ago on this forum, when the question was first posed, 'Why do DSLRs need an AA filter but DBs don't?'

The answer that made at least some sense, but wasn't entirely satisfactory, was that experienced and professional photographers, who would be the sort of people most likely to use MFDB systems, could be expected to have the knowledge, understanding and competence to deal with moire and aliasing issues whenever they arose, using whatever software that was available to fix the problem.

However, most amateurs buying DSLRs, might think there was something wrong with their camera on first seeing a few examples of obvious cases of moire. They might return their camera as being faulty.

This might be a plausible explanation in relation to the prosumer, cropped-fromat DSLR, but really doesn't explain why the significantly more expensive Canon 1Ds series of FF DSLR, aimed at the professional, also needs an AA filter.

The explanation that the primary reason why DBs (and the Leicas) do not have AA filters is because they would make the already-very-expensive camera or back even more expensive, sounds a much more plausible explanation to me.

Of course, marketing has a role to play here. If the disadvantages of not having an AA filter are perceived by many as outweighing the advantages, then work has to be done extolling the virtues of a lack of AA filter whilst downplaying the obvious disadvantages of moire problems.

This is what appears to have happened. That marginal, extra crispness of a good DB image where moire is not a problem, can also be easily confused with the generally higher resolution that the MFDB produces as a result of its greater pixel count.

When I first checked out the MaxMax site to see what a 5D image looks like when the camera's AA filter is removed, I was very disappointed in the degree of improvement. It was noticeable, but strictly for pixel-peepers in my opinion.
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2773



« Reply #122 on: November 22, 2010, 08:44:37 PM »
ReplyReply

The Leica M9 isn't MF, but it also has no AA filter. I think this is a matter of design philosophy (sizing-up the relative value of the trade-offs) and marketing. There are hardware and software approaches to this problem - when it occurs.

Mark,

Yes, but the M9 is not a dSLR and, because of the short lens to sensor distance, there might not be room for one. Remember, with the M8 there was not even room for an IR filter. One can reduce the color effects in Moire, but the actual aliasing is very difficult to remove. The old Kodak full frame dSLR also didn't have a blur filter and it was notorious for moire. As you say, there are trade offs. One can pay to have the blur filter removed from a dSLR, but few do.

Regards,

Bill
Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6891


WWW
« Reply #123 on: November 23, 2010, 09:03:45 AM »
ReplyReply

I see that I have upset you Mark. Sorry! But I did use smileys.

On a serious note, I do genuinely get the impression there's an element of self-deception amongst many photographers extolling the benefits of MFDB systems.

Take the issue of the AA filter for example. I remember well, a few years ago on this forum, when the question was first posed, 'Why do DSLRs need an AA filter but DBs don't?'

The answer that made at least some sense, but wasn't entirely satisfactory, was that experienced and professional photographers, who would be the sort of people most likely to use MFDB systems, could be expected to have the knowledge, understanding and competence to deal with moire and aliasing issues whenever they arose, using whatever software that was available to fix the problem.

However, most amateurs buying DSLRs, might think there was something wrong with their camera on first seeing a few examples of obvious cases of moire. They might return their camera as being faulty.

This might be a plausible explanation in relation to the prosumer, cropped-fromat DSLR, but really doesn't explain why the significantly more expensive Canon 1Ds series of FF DSLR, aimed at the professional, also needs an AA filter.

The explanation that the primary reason why DBs (and the Leicas) do not have AA filters is because they would make the already-very-expensive camera or back even more expensive, sounds a much more plausible explanation to me.

Of course, marketing has a role to play here. If the disadvantages of not having an AA filter are perceived by many as outweighing the advantages, then work has to be done extolling the virtues of a lack of AA filter whilst downplaying the obvious disadvantages of moire problems.

This is what appears to have happened. That marginal, extra crispness of a good DB image where moire is not a problem, can also be easily confused with the generally higher resolution that the MFDB produces as a result of its greater pixel count.

When I first checked out the MaxMax site to see what a 5D image looks like when the camera's AA filter is removed, I was very disappointed in the degree of improvement. It was noticeable, but strictly for pixel-peepers in my opinion.


I would need to be shown concrete evidence that cost is an issue for an AA filter in a high-end camera system whatever the size of the sensor. The more likely explanation why it is almost so universally used in the sub-MF formats is that the manufacturers simply don't want to deal with torrent of complaints they would get from moire, etc. if there were no AA filter. By the time we get to MF the market is so thin, and the people buying them are so comparatively well enough equipped to handle this issue without complaining, that the manufacturers need not fear massive consumer reaction. I think there is a tendency in discussions to exaggerate the strength of these filters and the impact they have on the images. For all we know, it could be a very mild filtering which has only a modest impact on overall sharpness and easily enough compensated in Capture sharpening. So I'm not surprised by your observation on the 5D.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6891


WWW
« Reply #124 on: November 23, 2010, 09:11:04 AM »
ReplyReply

First of all, i think it was a very nice and informative review -- thanks Nick, Mark, and Michael.

However, irrespective of whether joseph is a 'newbie' or not, he did seem to have a valid point that there is something wrong with the 100% crops. And after reading the responses it seems to me that the authors have not addressed this - right?  it is fine if a mistake was made, but it would be good to confirm this. Anyway, I am quite interested in the results -- not everyone has the opportunity to work with such amazing cameras :-)

Thanks again for taking time to do such review.
 

Thanks, glad you enjoyed the review and got something useful out of it.

If people are making valid, constructive points it is irrelevant to me whether they are new to the forum or not. My concern is with the substance. The judgmental issue concerning the crops and how they display in the article is whether they provide a reasonably correct comparative picture of what we are describing in the text. Only the people who have seen the original images can make this determination with confidence. Sorry, but - inevitably- that's just how it is. We are of the view that we've provided valid comparative visual guidance with these images, and unless it can be proven otherwise, I'll leave the discussion of this tangent at that.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6891


WWW
« Reply #125 on: November 23, 2010, 09:20:44 AM »
ReplyReply

Mark,

Yes, but the M9 is not a dSLR and, because of the short lens to sensor distance, there might not be room for one. Remember, with the M8 there was not even room for an IR filter. One can reduce the color effects in Moire, but the actual aliasing is very difficult to remove. The old Kodak full frame dSLR also didn't have a blur filter and it was notorious for moire. As you say, there are trade offs. One can pay to have the blur filter removed from a dSLR, but few do.

Regards,

Bill

Interesting points Bill. As you may know, Capture One has a Moire filter - the second to last tool of the Details panel. It allows us to adjust for the amount and the pattern, quite effectively.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Eric Myrvaagnes
Sr. Member
****
Offline Offline

Posts: 7897



WWW
« Reply #126 on: November 23, 2010, 10:04:09 AM »
ReplyReply

Mark and Nick,

I just want to say thank you for a very clear, very informative, and thoroughly enjoyable review and "shootout." You both did a fine job, and described very clearly what you were doing. For those of us who can read English, there were no problems at all. Please don't let the Monday morning quarterbacks and Pixel Partiers get on your nerves!

Cheers,

Eric
Logged

-Eric Myrvaagnes

http://myrvaagnes.com  Visit my website. New images each season.
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6891


WWW
« Reply #127 on: November 23, 2010, 01:42:42 PM »
ReplyReply

Mark and Nick,

I just want to say thank you for a very clear, very informative, and thoroughly enjoyable review and "shootout." You both did a fine job, and described very clearly what you were doing. For those of us who can read English, there were no problems at all. Please don't let the Monday morning quarterbacks and Pixel Partiers get on your nerves!

Cheers,

Eric

You are welcome Eric; glad you enjoyed it and found it useful. And indeed, no sleep is being lost at this end of the conversation. We go into these excursions of technical curiosity kind of anticipating the full spectrum of commentary on the way out. :-)
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
BernardLanguillier
Sr. Member
****
Offline Offline

Posts: 7872



WWW
« Reply #128 on: November 23, 2010, 06:10:38 PM »
ReplyReply

I would need to be shown concrete evidence that cost is an issue for an AA filter in a high-end camera system whatever the size of the sensor. The more likely explanation why it is almost so universally used in the sub-MF formats is that the manufacturers simply don't want to deal with torrent of complaints they would get from moire, etc. if there were no AA filter. By the time we get to MF the market is so thin, and the people buying them are so comparatively well enough equipped to handle this issue without complaining, that the manufacturers need not fear massive consumer reaction.

The Pentax representatives I had spoken with in March in Japan told me the same thing. They had first considered adding a AA filter to the sensor because of concerns with digital artifacts (moire being one of them) but testing showed that they could manage to control these through adequate processing and that the overall outcome was in their opinion superior for the landscape applications they are targeting.

Besides, we all know that the specification of a piece of gear (price being one of them) is a key consideration in the purchasing decision of many people, sometimes regardless of the measured performance of the equipment (high end Audio is a wonderful example where high spec and high price do most of the selling). I believe that many MF shooters would not even have considered the 645D had it had an AA filter, again regardless of the actual performance. They had no choice anyway. Smiley

Regards,
Bernard
Logged

A few images online here!
Vivec
Newbie
*
Offline Offline

Posts: 23


WWW
« Reply #129 on: November 23, 2010, 07:31:47 PM »
ReplyReply

The judgmental issue concerning the crops and how they display in the article is whether they provide a reasonably correct comparative picture of what we are describing in the text. Only the people who have seen the original images can make this determination with confidence. Sorry, but - inevitably- that's just how it is. We are of the view that we've provided valid comparative visual guidance with these images, and unless it can be proven otherwise, I'll leave the discussion of this tangent at that.

Thanks Mark for your clarification. This seems fair enough.

I hope you and Nick can get your hands on a S2 too and do a shootout between the three :-)
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8877


« Reply #130 on: November 23, 2010, 09:29:18 PM »
ReplyReply

The Pentax representatives I had spoken with in March in Japan told me the same thing. They had first considered adding a AA filter to the sensor because of concerns with digital artifacts (moire being one of them) but testing showed that they could manage to control these through adequate processing and that the overall outcome was in their opinion superior for the landscape applications they are targeting.

Besides, we all know that the specification of a piece of gear (price being one of them) is a key consideration in the purchasing decision of many people, sometimes regardless of the measured performance of the equipment (high end Audio is a wonderful example where high spec and high price do most of the selling). I believe that many MF shooters would not even have considered the 645D had it had an AA filter, again regardless of the actual performance. They had no choice anyway. Smiley

Regards,
Bernard

This is very strange indeed, Bernard. If the cost of adding an AA filter is not a major issue, and manufacturers of MFDB equipment can adequately control the effects of aliasing in software, one wonders why manufacturers of top-end DSLRs cannot do the same.

Doesn't make sense to me, but of course I'm rational.

I would never buy anything on the basis that the high price alone indicated it was a superior product.
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7323


WWW
« Reply #131 on: November 23, 2010, 10:29:06 PM »
ReplyReply

Hi,

I have looked at a lot of Leica S2 sample pictures on Lloyd Chambers DAP site. In many cases color moiré is quit obvious on those images. Stopping down to f/11 eliminates the problem by and large. On the other hand, the Leica lenses seem to be close to diffraction limited at f/5.6. The moiré effects are ugly at actual pixels but I don't know if they would be visible in prints. The kind of areas they show up are thin details with relatively high contrast, tree branches, straws of grass and so on.

As a comment on what Bernard says, AA-filtered images require significantly more sharpening than non filtered images. The recommendation is like 300-500% at very small radius. That kind of sharpening also affects noise, which we can reduce by applying capture sharpening with some edge masking.

This may be one of the small almost mythical effects that possibly give MF an appearance of better image quality.

  • Resolution (in terms of LP per picture height is higher)
  • More modulation is transferred for a small feature of give size because MTF is higher at lower frequencies
  • The image is smoother because less sharpening is needed

It may be that neither of the above factors is decisive but the combined advantages may turn into a significant improvement of perceived image quality.

Note also that none of the above factors is measured by DxO-mark! This may also explain why MFDBs perform so well at low ISO but far less optimally at high ISO. The sensor may be a bit more noisy than CMOS DSLR sensors on it's own. But, on DSLR we need to apply more excessive sharpening that amplifies noise. At higher ISOs the readout noise in CCDs may get obvious even with the small amount of sharpening needed MFDBs.

Best regards
Erik
The Pentax representatives I had spoken with in March in Japan told me the same thing. They had first considered adding a AA filter to the sensor because of concerns with digital artifacts (moire being one of them) but testing showed that they could manage to control these through adequate processing and that the overall outcome was in their opinion superior for the landscape applications they are targeting.

Besides, we all know that the specification of a piece of gear (price being one of them) is a key consideration in the purchasing decision of many people, sometimes regardless of the measured performance of the equipment (high end Audio is a wonderful example where high spec and high price do most of the selling). I believe that many MF shooters would not even have considered the 645D had it had an AA filter, again regardless of the actual performance. They had no choice anyway. Smiley

Regards,
Bernard

Logged

BernardLanguillier
Sr. Member
****
Offline Offline

Posts: 7872



WWW
« Reply #132 on: November 23, 2010, 10:41:16 PM »
ReplyReply

As a comment on what Bernard says, AA-filtered images require significantly more sharpening than non filtered images. The recommendation is like 300-500% at very small radius. That kind of sharpening also affects noise, which we can reduce by applying capture sharpening with some edge masking.

Indeed. This is why very low noise at base ISO is one important characteristics of DSLR sensors. It does affect the ability to sharpen and therefore the perceived resolution.

It should be clear by now that perceived resolution is affected both by the native lack of softness but also by the extend to which softness can be compensated by sharpening without introducing unwanted side effects.

Cheers,
Bernard
Logged

A few images online here!
John Camp
Sr. Member
****
Offline Offline

Posts: 1258


« Reply #133 on: November 24, 2010, 12:36:58 AM »
ReplyReply

Mark,

Yes, but the M9 is not a dSLR and, because of the short lens to sensor distance, there might not be room for one. Remember, with the M8 there was not even room for an IR filter. One can reduce the color effects in Moire, but the actual aliasing is very difficult to remove. The old Kodak full frame dSLR also didn't have a blur filter and it was notorious for moire. As you say, there are trade offs. One can pay to have the blur filter removed from a dSLR, but few do.


The M8 did have an IR filter, though not an AA filter. (A mistake often made -- even the Wiki says there was no IR filter.) The filter was simply not strong enough, necessitating the use of the on-lens filters.
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8877


« Reply #134 on: November 24, 2010, 12:44:47 AM »
ReplyReply

Indeed. This is why very low noise at base ISO is one important characteristics of DSLR sensors. It does affect the ability to sharpen and therefore the perceived resolution.

It should be clear by now that perceived resolution is affected both by the native lack of softness but also by the extend to which softness can be compensated by sharpening without introducing unwanted side effects.

Cheers,
Bernard


That's absolutely true. Without sharpening entering into the equation, the P65+ has lower SNR, better tonality and better color sensitivity than the D3X, at base ISO. The D3X excels in comparison with the P65, only with regard to DR at the pixel level, and at normalised print sizes that don't involve interpolation of the D3X file.

If one attempts to increase the sharpening of the D3X image in an attempt to match the resolution of the P65+ image, then the gap that already exists between the D3X and the P65+ with regard to SNR and tonality will increase.

However, this difference is not nearly as 'marked' comparing the P645D with the D3X. At normalised print sizes, the SNR, tonality and color sensitivity is very closely the same, yet the DR of the D3X is a full stop higher.

To what extent increased sharpening of the D3X image might increase noise, and to what extent such increase in noise would be apparent on a print, remains to be demonstrated.

I've seen no such comparisons. But I have heard of a comparison at A3+ size between the Canon G10 and the P65+. And Michael can testify that experienced photographers could not tell the difference in the prints, apart from the shallower DoF of the P65+ because Michael had not used the appropriate F stops to equalize DoF.

Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2773



« Reply #135 on: November 24, 2010, 09:06:46 AM »
ReplyReply

Interesting points Bill. As you may know, Capture One has a Moire filter - the second to last tool of the Details panel. It allows us to adjust for the amount and the pattern, quite effectively.

Mark,

I don't use Capture 1, so I have no experience with the tool. Nikon Capture does have such a tool, but it is nearly worthless. Did you employ the Moire filter in the images you posted? In looking at them I was not able to detect any Moire. Moire shows up most clearly when the subject contains regularly repeating details near the resolution limit of the sensor. Such details do not often occur in landscapes, but do show up in architecture and fashion photography with fabrics. I'm not sure about bricks Smiley.

Those who subscribe to Digilloyd's DAP forum should look at his stunning image of the North part of the Golden Gate Bridge taken with the Leica S2. Moire is visible only in the grilles of some air compressors near one of the footings of the bridge (he posts the DNG file for download, but does not permit publishing of the images). The artifacts might not appear in a print, but could be removed with local editing.

On the other hand, his shots of the mosaic on the church demonstrates strong Moire when the image is viewed at 100%. The Nikon D3 in that report demonstrated less aliasing, but the per pixel image was soft. The S2 aliasing might not be visible in a normal sized print because of down sampling of the large megapixel file. On the other hand, with a low megapixel camera such as the Nikon D70 (a 6 MP camera with a weak blur filter), aliasing is not down sampled and appears full strength in an 8*10 inch print. I had a number of prints ruined with that camera, which was my first experience with digital.

Opinions vary on the effect of aliasing. Even when it is not apparent as Moire, some think that the false detail it produces lends a harsh look to the image which may be misinterpreted as sharpness. Others disagree. Diglloyd's initial impression was that the D3x would be better off for landscape work without the filter. In an addendum, he demonstrated how much of the softness imparted by the blur filter on the D3x could be removed by deconvolution sharpening.

So in the final analysis, the Moire of the MFDBs can be removed by software when it is objectionable. The MFDBs require much less sharpening. The softening of the blur filter on the D3x can be mitigated with deconvolution sharpening, but the sharpening may produce artifacts.

Diglloyd is one of the few testers presenting 100% views of files and also supplies the raw files. Such data from MFDBs are rarely seen online. I have yet to see a MFDB shot of a properly exposed Stouffer wedge.

Regards,

Bill

Logged
ndevlin
Sr. Member
****
Offline Offline

Posts: 504



WWW
« Reply #136 on: November 24, 2010, 09:44:18 AM »
ReplyReply


Diglloyd is one of the few testers presenting 100% views of files and also supplies the raw files. Such data from MFDBs are rarely seen online. I have yet to see a MFDB shot of a properly exposed Stouffer wedge.


I've never been to Stouffer Wedge. Is is pretty this time of year, or should I wait till Spring? That's when Adams made most of his best shots there, if I remember his biography correctly.

- N.
Logged

Nick Devlin   @onelittlecamera
michael
Administrator
Sr. Member
*****
Offline Offline

Posts: 4803



« Reply #137 on: November 24, 2010, 10:10:00 AM »
ReplyReply

As I recall Nick, it's located in MFT (Mountain Fiord Tundra), which means there's little light this time of year. Better wait till summer.

Michael
Logged
Eric Myrvaagnes
Sr. Member
****
Offline Offline

Posts: 7897



WWW
« Reply #138 on: November 24, 2010, 04:41:51 PM »
ReplyReply

I've never been to Stouffer Wedge. Is is pretty this time of year, or should I wait till Spring? That's when Adams made most of his best shots there, if I remember his biography correctly.

- N.
I've always found Stouffer Wedge much more impressive than the common Stouff Wedge, but Stouffest Wedge is the most spectacular of all. Grin

-E.
Logged

-Eric Myrvaagnes

http://myrvaagnes.com  Visit my website. New images each season.
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6891


WWW
« Reply #139 on: November 24, 2010, 04:43:45 PM »
ReplyReply

Mark,

I don't use Capture 1, so I have no experience with the tool. Nikon Capture does have such a tool, but it is nearly worthless. Did you employ the Moire filter in the images you posted? In looking at them I was not able to detect any Moire. Moire shows up most clearly when the subject contains regularly repeating details near the resolution limit of the sensor. Such details do not often occur in landscapes, but do show up in architecture and fashion photography with fabrics. I'm not sure about bricks Smiley.

Those who subscribe to Digilloyd's DAP forum should look at his stunning image of the North part of the Golden Gate Bridge taken with the Leica S2. Moire is visible only in the grilles of some air compressors near one of the footings of the bridge (he posts the DNG file for download, but does not permit publishing of the images). ..............

Regards,

Bill



Bill - no moire from bricks therefore no moire filter used on the images.

I have seen moire in a Leica M9 shot containing a metallic air conditioning compressor sitting on a roof. In that particular case, visible evidence of the moire was removed by de-saturating that one item.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Pages: « 1 ... 5 6 [7] 8 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad