This topic is in the 2015 archive. You can read it and like its posts here. It is read only.
Archive 2015No low pass filter
p.4 #1No low pass filter
Well ... as to the concept that the RAW shouldn't really need any sharpening applied, that's a bit counter to the entire concept of RAW.
Film mfr's designed diff films with diff levels of contrast response in the film. Then the processing continued to extend that contrast to a standardized amount or a custom amount depending on what a person wanted to achieve. Then in the case of prints, an additional step of contrast consideration was at ones disposal.
So, in reality there has NEVER been a scenario (okay maybe an argument for Polaroid) in which a truly consistent standard existed for what the contrast levels ooc should be ... just that with chromes (which I also shot) the fiddling was all "hands off" @ the lab doing it to a standardized process for a given film (i.e. Ektachrome vs. Kodachrome vs. Fujichrome, etc.) Which of those was the "right" one ... folks are probably still holding that debate in a think tank somewhere.
The point of establishing RAW wasn't to eliminate the need for additional processing, but rather to provide a base from which we can proceed to our own design of additional processing. In that regard ... by intent and design ... RAW is what it was intended to be, so that we in essence are allowed to be our own (film) profile designers and lab techs.
As such, sharpening is a function of contrast. Expecting RAW files to already be sharpened is like saying you expect the contrast level to already be set in RAW. Raw has no way of knowing if you want a high or low contrast level. It has no way of knowing if your subject and/or lighting is high or low contrast.
In that regard, the entire point of RAW is that WE get to choose the parameters ourselves. This is what the photographic community clamored for along the way. And we now have it. The expectation that RAW "ought" to be already sharpened sooc is the antithesis of RAW.
I mean, if you really want to grill your own (raw) steak ... you don't expect it to be already cooked when you get it from the butcher. That's kinda silly to even consider that the butcher would give you a steak cooked and prepared specifically to your liking. But, that's kinda what your'e saying when you suggest that files should already be sharpened (read contrast level established) to your liking sooc, when the baseline for RAW is "uncooked" so that we can all cook to our own desired levels. Some folks like a medium rare steak and others a well done one ... so the butcher gives each a raw one and they can cook it to their own liking.
I empathize with Ben greatly regarding vision. But, that doesn't mean that the approach and design of RAW was somehow done in error. For those who do NOT want to cook their own steak, they go to a restaurant and subscribe to the chef's approach, with a degree of input given to how they would like the outcome, but not to the matter of how to achieve it.
If we don't want to do our own sharpening, then we have the option to ascribe an input to how we would like the outcome @ some, little, lot of sharpening and leave it in the hands of the OEM mfr, LR, C1, DXO or some other programming that has been done for us. If that doesn't satisfy us, we can jump around to different programming (i.e. diff chef's) to find someone that satisfies us. In the end, it is either do your own, or find someone else to do it for you.
But, in either case ... RAW is the baseline from which to afford the nearly infinite variability that the industry desired from which to function.
Now, as to the "cancelling" AA filters ... my .02 is "pi$$ on that". What's the point of fuzzing it twice even if we are able to retrieve a degree of cancellation offset. To me, that remains a (somewhat) wasted effort from what it could be if you simply omitted the AA filter (or plain glass) If that is the direction that the mfr takes it, so be it ... but if that's all you got, I'll just stick with the regular AA filter and use more sharpening all around.
As is my SLR/C and my FZ1000 are sans a "cancelling" AA filter. To the best of my knowledge, MFD doesn't play around with "cancelling" AA filters either. It is either in or it's out. To me, our sharpening algorithms are omni-directional. Either use an AA filter that is omni-directional blur or none at all. I've never been impressed by the look of the D800E files, despite the difference from the D800. It still looks contrast comprised from from the "double pass" to me. Seems "half-baked" to me ... kinda like a Ferrari with training wheels still on it. You either want max performance (and it's inherent perils that need to be learned how to handle) or you want safety and ease of utility @ a Camry.
Edited on Feb 05, 2015 at 04:26 PM · View previous versions
Film mfr's designed diff films with diff levels of contrast response in the film. Then the processing continued to extend that contrast to a standardized amount or a custom amount depending on what a person wanted to achieve. Then in the case of prints, an additional step of contrast consideration was at ones disposal.
So, in reality there has NEVER been a scenario (okay maybe an argument for Polaroid) in which a truly consistent standard existed for what the contrast levels ooc should be ... just that with chromes (which I also shot) the fiddling was all "hands off" @ the lab doing it to a standardized process for a given film (i.e. Ektachrome vs. Kodachrome vs. Fujichrome, etc.) Which of those was the "right" one ... folks are probably still holding that debate in a think tank somewhere.
The point of establishing RAW wasn't to eliminate the need for additional processing, but rather to provide a base from which we can proceed to our own design of additional processing. In that regard ... by intent and design ... RAW is what it was intended to be, so that we in essence are allowed to be our own (film) profile designers and lab techs.
As such, sharpening is a function of contrast. Expecting RAW files to already be sharpened is like saying you expect the contrast level to already be set in RAW. Raw has no way of knowing if you want a high or low contrast level. It has no way of knowing if your subject and/or lighting is high or low contrast.
In that regard, the entire point of RAW is that WE get to choose the parameters ourselves. This is what the photographic community clamored for along the way. And we now have it. The expectation that RAW "ought" to be already sharpened sooc is the antithesis of RAW.
I mean, if you really want to grill your own (raw) steak ... you don't expect it to be already cooked when you get it from the butcher. That's kinda silly to even consider that the butcher would give you a steak cooked and prepared specifically to your liking. But, that's kinda what your'e saying when you suggest that files should already be sharpened (read contrast level established) to your liking sooc, when the baseline for RAW is "uncooked" so that we can all cook to our own desired levels. Some folks like a medium rare steak and others a well done one ... so the butcher gives each a raw one and they can cook it to their own liking.
I empathize with Ben greatly regarding vision. But, that doesn't mean that the approach and design of RAW was somehow done in error. For those who do NOT want to cook their own steak, they go to a restaurant and subscribe to the chef's approach, with a degree of input given to how they would like the outcome, but not to the matter of how to achieve it.
If we don't want to do our own sharpening, then we have the option to ascribe an input to how we would like the outcome @ some, little, lot of sharpening and leave it in the hands of the OEM mfr, LR, C1, DXO or some other programming that has been done for us. If that doesn't satisfy us, we can jump around to different programming (i.e. diff chef's) to find someone that satisfies us. In the end, it is either do your own, or find someone else to do it for you.
But, in either case ... RAW is the baseline from which to afford the nearly infinite variability that the industry desired from which to function.
Now, as to the "cancelling" AA filters ... my .02 is "pi$$ on that". What's the point of fuzzing it twice even if we are able to retrieve a degree of cancellation offset. To me, that remains a (somewhat) wasted effort from what it could be if you simply omitted the AA filter (or plain glass) If that is the direction that the mfr takes it, so be it ... but if that's all you got, I'll just stick with the regular AA filter and use more sharpening all around.
As is my SLR/C and my FZ1000 are sans a "cancelling" AA filter. To the best of my knowledge, MFD doesn't play around with "cancelling" AA filters either. It is either in or it's out. To me, our sharpening algorithms are omni-directional. Either use an AA filter that is omni-directional blur or none at all. I've never been impressed by the look of the D800E files, despite the difference from the D800. It still looks contrast comprised from from the "double pass" to me. Seems "half-baked" to me ... kinda like a Ferrari with training wheels still on it. You either want max performance (and it's inherent perils that need to be learned how to handle) or you want safety and ease of utility @ a Camry.
Edited on Feb 05, 2015 at 04:26 PM · View previous versions
Feb 05, 2015 at 04:10 PM
p.4 #2No low pass filter
I understand the concept of raw and I have never done jpg. But it seems to me a raw ought to at least capture what entered the lens and we can tweak from there. It ought to be as sharp as the lens and focus allows, and the color ought to be correct.
For me post processing ought to be about taking what was there and tweaking it to some enhanced level. More saturation, boosting shadows, cropping, removing dust or other stuff we don't want.
I understand we need an AA filter to perform the anti aliasing function. So design a sensor that does not alias.
I never sharpened my Kodachromes and yes, I stood inches from the screen and they were too often soft. but usually because they were out of focus. Now I sharpen every image by exactly the same formula because the camera and lens and method of taking are all the same.
I guess I was an engineer too long and I always hated the crippling junk that marketing or accounting forced on us. I am not an optical engineer and don't pretend to understand this. But I did design space level microwave switches and we did not cripple them. But when I went to commercial firms, it was always cost savings.
For me post processing ought to be about taking what was there and tweaking it to some enhanced level. More saturation, boosting shadows, cropping, removing dust or other stuff we don't want.
I understand we need an AA filter to perform the anti aliasing function. So design a sensor that does not alias.
I never sharpened my Kodachromes and yes, I stood inches from the screen and they were too often soft. but usually because they were out of focus. Now I sharpen every image by exactly the same formula because the camera and lens and method of taking are all the same.
I guess I was an engineer too long and I always hated the crippling junk that marketing or accounting forced on us. I am not an optical engineer and don't pretend to understand this. But I did design space level microwave switches and we did not cripple them. But when I went to commercial firms, it was always cost savings.
Feb 05, 2015 at 04:25 PM
p.4 #3No low pass filter
Ben,
I long for the tiered Foveon-like sensor without an AA filter that Canon has in R&D. Fuji has their X-Trans and there is the Foveon itself ... but until Canon or Nikon breaks away from the Bayer array, it doesn't look hopeful that they will fully omit the AA filter as Leica has.
That said ... you may want to try using DPP processing for your sharpening. At least that way you are matching your OEM RAW with OEM sharpening algorithms. Compared to LR or PS, I felt like DPP (back when I was comparing them) did a better job @ cleanly sharpening my 5D files. But, I got "lazy" at the workflow of moving from DPP to PS, so I developed my own sharpening workflow in PS instead.
Given your visual challenges ... DPP might be worth looking into as a hybrid approach.
I long for the tiered Foveon-like sensor without an AA filter that Canon has in R&D. Fuji has their X-Trans and there is the Foveon itself ... but until Canon or Nikon breaks away from the Bayer array, it doesn't look hopeful that they will fully omit the AA filter as Leica has.
That said ... you may want to try using DPP processing for your sharpening. At least that way you are matching your OEM RAW with OEM sharpening algorithms. Compared to LR or PS, I felt like DPP (back when I was comparing them) did a better job @ cleanly sharpening my 5D files. But, I got "lazy" at the workflow of moving from DPP to PS, so I developed my own sharpening workflow in PS instead.
Given your visual challenges ... DPP might be worth looking into as a hybrid approach.
Feb 05, 2015 at 04:37 PM
p.4 #4No low pass filter
ben egbert wrote:
But it seems to me a raw ought to at least capture what entered the lens and we can tweak from there.
Basically it does, and we tweak from there.
But it captures what exited the lens. The lens has many effects on the image.
ben egbert wrote:
It ought to be as sharp as the lens and focus allows, and the color ought to be correct.
Then you have a gap in understanding the technology. To be
as sharp as the modern sharp lenses that you own or aspire
to own, it has to have at least twice the line pair resolution
of the lens. Twice because the line pairs are pairs, and twice
again because of the Nyquist Sampling Theorem which shows
that to avoid aliasing and false data you have to sample at
twice the highest frequency -- twice the highest line pair rate.
The line pairs are square waveforms, but at the limits of lens
resolution they are effectively sine waves.
Such a sensor would have about 340 sensels per mm or
12,000 wide x 8,000 high or 96 MPixel. Something to
aim for, but by that time lens technology (and expense)
may have raised the bar.
The colour must be corrected because the sensor doesn't
sense colour. It is similar to how the signal coming from
the human optical sensor is uncorrected for colour and
needs to be corrected by the visual cortex and experience.
We see a white paper as white in sunlight and in tungsten
(incandescent) light because our visual cortex colour corrects.
ben egbert wrote:
So design a sensor that does not alias
Easy. Use twice the linear resolution of the finest lens you
might ever own in a lifetime. It's a mathematical fact. See
Nyquist Theorem.
But it seems to me a raw ought to at least capture what entered the lens and we can tweak from there.
Basically it does, and we tweak from there.
But it captures what exited the lens. The lens has many effects on the image.
ben egbert wrote:
It ought to be as sharp as the lens and focus allows, and the color ought to be correct.
Then you have a gap in understanding the technology. To be
as sharp as the modern sharp lenses that you own or aspire
to own, it has to have at least twice the line pair resolution
of the lens. Twice because the line pairs are pairs, and twice
again because of the Nyquist Sampling Theorem which shows
that to avoid aliasing and false data you have to sample at
twice the highest frequency -- twice the highest line pair rate.
The line pairs are square waveforms, but at the limits of lens
resolution they are effectively sine waves.
Such a sensor would have about 340 sensels per mm or
12,000 wide x 8,000 high or 96 MPixel. Something to
aim for, but by that time lens technology (and expense)
may have raised the bar.
The colour must be corrected because the sensor doesn't
sense colour. It is similar to how the signal coming from
the human optical sensor is uncorrected for colour and
needs to be corrected by the visual cortex and experience.
We see a white paper as white in sunlight and in tungsten
(incandescent) light because our visual cortex colour corrects.
ben egbert wrote:
So design a sensor that does not alias
Easy. Use twice the linear resolution of the finest lens you
might ever own in a lifetime. It's a mathematical fact. See
Nyquist Theorem.
Feb 05, 2015 at 04:39 PM
p.4 #5No low pass filter
Monito wrote:
Basically it does, and we tweak from there.
But it captures what exited the lens. The lens has many effects on the image.
Then you have a gap in understanding the technology. To be
as sharp as the modern sharp lenses that you own or aspire
to own, it has to have at least twice the line pair resolution
of the lens. Twice because the line pairs are pairs, and twice
again because of the Nyquist Sampling Theorem which shows
that to avoid aliasing and false data you have to sample at
twice the highest frequency -- twice the highest line pair rate.
The line pairs are square waveforms,...Show more →
I should have said as sharp as the lens, focus and current sensor allows. I think 50MP is plenty of resolution for the prints I aspire to make, maybe even with the AA filter. I may stop sharpening altogether. I actually do very little now.
I think lenses still have issues that I can see even at 22mpix, corners and some Spherical aberrations. Which by the way the aberrations and distortion can be auto corrected when I open the file. I can also apply my camera color profile. So maybe I will just sharpen in ACR when I open the file per a preset amount.
ben egbert wrote
But it seems to me a raw ought to at least capture what entered the lens and we can tweak from there.But it captures what exited the lens. The lens has many effects on the image.
ben egbert wrote
It ought to be as sharp as the lens and focus allows, and the color ought to be correct.as sharp as the modern sharp lenses that you own or aspire
to own, it has to have at least twice the line pair resolution
of the lens. Twice because the line pairs are pairs, and twice
again because of the Nyquist Sampling Theorem which shows
that to avoid aliasing and false data you have to sample at
twice the highest frequency -- twice the highest line pair rate.
The line pairs are square waveforms,...Show more →
ben egbert wrote
So design a sensor that does not aliasI should have said as sharp as the lens, focus and current sensor allows. I think 50MP is plenty of resolution for the prints I aspire to make, maybe even with the AA filter. I may stop sharpening altogether. I actually do very little now.
I think lenses still have issues that I can see even at 22mpix, corners and some Spherical aberrations. Which by the way the aberrations and distortion can be auto corrected when I open the file. I can also apply my camera color profile. So maybe I will just sharpen in ACR when I open the file per a preset amount.
Feb 05, 2015 at 05:24 PM
p.4 #6No low pass filter
racer-ex wrote:
Okay I am a learning photographer but I do not know what no low pass filter is be kind please...
It's something every camera maker (except Canon) offers...
Okay I am a learning photographer but I do not know what no low pass filter is be kind please...
It's something every camera maker (except Canon) offers...
Feb 06, 2015 at 01:35 PM
p.4 #7No low pass filter
racer-ex wrote:
Okay I am a learning photographer but I do not know what no low pass filter is be kind please...
molson wrote:
It's something every camera maker (except Canon) offers...
Read the news, molson.
Okay I am a learning photographer but I do not know what no low pass filter is be kind please...
molson wrote:
It's something every camera maker (except Canon) offers...
Read the news, molson.
Feb 06, 2015 at 01:39 PM
p.4 #8No low pass filter
Monito wrote:
Read the news, molson.
The fanboy news, where the deluded masses think that Canon has finally figured out how to build a camera without a blur filter?
Or the real news, that Canon is using the same modified low-pass filter that Nikon used on the D800E, instead of eliminating the low-pass filter like they did on the D810 and D7100 (and what every other manufacturer is now offering on at least some of their camera...)?
racer-ex wrote
Okay I am a learning photographer but I do not know what no low pass filter is be kind please...molson wrote
It's something every camera maker (except Canon) offers...The fanboy news, where the deluded masses think that Canon has finally figured out how to build a camera without a blur filter?
Or the real news, that Canon is using the same modified low-pass filter that Nikon used on the D800E, instead of eliminating the low-pass filter like they did on the D810 and D7100 (and what every other manufacturer is now offering on at least some of their camera...)?
Feb 07, 2015 at 05:21 PM
p.4 #9No low pass filter
AA filters make a lot of sense on bayer array sensors in order to prevent moire from rearing its ugly head. That said, with the huge increases in resolution, I would hope that canon is putting a lighter AA filter in the 5Ds than that currently found in the 5D3.
It would be nice if we stopped using bayer sensors.
It would be nice if we stopped using bayer sensors.
Feb 07, 2015 at 05:33 PM
p.4 #10No low pass filter
molson wrote:
Read the news, molson.The fanboy news, where the deluded masses think that Canon has finally figured out how to build a camera without a blur filter?
Or the real news, that Canon is using the same modified low-pass filter that Nikon used on the D800E, instead of eliminating the low-pass filter like they did on the D810 and D7100 (and what every other manufacturer is now offering on at least some of their camera...)?
+1 @ diff between offset / cancelling "double blur"
vs. a true absence of any AA blurring.
Monito wrote
racer-ex wrote
Okay I am a learning photographer but I do not know what no low pass filter is be kind please...molson wrote
It's something every camera maker (except Canon) offers...Or the real news, that Canon is using the same modified low-pass filter that Nikon used on the D800E, instead of eliminating the low-pass filter like they did on the D810 and D7100 (and what every other manufacturer is now offering on at least some of their camera...)?
+1 @ diff between offset / cancelling "double blur"
vs. a true absence of any AA blurring. Feb 07, 2015 at 05:51 PM
p.4 #11No low pass filter
kezeka wrote:
AA filters make a lot of sense on bayer array sensors in order to prevent moire from rearing its ugly head. That said, with the huge increases in resolution, I would hope that canon is putting a lighter AA filter in the 5Ds than that currently found in the 5D3.
Bayer has nothing to do with moiré or aliasing.
Aliasing (moiré is one example) is caused by sampling data
that includes (spatial) frequencies more than half as the
frequency of the sampler.
You get moiré from Foveons too.
The human retina has sensels (rods and cones) in a Poisson
distribution, not a regular array. As such it too gets aliasing,
but in a random way so that repetitive scenes (such as fabric)
are not weird looking.
AA filters make a lot of sense on bayer array sensors in order to prevent moire from rearing its ugly head. That said, with the huge increases in resolution, I would hope that canon is putting a lighter AA filter in the 5Ds than that currently found in the 5D3.
Bayer has nothing to do with moiré or aliasing.
Aliasing (moiré is one example) is caused by sampling data
that includes (spatial) frequencies more than half as the
frequency of the sampler.
You get moiré from Foveons too.
The human retina has sensels (rods and cones) in a Poisson
distribution, not a regular array. As such it too gets aliasing,
but in a random way so that repetitive scenes (such as fabric)
are not weird looking.
Feb 07, 2015 at 10:06 PM
p.4 #12No low pass filter
Ah, molson, I see now that you were referring to the implementation
detail (cancellation versus elimination), while I thought the question
asker was referring to the concept (AA filter versus no AA).
detail (cancellation versus elimination), while I thought the question
asker was referring to the concept (AA filter versus no AA).
Feb 07, 2015 at 10:08 PM
p.4 #13No low pass filter
ben egbert wrote:
I shot 25. But tell me, why does an image need sharpening if the lens is good and the focus was good?
One issue is that there is no single right way to sharpen. RAW leaves that decision (or, should I say, those decisions) to the photographer rather than locking him/her out of that creative decision.
RAW (more or less) represents actual data from the sensor with no (or little) interpretation by the camera or other software. (Obviously, since bits are not colors, some level of interpretation is required to display an image at all.)
The term "sharpening" in digital photography could be said to be a misnomer. What we do does increase the apparent sharpness, but the operation is really one that increases contrast. Along edges (to keep it simple) when there is a darker and a lighter "side" to the edge, it lightens the lighter side and darkens the darker side, creating greater luminosity contrast along that edge. (Technically these things are not limited to what we think of as edges.)
This is done by post-processing (e.g - further processing) the raw data.
In the end, aside from curiosity, it sort of doesn't matter "why" it works this way — you could simply live with the fact that it does! On the other hand, I'd point out that understanding how it works can help you make best decisions about how to optimize image quality in post.
The concept of straight out of camera is perhaps one best left behind.
Dan,
who was tempted for a moment to format his entire message like this and force others to adapt to his personal preference:
The terms "sharp-
ening" in digital
photography could
be said to be a mis-
nomer. What we
do does increase
the apparent sharp
-ness, but the oper-
ation is really one
that increases con-
trast. Along edges
(to keep it simple)
when there is a
darker and a
lighter "side" to
the edge, it light-
ens the lighter side
and darkens the
darker side, creat-
ing greater luminos-
ity contrast along
that edge. (Tech-
nically these things
are not limited to
what we think of
as edges.)
Edited on Feb 10, 2015 at 10:25 AM · View previous versions
I shot 25. But tell me, why does an image need sharpening if the lens is good and the focus was good?
One issue is that there is no single right way to sharpen. RAW leaves that decision (or, should I say, those decisions) to the photographer rather than locking him/her out of that creative decision.
RAW (more or less) represents actual data from the sensor with no (or little) interpretation by the camera or other software. (Obviously, since bits are not colors, some level of interpretation is required to display an image at all.)
The term "sharpening" in digital photography could be said to be a misnomer. What we do does increase the apparent sharpness, but the operation is really one that increases contrast. Along edges (to keep it simple) when there is a darker and a lighter "side" to the edge, it lightens the lighter side and darkens the darker side, creating greater luminosity contrast along that edge. (Technically these things are not limited to what we think of as edges.)
This is done by post-processing (e.g - further processing) the raw data.
In the end, aside from curiosity, it sort of doesn't matter "why" it works this way — you could simply live with the fact that it does! On the other hand, I'd point out that understanding how it works can help you make best decisions about how to optimize image quality in post.
The concept of straight out of camera is perhaps one best left behind.
Dan,
who was tempted for a moment to format his entire message like this and force others to adapt to his personal preference:
The terms "sharp-
ening" in digital
photography could
be said to be a mis-
nomer. What we
do does increase
the apparent sharp
-ness, but the oper-
ation is really one
that increases con-
trast. Along edges
(to keep it simple)
when there is a
darker and a
lighter "side" to
the edge, it light-
ens the lighter side
and darkens the
darker side, creat-
ing greater luminos-
ity contrast along
that edge. (Tech-
nically these things
are not limited to
what we think of
as edges.)
Edited on Feb 10, 2015 at 10:25 AM · View previous versions
Feb 07, 2015 at 10:33 PM
p.4 #14No low pass filter
It
would be
simplest if Fred
formatted text automatically
for us (perhaps to user selectable width).
But it must be said that Fred has made an excellent site.
I browse
very wide so
that I can see the
photos, and for other
sites (most of which format text to a reasonable width).
would be
simplest if Fred
formatted text automatically
for us (perhaps to user selectable width).
But it must be said that Fred has made an excellent site.
I browse
very wide so
that I can see the
photos, and for other
sites (most of which format text to a reasonable width).
Feb 07, 2015 at 10:52 PM
p.4 #15No low pass filter
+1 @ sharpening = contrast enhancement
Ben,
When we shot slides ... the image projected onto the light sensitive film was further enhanced chemically through processing. It was "hands off" to us via standard lab processing, but it was still a necessary process.
We made our choices @ film (Ektachrome vs. Kodachrome vs. Fujichrome) ... each with different base and chemical formula to generate the contrast levels (and thus perceived sharpness) variably desired. Today, the "unsharpened" raw affords us (or requires us & half-empty / half-full) the opportunity to input to this variably as we desire.
Ben,
When we shot slides ... the image projected onto the light sensitive film was further enhanced chemically through processing. It was "hands off" to us via standard lab processing, but it was still a necessary process.
We made our choices @ film (Ektachrome vs. Kodachrome vs. Fujichrome) ... each with different base and chemical formula to generate the contrast levels (and thus perceived sharpness) variably desired. Today, the "unsharpened" raw affords us (or requires us & half-empty / half-full) the opportunity to input to this variably as we desire.
Feb 07, 2015 at 11:24 PM
p.4 #16No low pass filter
EB-1 wrote:
Yeah. I see it even with the D7100 which has a smaller pixel pitch. Sometimes there are colored sparkles around the individual hairs.
EBH
Yep, the D7100 sucks ass for that reason alone. I don't much care for that sensor's crunchy/crisp rendering. When the time comes,I might still go with a D810 because the D750's controls are a bit skimpy but I really do prefer the artefact-free fuzziness provided by an AA filter. Dilemma.
Christian H wrote
With wildlife (fur, feathers) I routinely get the 'false color' moire even with the D800. Not looking forward to a body without an AA filter when my D800 dies.EBH
Yep, the D7100 sucks ass for that reason alone. I don't much care for that sensor's crunchy/crisp rendering. When the time comes,I might still go with a D810 because the D750's controls are a bit skimpy but I really do prefer the artefact-free fuzziness provided by an AA filter. Dilemma.
Feb 08, 2015 at 07:23 AM
p.4 #17No low pass filter
While we don't have any side by side comps of two cameras for a given scene / lens at AA vs. "cancelling" AA ... I think that the fine detail even with the AA filter of the 5Ds is likely sufficient for a majority of applications.
Of course, being the non-AA oriented person that I am ... I'll certainly scrutinize the diff once we have RAW side by sides. Mostly, I'll be looking for the amount of contrast degradation between the two. I suspect that the difference will be within a very "bridgeable" amount via post at these resolutions.
Of course, being the non-AA oriented person that I am ... I'll certainly scrutinize the diff once we have RAW side by sides. Mostly, I'll be looking for the amount of contrast degradation between the two. I suspect that the difference will be within a very "bridgeable" amount via post at these resolutions.
Feb 08, 2015 at 07:37 AM
p.4 #18No low pass filter
RustyBug wrote:
+1 @ sharpening = contrast enhancement
Ben,
When we shot slides ... the image projected onto the light sensitive film was further enhanced chemically through processing. It was "hands off" to us via standard lab processing, but it was still a necessary process.
We made our choices @ film (Ektachrome vs. Kodachrome vs. Fujichrome) ... each with different base and chemical formula to generate the contrast levels (and thus perceived sharpness) variably desired. Today, the "unsharpened" raw affords us (or requires us & half-empty / half-full) the opportunity to input to this variably as we desire.
Yep and it's hands off for me too. I sharpen blindly, via a formula hoping that all the eagle eyed folks here will not realize it.
In fact I got my formula from GDan and now I see he has modified it from a couple years back. Thats ok, I did too.
The fact that I use the same lenses the same camera and shoot the same subjects and don't do shallow dof makes global sharpening easy. I just need to avoid sharpening noise.
+1 @ sharpening = contrast enhancement
Ben,
When we shot slides ... the image projected onto the light sensitive film was further enhanced chemically through processing. It was "hands off" to us via standard lab processing, but it was still a necessary process.
We made our choices @ film (Ektachrome vs. Kodachrome vs. Fujichrome) ... each with different base and chemical formula to generate the contrast levels (and thus perceived sharpness) variably desired. Today, the "unsharpened" raw affords us (or requires us & half-empty / half-full) the opportunity to input to this variably as we desire.
Yep and it's hands off for me too. I sharpen blindly, via a formula hoping that all the eagle eyed folks here will not realize it.
In fact I got my formula from GDan and now I see he has modified it from a couple years back. Thats ok, I did too.
The fact that I use the same lenses the same camera and shoot the same subjects and don't do shallow dof makes global sharpening easy. I just need to avoid sharpening noise.
Feb 08, 2015 at 10:56 AM
p.4 #19No low pass filter
ben egbert wrote:
In fact I got my formula from GDan and now I see he has modified it from a couple years back. Thats ok, I did too.
:-)
The main change I made to my starting point for sharpening is that rather than doing a USM layer in Photoshop, I now use the "clarity" slider in ACR. (Sometimes in LR, instead.)
Dan
In fact I got my formula from GDan and now I see he has modified it from a couple years back. Thats ok, I did too.
:-)
The main change I made to my starting point for sharpening is that rather than doing a USM layer in Photoshop, I now use the "clarity" slider in ACR. (Sometimes in LR, instead.)
Dan
Feb 08, 2015 at 11:33 AM
p.4 #20No low pass filter
gdanmitchell wrote:
:-)
The main change I made to my starting point for sharpening is that rather than doing a USM layer in Photoshop, I now use the "clarity" slider in ACR. (Sometimes in LR, instead.)
Dan
At some point I had the values 300,.3,0 usm and 12,50,1 for large radius high amount. I thought I got those from you.
But I started using a bit of sharpening in ACR, 50,0.6,7 mask 30. so I backed off the final settings to 70%. I also do the final usm in lab.
I used to use clarity, but stopped, how much do you use, or what range?
ben egbert wrote
In fact I got my formula from GDan and now I see he has modified it from a couple years back. Thats ok, I did too.The main change I made to my starting point for sharpening is that rather than doing a USM layer in Photoshop, I now use the "clarity" slider in ACR. (Sometimes in LR, instead.)
Dan
At some point I had the values 300,.3,0 usm and 12,50,1 for large radius high amount. I thought I got those from you.
But I started using a bit of sharpening in ACR, 50,0.6,7 mask 30. so I backed off the final settings to 70%. I also do the final usm in lab.
I used to use clarity, but stopped, how much do you use, or what range?
Feb 08, 2015 at 11:48 AM