E-1 Noise v. 300D Noise

[Anthony Cheh]

I noticed that Phil Askey's luminance noise graphs show the E-1's levels to be equal to or even slightly better than the Canon 300D's (hopefully I read the graphs correctly). He also points out that the higher default sharpness, saturation and contrast settings for the 300D result in elevated noise levels compared with the softer default settings for the Canon 10D. Sean Reid mentioned that he found the E-1's color saturation level to be somewhat higher than the 10D's.

So far no one seems to be complaining about 300D noice levels. Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the E-1, 10D, D100 and S2 were equalized for sharpness, saturation and contrast, rather than simply tested at the factory default settings.

One would think that a camera with relatively low default settings would have an advantage if the default setting was used for test comparisons. But that advantange might not be representative of real world results, with a greater degree of sharpening, saturation and contrast required in camera or post processing to achieve a visually pleasing result.

Tony

 

[Bram Bos]

Hi,

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso 1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Tan35125]

Anthony Cheh wrote:
Sean Reid mentioned that he

found the E-1's color saturation level to be somewhat higher than
the 10D's.

CS2 (Color Saturation level 2), the camera's default maybe stronger than some would like it to be. Personally I have been using CS1. It gives me what I think for me, to be the best combination of not overly hot, yet not overly dull colours.

 

[Anthony Cheh]

Bram:

If you have a chance to read my other posts on this forum, you will see that I only shoot at ISO 100 to 400 and I am quite pleased with my new E-1 and 14-54 Lens. I bought the E-1 instead of a Canon 10D, Nikon D100 or Pentax istD because the E-1 and the 14-54 lens are in fact a much better match to my shooting style and preferences.

However, I don't assume that everyone shoots pictures under the same lighting conditions that I do, and I don't assume that most of the people who have a concern about noise levels are "professional grey patch shooters". From the tone and seriousness of their comments, information and questions, I instead assume most of them are serious picture takers/makers. Unlike you and I, some eople have a real need for low noise in low light.

Given the interest in this area, I found the fact that the 300D (which no one has described as "noisy") has apparently the same high ISO luminance noise levels as the E-1, to be interesting. I also found it interesting that the noise level comparisons do not actually equalize the sharpness, contrast or saturation levels. These factors are not the same for different cameras but they do affect the noise test results. I do not know if this kind of equalization can be achieved.

Sean Reid's thorough and much appreciated tests indicated that the E-1 has a higher saturation level at the default setting than the Canon 10D. No one has compared the other default settings for sharpness and contrast.

What I was trying to get at is perhaps the E-1 is not at such a disadvantage with higher ISO noise levels and that only a fully equalized test could establish that issue.

Finally, this is an open forum for matters of interest concerning Olympus equipment on a website dealing primarily with technical, rather than artistic matters. There are many other sites where the balance shifts the other way. I don't find everything here relevant or interesting and the subject line usually guides me.

If, unlike some others, you find discussions about noise to be irrelevant or upsetting, you might want to avoid threads with that subject matter, rather than condemning others for whom it is relevant and interesting and incorrectly assuming that they don't really take pictures. I've been shooting for almost 40 years, long before digital cameras let alone ISO ratings and associated digital noise even existed for photographers. But photography, and digital photography in particular, is a combination of technology and art; neither can exist without the other and they are complimentary rather than antagonistic.

Tony

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Anthony Cheh]

Tan:

Thanks for the information. Interestingly, the menu on my E-1 (US model) does not have CS settings and the default is instead 0, with the range from -2 to +2.

Perhaps the difference is due to firmware changes or localization? My firmware is v. 1.0, D4040 according to the EXIF information.

Tony

found the E-1's color saturation level to be somewhat higher than
the 10D's.

CS2 (Color Saturation level 2), the camera's default maybe stronger
than some would like it to be. Personally I have been using CS1. It
gives me what I think for me, to be the best combination of not
overly hot, yet not overly dull colours.

 

[PerL]

I think this a very interesting question. I think the E-1 due to its small size and weight is ideal for traveling PJ-work, but then you have to have good low-light ability and reasonable low noise at 1600 ISO and higher.

Now the perception is that E-1 has more noise than the competition (10D, 1D, D100, D1/2H), but as you point out, its the noise after the postprocessing that counts in the real world.

I noticed that Phil Askey's luminance noise graphs show the E-1's
levels to be equal to or even slightly better than the Canon 300D's
(hopefully I read the graphs correctly). He also points out that
the higher default sharpness, saturation and contrast settings for
the 300D result in elevated noise levels compared with the softer
default settings for the Canon 10D. Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

One would think that a camera with relatively low default settings
would have an advantage if the default setting was used for test
comparisons. But that advantange might not be representative of
real world results, with a greater degree of sharpening, saturation
and contrast required in camera or post processing to achieve a
visually pleasing result.

Tony

--
http://www.pbase.com/interactive

 

[Bram Bos]

If, unlike some others, you find discussions about noise to be
irrelevant or upsetting, you might want to avoid threads with that
subject matter, rather than condemning others for whom it is
relevant and interesting and incorrectly assuming that they don't
really take pictures.

Perhaps you misinterpreted the tone of my message, and it could
be caused by the fact that English is not my native language (dutch
is). The noise-thing has been going on so long here that I probably
got tired of the issue - which I don't think is an issue at all, hence
my post.

I don't think it's irrelevant, but I do think it's not as important as
some people make it sound. Especially if the phenomenon isn't even
visible until you hit ISO 800. But that's also your point if I understand
it correctly. And the E-1 and 300D aim at very different users. Maybe
people who demand optimum noise characteristics at high ISO levels
should look a little higher than an entry level digital like the 300D?

Anyway, if my post offended you, I apologise, but it wasn't intentional.
I'm just too tired of beating a dead horse over and over again.

Regards,

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[bob oharra]

Given the constant attention the noise issue has gotten on this forum, I'm inclined to agree with Bram's characterization of "obession" with noise.

Every camera has strenths and weaknesses. The E series cameras do a superb job capturing natural colors. This is much more important to many people than a little noise on the rare occasion one really needs a high ISO setting.

I do not think the typical E-1 buyer will be put off over the noise issue.The E-1 should be a commercial sucess for Olympus (as will subsequent E models).

While I will follow the E-1 with interest, personally, I will continue to enjoy my outdated E-10........ despite all it's short coming. To my amazement the damn thing continues to produce great pictures.

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[simona]

I have to agree with Bob here. I also own an E10 and I have been generally very happy with it. Noise at high ISOs has been a problem, so I am looking for an improvement there, but the E1 seems to deliver that in spades. Of much greater concern though is the speed of the camera (start up, write times etc) and chromatic aberation. I found CA a real problem, particularly when using the TCON300S, and from all accounts, the E1 seems to deal with this issue admirably. This and the improvement in immunity to dust contamination (and general weather sealing) is an area where I think that the E1 is going to have a significant edge over competitors that also use lenses designed for 35mm bodies. Even the 1Ds exhibits noticable CA and soft edges when using wide angled L series lenses. This is not the fault of the camera or the lens, it is a simple consequence of the high refractive index of silicon dioxide within the sensor when combined with non-collumated light. From my perspective, the small noise penalty of the E1 at ISO800 and above when compared to the 10D etc is far outweighed by the advantages.

Regards

Simon

Every camera has strenths and weaknesses. The E series cameras do a
superb job capturing natural colors. This is much more important to
many people than a little noise on the rare occasion one really
needs a high ISO setting.

I do not think the typical E-1 buyer will be put off over the noise
issue.The E-1 should be a commercial sucess for Olympus (as will
subsequent E models).

While I will follow the E-1 with interest, personally, I will
continue to enjoy my outdated E-10........ despite all it's short
coming. To my amazement the damn thing continues to produce great
pictures.

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Sam Berry]

I completely agree on noise graphs. They are pretty pointless. Tone curves and many other factors influence the outcome too much, and so looking at real pictures is by far the best thing, provided white-balance is taken into account.

But...why does everyone keep saying 'complete lack of CA' when it is simply not true?

Like on Sean's shot here:
http://forums.dpreview.com/forums/read.asp?forum=1022&message=6313117

[The red line around the white] is not awful, but it looks a lot like lateral CA to me. There have been other examples but I forget which. Repeating marketing hype over and over does not make it true.
The colour moire is also a little odd too.

Sean

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Sam Berry]

I have to agree with Bob here. I also own an E10 and I have been
generally very happy with it. Noise at high ISOs has been a
problem, so I am looking for an improvement there, but the E1 seems
to deliver that in spades. Of much greater concern though is the
speed of the camera (start up, write times etc) and chromatic
aberation. I found CA a real problem, particularly when using the
TCON300S, and from all accounts, the E1 seems to deal with this
issue admirably. This and the improvement in immunity to dust
contamination (and general weather sealing) is an area where I
think that the E1 is going to have a significant edge over
competitors that also use lenses designed for 35mm bodies. Even the
1Ds exhibits noticable CA and soft edges when using wide angled L
series lenses. This is not the fault of the camera or the lens, it
is a simple consequence of the high refractive index of silicon
dioxide within the sensor when combined with non-collumated light.

Can you reference this? I have been looking for an explanation of why CA was attributed to the sensor and not to the lens. Although I cannot see how this explains longitudinal CA at all.

From my perspective, the small noise penalty of the E1 at ISO800
and above when compared to the 10D etc is far outweighed by the
advantages.

Regards

Simon

Every camera has strenths and weaknesses. The E series cameras do a
superb job capturing natural colors. This is much more important to
many people than a little noise on the rare occasion one really
needs a high ISO setting.

I do not think the typical E-1 buyer will be put off over the noise
issue.The E-1 should be a commercial sucess for Olympus (as will
subsequent E models).

While I will follow the E-1 with interest, personally, I will
continue to enjoy my outdated E-10........ despite all it's short
coming. To my amazement the damn thing continues to produce great
pictures.

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[simona]

I have to agree with Bob here. I also own an E10 and I have been
generally very happy with it. Noise at high ISOs has been a
problem, so I am looking for an improvement there, but the E1 seems
to deliver that in spades. Of much greater concern though is the
speed of the camera (start up, write times etc) and chromatic
aberation. I found CA a real problem, particularly when using the
TCON300S, and from all accounts, the E1 seems to deal with this
issue admirably. This and the improvement in immunity to dust
contamination (and general weather sealing) is an area where I
think that the E1 is going to have a significant edge over
competitors that also use lenses designed for 35mm bodies. Even the
1Ds exhibits noticable CA and soft edges when using wide angled L
series lenses. This is not the fault of the camera or the lens, it
is a simple consequence of the high refractive index of silicon
dioxide within the sensor when combined with non-collumated light.

Can you reference this? I have been looking for an explanation of
why CA was attributed to the sensor and not to the lens. Although I
cannot see how this explains longitudinal CA at all.

I am not too sure whether I can find a specific reference, but I will look into it. I guess, the truth is that there is more than one source of CA. The lens can certainly be one, but achilles heel of many digicams is chromatic aberation introduced by the sensor as result of diffraction of non collumated light leading to bleeding of the shorter wavelengths into adjacent pixels. My understanding was that this is the source of the dreaded purple fringe in high contrast situations. I will look into whether I can find a specific reference for this.

From my perspective, the small noise penalty of the E1 at ISO800
and above when compared to the 10D etc is far outweighed by the
advantages.

Regards

Simon

Every camera has strenths and weaknesses. The E series cameras do a
superb job capturing natural colors. This is much more important to
many people than a little noise on the rare occasion one really
needs a high ISO setting.

I do not think the typical E-1 buyer will be put off over the noise
issue.The E-1 should be a commercial sucess for Olympus (as will
subsequent E models).

While I will follow the E-1 with interest, personally, I will
continue to enjoy my outdated E-10........ despite all it's short
coming. To my amazement the damn thing continues to produce great
pictures.

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[simona]

Hi Sam,

I think that you are right. There does seem to be visible lateral CA in Sean's shot. It is not too severe though, but claiming a complete lack of CA is not credible. The key question in my mind is how it performs relative to other cameras in this price backet with equivalent 'quality' lenses. From what I see, it does seem to deliver a significant improvement over my E10.

For me, this is a bigger issue than whether one cameras is marginally noisier at ISO800 than another.

Regards

Simon

But...why does everyone keep saying 'complete lack of CA' when it
is simply not true?

Like on Sean's shot here:
http://forums.dpreview.com/forums/read.asp?forum=1022&message=6313117

[The red line around the white] is not awful, but it looks a lot
like lateral CA to me. There have been other examples but I forget
which. Repeating marketing hype over and over does not make it true.
The colour moire is also a little odd too.

Sean

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Sam Berry]

I am not too sure whether I can find a specific reference, but I
will look into it. I guess, the truth is that there is more than
one source of CA. The lens can certainly be one, but achilles heel
of many digicams is chromatic aberation introduced by the sensor as
result of diffraction of non collumated light leading to bleeding
of the shorter wavelengths into adjacent pixels. My understanding
was that this is the source of the dreaded purple fringe in high
contrast situations. I will look into whether I can find a specific
reference for this.

Purple fringing is a phenomenon caused on both digital and film by IIRC longitudinal CA. The futher away from the centre of the spectrum the lens was designed for, the further the lens' focus shifts. So while you may have a specular highlight in focus in the bulk of the visible spectrum, the far blue/UV and IR may be significantly OOF. Film is normally partially sensitive to UV, and shows purple fringing as blue normally. CCDs, on the other hand are sensitive to a little UV and a lot of IR. A hot mirror is meant to block IR but it obviously is not perfect. Where IR starts is not concrete. An apochromatic lens mostly solves this problem, and most decent tele lenses are made this way. The problem is probably less of an issue with film largely because of better lenses. DSLRs rarely experience purple fringing, and never with decent lenses.

Many have said it is due to micro-lenses, but I have seen nothing to back this up.

I have two 50mm tessar SLR lenses, which are about as un-telecentric as you can get. I have no issues with either that aren't purely a fuction of the lens.

I have never heard of the refractive index of SiO2 being an issue. Or silicon for that matter.

A guy called Joseph Wisniewski who posts here knows a lot more about this than me, and had mady a few posts saying that different sensors affect the image due to corners being selective about which bits of the exit pupil it accepts light from, and that it may emphasize the pheripheral rays over the chief rays due to the reduced angle of incidence to some of the peripheral rays. Since the peripheral rays suffer more abberations this can adversely affect quality. Doesn't really apply to a cropped sensor much though.

From my perspective, the small noise penalty of the E1 at ISO800
and above when compared to the 10D etc is far outweighed by the
advantages.

Regards

Simon

Every camera has strenths and weaknesses. The E series cameras do a
superb job capturing natural colors. This is much more important to
many people than a little noise on the rare occasion one really
needs a high ISO setting.

I do not think the typical E-1 buyer will be put off over the noise
issue.The E-1 should be a commercial sucess for Olympus (as will
subsequent E models).

While I will follow the E-1 with interest, personally, I will
continue to enjoy my outdated E-10........ despite all it's short
coming. To my amazement the damn thing continues to produce great
pictures.

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[simona]

Hi Sam,

You may be right in asserting that the issue is largely due the performance of the lens, but my comments on the canon 1Ds partially came from Phil's review of the camera.

"Chromatic aberrations

Even using the best of Canon's L lenses we couldn't get away from the occasional chromatic aberration effect, that said it does require a particular high contrast situation (such as branches against a sky background) to be apparent. Also because of the 3:2 ratio of the image these artifacts are more likely to be visible on the left or right edges of the frame."

If you look at the sample images, purple fringing is very noticeable. Did you really mean to say that "DSLRs rarely experience purple fringing, and never with decent lenses"? or did you really mean to refer to film SLRs?

It is an interesting subject.

Thanks for your insightful comments

Regards

Simon

I am not too sure whether I can find a specific reference, but I
will look into it. I guess, the truth is that there is more than
one source of CA. The lens can certainly be one, but achilles heel
of many digicams is chromatic aberation introduced by the sensor as
result of diffraction of non collumated light leading to bleeding
of the shorter wavelengths into adjacent pixels. My understanding
was that this is the source of the dreaded purple fringe in high
contrast situations. I will look into whether I can find a specific
reference for this.

Purple fringing is a phenomenon caused on both digital and film by
IIRC longitudinal CA. The futher away from the centre of the
spectrum the lens was designed for, the further the lens' focus
shifts. So while you may have a specular highlight in focus in the
bulk of the visible spectrum, the far blue/UV and IR may be
significantly OOF. Film is normally partially sensitive to UV, and
shows purple fringing as blue normally. CCDs, on the other hand
are sensitive to a little UV and a lot of IR. A hot mirror is meant
to block IR but it obviously is not perfect. Where IR starts is not
concrete. An apochromatic lens mostly solves this problem, and most
decent tele lenses are made this way. The problem is probably less
of an issue with film largely because of better lenses. DSLRs
rarely experience purple fringing, and never with decent lenses.

Many have said it is due to micro-lenses, but I have seen nothing
to back this up.

I have two 50mm tessar SLR lenses, which are about as
un-telecentric as you can get. I have no issues with either that
aren't purely a fuction of the lens.

I have never heard of the refractive index of SiO2 being an issue.
Or silicon for that matter.

A guy called Joseph Wisniewski who posts here knows a lot more
about this than me, and had mady a few posts saying that different
sensors affect the image due to corners being selective about which
bits of the exit pupil it accepts light from, and that it may
emphasize the pheripheral rays over the chief rays due to the
reduced angle of incidence to some of the peripheral rays. Since
the peripheral rays suffer more abberations this can adversely
affect quality. Doesn't really apply to a cropped sensor much
though.

 

[Sam Berry]

I agree, although for me personally the noise issue is huge. I shot about 150 shots in a nightclub two nights ago, about 5% ISO 400, 30% ISO 800 and the rest mostly 1600. A few 3200s. Most are RAW and will need pushing up to a stop in post. Making colour images out of ISO 4000-8000 shots is something I need to do. Every shot was at f/1.8 or f/2 and 1/20s-1/80s handheld. I am in a minority though, I know.

Anyway, I expect fringing in non-pro lenses either wide-open on very-bright highlights or lat-CA very stopped down on edges far away from the focus point. Expecting CA to be fully corrected for all distances when focussing hyperfocal is too much to expect IMHE of a non-pro lens. What is worrying about this shot is that it does not fall into either catagory. Sean seems both experienced and fairly unbiased, so if he says that the lenses are Canon L quality then I tend to believe him, but I have not seen anything with my own eyes to back that up. I have not seen any many wide-open shots yet, just lots of f/5 or so ones.

I am prepared to believe they are, just havn't seen it yet. For the money they seem pretty good considering the range in any case. And telecentricity has nothing to do with it.

For me, this is a bigger issue than whether one cameras is
marginally noisier at ISO800 than another.

Regards

Simon

But...why does everyone keep saying 'complete lack of CA' when it
is simply not true?

Like on Sean's shot here:
http://forums.dpreview.com/forums/read.asp?forum=1022&message=6313117

[The red line around the white] is not awful, but it looks a lot
like lateral CA to me. There have been other examples but I forget
which. Repeating marketing hype over and over does not make it true.
The colour moire is also a little odd too.

Sean

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Sam Berry]

Yes, I noticed that in the review too. You bring up a good point.

In this particular case however:

The 1Ds is full-frame. Any issues which are not serious here will be inconsequential on a cropped sensor. The incidence angles are significantly more normal. But the big thing here is that the 17-35mm lens used is pretty poor at its wide end (particularly in the periphery), just like the 20-35mm before it and the 16-35mm after it. Canon, with the exception of the newish 17-40 are not good with wide-angles. The use of an ultra-wide partially invalidates the test. Due to the mirror, nothing can get very un-telecentric in an SLR. A 50mm lens doesn't have a much steeper angle of incidence than than UWA. It is also faster=bigger exit pupil=larger angle. An UWA does suffer from lots of physics which cause it to be optically poor compared to a longer lens. Nikon offering is a world better by all accounts. Although Phil states 'the best of Canon's L lenses' he is probably using the second-worst L lens. The first being the 14mm. 17mm on film rarely looks doesn't usually look so hot in the corners either, in general. And film has got rangefinder cameras, which help enormously.

Oh, and the trees are almost certainly significantly in-front of the focus point.

The 1Ds may experience problems which cannot be attributed to the lens, but this is not the way to test it.

"Chromatic aberrations
Even using the best of Canon's L lenses we couldn't get away from
the occasional chromatic aberration effect, that said it does
require a particular high contrast situation (such as branches
against a sky background) to be apparent. Also because of the 3:2
ratio of the image these artifacts are more likely to be visible on
the left or right edges of the frame."

If you look at the sample images, purple fringing is very
noticeable. Did you really mean to say that "DSLRs rarely
experience purple fringing, and never with decent lenses"? or did
you really mean to refer to film SLRs?

It is an interesting subject.

Thanks for your insightful comments

Regards

Simon

I am not too sure whether I can find a specific reference, but I
will look into it. I guess, the truth is that there is more than
one source of CA. The lens can certainly be one, but achilles heel
of many digicams is chromatic aberation introduced by the sensor as
result of diffraction of non collumated light leading to bleeding
of the shorter wavelengths into adjacent pixels. My understanding
was that this is the source of the dreaded purple fringe in high
contrast situations. I will look into whether I can find a specific
reference for this.

Purple fringing is a phenomenon caused on both digital and film by
IIRC longitudinal CA. The futher away from the centre of the
spectrum the lens was designed for, the further the lens' focus
shifts. So while you may have a specular highlight in focus in the
bulk of the visible spectrum, the far blue/UV and IR may be
significantly OOF. Film is normally partially sensitive to UV, and
shows purple fringing as blue normally. CCDs, on the other hand
are sensitive to a little UV and a lot of IR. A hot mirror is meant
to block IR but it obviously is not perfect. Where IR starts is not
concrete. An apochromatic lens mostly solves this problem, and most
decent tele lenses are made this way. The problem is probably less
of an issue with film largely because of better lenses. DSLRs
rarely experience purple fringing, and never with decent lenses.

Many have said it is due to micro-lenses, but I have seen nothing
to back this up.

I have two 50mm tessar SLR lenses, which are about as
un-telecentric as you can get. I have no issues with either that
aren't purely a fuction of the lens.

I have never heard of the refractive index of SiO2 being an issue.
Or silicon for that matter.

A guy called Joseph Wisniewski who posts here knows a lot more
about this than me, and had mady a few posts saying that different
sensors affect the image due to corners being selective about which
bits of the exit pupil it accepts light from, and that it may
emphasize the pheripheral rays over the chief rays due to the
reduced angle of incidence to some of the peripheral rays. Since
the peripheral rays suffer more abberations this can adversely
affect quality. Doesn't really apply to a cropped sensor much
though.

 

[simona]

Hi agian Sam,

Regarding noise, it is all about what you need the camera for. Given your usage, I can see that high ISO noise IS a big issue. I do quite a bit of mountaineering, so most of what I do is landscape photography. I have recently become interested in motorsport photogarphy and here the limitations of my E10 (speed & battery life in particular) are been severely exposed. Up to ISO800 is useful to get a decent shutter speed in poor lighting conditions, but it is not too much of an issue. With my E10, I rarely used it above ISO80 becuase the noise was a real killer.

Regards

Simon

http://www.photosig.com/go/users/userphotos?id=92014

Anyway, I expect fringing in non-pro lenses either wide-open on
very-bright highlights or lat-CA very stopped down on edges far
away from the focus point. Expecting CA to be fully corrected for
all distances when focussing hyperfocal is too much to expect IMHE
of a non-pro lens. What is worrying about this shot is that it does
not fall into either catagory. Sean seems both experienced and
fairly unbiased, so if he says that the lenses are Canon L quality
then I tend to believe him, but I have not seen anything with my
own eyes to back that up. I have not seen any many wide-open shots
yet, just lots of f/5 or so ones.

I am prepared to believe they are, just havn't seen it yet. For the
money they seem pretty good considering the range in any case. And
telecentricity has nothing to do with it.

For me, this is a bigger issue than whether one cameras is
marginally noisier at ISO800 than another.

Regards

Simon

But...why does everyone keep saying 'complete lack of CA' when it
is simply not true?

Like on Sean's shot here:
http://forums.dpreview.com/forums/read.asp?forum=1022&message=6313117

[The red line around the white] is not awful, but it looks a lot
like lateral CA to me. There have been other examples but I forget
which. Repeating marketing hype over and over does not make it true.
The colour moire is also a little odd too.

Sean

Sean Reid mentioned that he
found the E-1's color saturation level to be somewhat higher than
the 10D's.

So far no one seems to be complaining about 300D noice levels.
Could this be due to other factors such as chrominance noise?

I also wonder what comparative noise graphs would look like if the
E-1, 10D, D100 and S2 were equalized for sharpness, saturation and
contrast, rather than simply tested at the factory default settings.

No offense, but why is everyone so obsessed with noise? Are you all
professional grey-patch shooters working in lab-setups? If so, then
all
this discussion is probably useful, but I rather prefer to look at the
real-world shots that are shown, because that's probably what I would
use the camera for. If I look at the shots shown from the E-1, I see
great quality 5mpx shots, with smoots skies, great colours, and a wide
dynamic range (lots of shadow detail and good highlight handling). Oh,
and did I mention a complete lack of CA? See, you don't need noise
graphs and a fourier analisys to figure that out..

I have never had a problem with the noise in my E-10 shots. I see
it is
there ofcourse, but it doesn't get in the way when I print. The E-1 is
vastly superior to the E-10 when it comes to noise. So I know that it
will be great wrt noise. And how often are you planning to use iso
1600
in the first place?

But that's just my 2 eurocents..

Bram

--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Bram Bos]

Hi

useful to get a decent shutter speed in poor lighting conditions,
but it is not too much of an issue. With my E10, I rarely used it
above ISO80 becuase the noise was a real killer.

Mine is permanently locked on ISO80, and while a higher ISO would
be convenient a lot can be done with low ISOs when you have a
steady hand. For example, I've done a lot of shots like the one below,
all handheld, all ISO 80.. This is an interior detail of a Buddhist monastery
in Ladakh (with an E-10 by the way):



--------------------------------------------------------------------
My Travel Galleries (asia, middle east, latin america)
http://www.pbase.com/brambos

 

[Sam Berry]

You've got a nice gallery there, from what I can tell from the thumbnails The mountain shots are paricularly nice. As a regular visitor to Switzerland and can appreciate the alps! I have no issues with the E-1 low-ISO noise. ISO 100 is noiseless for all intents and purposes that I can think of.

Oh, and by the way, I don't use Canon's normal output method. I convert to Tiff w/o sharpening (canon performs selective sharpening to reduce noise normally), sharpen using a convoution kernel and then run though neat-image. You get much nicer textures this way.

Oh, were the racing shots with a tele-converter or just your Oly?