Canon's 50mp DSLR. Higher res than D810, but more moire andnoise
 

On 8/05/2015 5:49 p.m., android wrote:
In article ,
RichA wrote:

IMO, the resolution is noticeable but not huge because the linear resolution
increase isn't that great. It sounds like a lot when you say, "50mp compared
to 36" but that's not how resolution works. If you want to double
resolution, you need to quadruple pixel count. It does show more moire than
the D810, I don't know why. Noise is perhaps 1/3 to 1/2 stop worse than the
Nikon. Because file sizes are dependent on area, unlike resolution, file
sizes are considerably larger.

I don't think that the moire is "worse", just different because of
sampling frequency on the patterns on the test target.
Bitching about whether an AA filter is still needed or not (even at only
36mp) should be made after adding the D800 to the comparison tool. Moire
appears in the D800 image in exactly the same place as with the D800E
and D810, it's certainly not eliminated by the AA filter, only reduced.
I'd expect to see the same difference between the two 5DS models when
DPReview add the 5Ds to the compare tool.


http://www.dpreview.com/reviews/imag...&attr13_0=cano
n_eos5dsr&attr13_1=nikon_d810&attr13_2=pentax_645z &attr13_3=phaseone_iq180&att
r15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&a ttr16_0=200&attr16_1=200&att
r16_2=200&attr16_3=35&normalization=full&widget=1& x=-0.5172853599968508&y=0.26
069086651053863

If you check the noise in DxO you'll se that high ISO noise is lover in
1Dx than in the D810. I also downloaded the ISO 6400 files for the 5Dsr
and 1Dx from Imaging Resource and found that noise vise, at the same
imagesize they there not to tell apart...

Click on the measurements tab:

http://tinyurl.com/lng7t2c

http://www.dxomark.com/Cameras/Compa...810-versus-Nik
on-D4-versus-Canon-EOS-1Dx___963_767_753

Really?
That link to DXO shows very little difference at all in high ISO noise
between any of those 3 cameras. The "score" - Sports (Low Light ISO)
are 2786, 2853, and 2956. If you were able to set your camera to ISO
2786 and 2956 and compare the result (image noise) do you think you'd be
able to pick the difference? Even if you could, then what practical
difference would it make anyway? File size might make a practical
difference, so might resolution.

Canon's 50mp DSLR. Higher res than D810, but more moire and noise
 

In article , Me
wrote:

On 8/05/2015 5:49 p.m., android wrote:
In article ,
RichA wrote:

IMO, the resolution is noticeable but not huge because the linear
resolution
increase isn't that great. It sounds like a lot when you say, "50mp
compared
to 36" but that's not how resolution works. If you want to double
resolution, you need to quadruple pixel count. It does show more moire
than
the D810, I don't know why. Noise is perhaps 1/3 to 1/2 stop worse than
the
Nikon. Because file sizes are dependent on area, unlike resolution, file
sizes are considerably larger.

I don't think that the moire is "worse", just different because of
sampling frequency on the patterns on the test target.
Bitching about whether an AA filter is still needed or not (even at only
36mp) should be made after adding the D800 to the comparison tool. Moire
appears in the D800 image in exactly the same place as with the D800E
and D810, it's certainly not eliminated by the AA filter, only reduced.
I'd expect to see the same difference between the two 5DS models when
DPReview add the 5Ds to the compare tool.


http://www.dpreview.com/reviews/imag...ght&attr13_0=c
ano
n_eos5dsr&attr13_1=nikon_d810&attr13_2=pentax_645z &attr13_3=phaseone_iq180&
att
r15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&a ttr16_0=200&attr16_1=200&
att
r16_2=200&attr16_3=35&normalization=full&widget=1& x=-0.5172853599968508&y=0
.26
069086651053863

If you check the noise in DxO you'll se that high ISO noise is lover in
1Dx than in the D810. I also downloaded the ISO 6400 files for the 5Dsr
and 1Dx from Imaging Resource and found that noise vise, at the same
imagesize they there not to tell apart...

Click on the measurements tab:

http://tinyurl.com/lng7t2c

http://www.dxomark.com/Cameras/Compa...810-versus-Nik
on-D4-versus-Canon-EOS-1Dx___963_767_753

Really?
That link to DXO shows very little difference at all in high ISO noise
between any of those 3 cameras.

That's the point. Evaluation of early 5Dsr files indicates that it will
produce better pictures at ISO 6400 than the 1Dx and D4. At the same
image seize.

The "score" - Sports (Low Light ISO)
are 2786, 2853, and 2956. If you were able to set your camera to ISO
2786 and 2956 and compare the result (image noise) do you think you'd be
able to pick the difference?

I don't care about DxO "scores". You want scores? Go to senscore.org!
scores without supported and structured data is pointless. I care about
the DxO measurements and makes my own conclusions.

Even if you could, then what practical
difference would it make anyway? File size might make a practical
difference, so might resolution.

Se the above...
--
teleportation kills

Canon's 50mp DSLR. Higher res than D810, but more moire andnoise
 

On 10/05/2015 6:21 p.m., android wrote:
In article , Me
wrote:

On 8/05/2015 5:49 p.m., android wrote:
In article ,
RichA wrote:

IMO, the resolution is noticeable but not huge because the linear
resolution
increase isn't that great. It sounds like a lot when you say, "50mp
compared
to 36" but that's not how resolution works. If you want to double
resolution, you need to quadruple pixel count. It does show more moire
than
the D810, I don't know why. Noise is perhaps 1/3 to 1/2 stop worse than
the
Nikon. Because file sizes are dependent on area, unlike resolution, file
sizes are considerably larger.

I don't think that the moire is "worse", just different because of
sampling frequency on the patterns on the test target.
Bitching about whether an AA filter is still needed or not (even at only
36mp) should be made after adding the D800 to the comparison tool. Moire
appears in the D800 image in exactly the same place as with the D800E
and D810, it's certainly not eliminated by the AA filter, only reduced.
I'd expect to see the same difference between the two 5DS models when
DPReview add the 5Ds to the compare tool.


http://www.dpreview.com/reviews/imag...ght&attr13_0=c
ano
n_eos5dsr&attr13_1=nikon_d810&attr13_2=pentax_645z &attr13_3=phaseone_iq180&
att
r15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&a ttr16_0=200&attr16_1=200&
att
r16_2=200&attr16_3=35&normalization=full&widget=1& x=-0.5172853599968508&y=0
.26
069086651053863

If you check the noise in DxO you'll se that high ISO noise is lover in
1Dx than in the D810. I also downloaded the ISO 6400 files for the 5Dsr
and 1Dx from Imaging Resource and found that noise vise, at the same
imagesize they there not to tell apart...

Click on the measurements tab:

http://tinyurl.com/lng7t2c

http://www.dxomark.com/Cameras/Compa...810-versus-Nik
on-D4-versus-Canon-EOS-1Dx___963_767_753

Really?
That link to DXO shows very little difference at all in high ISO noise
between any of those 3 cameras.

That's the point. Evaluation of early 5Dsr files indicates that it will
produce better pictures at ISO 6400 than the 1Dx and D4. At the same
image seize.

The "score" - Sports (Low Light ISO)
are 2786, 2853, and 2956. If you were able to set your camera to ISO
2786 and 2956 and compare the result (image noise) do you think you'd be
able to pick the difference?

I don't care about DxO "scores". You want scores? Go to senscore.org!
scores without supported and structured data is pointless. I care about
the DxO measurements and makes my own conclusions.

Even if you could, then what practical
difference would it make anyway? File size might make a practical
difference, so might resolution.

Se the above...

I don't think we are arguing about anything.
Raw data for a selection of current/recent full frame cameras here,
based on a standard output size, charted as dynamic range at a standard
signal:noise ratio: http://i.imgur.com/BDKPNu2.png
Above ISO 1600 and above, they're basically all the same. No surprise
there, they all have similar quantum efficiency - so photon noise is
about the same.
Below ISO 1600, then things change.
Wavy lines are for some cameras which apply gain after sensor readout
for intermediate ISO settings.
I don't think there's any "spin" applied to that data - as used in the
weighted scores by DXO.
Senscore is a bit nutty. Resolution is weighted to contribute 20% of
the final score - so a sensor with only 1 pixel could score quite well -
yet the "other half" of their site - "lenscore" is obsessed with
resolution performance of lenses.

Canon's 50mp DSLR. Higher res than D810, but more moire andnoise
 

On 9/05/2015 6:15 p.m., RichA wrote:


The lowest noise camera ever was the Nikon D3s. The D4 and NO Canon ever matched it.

Not correct on both counts.

Canon's 50mp DSLR. Higher res than D810, but more moire and noise
 

In article , Me
wrote:

On 10/05/2015 6:21 p.m., android wrote:
In article , Me
wrote:

On 8/05/2015 5:49 p.m., android wrote:
In article ,
RichA wrote:

IMO, the resolution is noticeable but not huge because the linear
resolution
increase isn't that great. It sounds like a lot when you say, "50mp
compared
to 36" but that's not how resolution works. If you want to double
resolution, you need to quadruple pixel count. It does show more moire
than
the D810, I don't know why. Noise is perhaps 1/3 to 1/2 stop worse than
the
Nikon. Because file sizes are dependent on area, unlike resolution,
file
sizes are considerably larger.

I don't think that the moire is "worse", just different because of
sampling frequency on the patterns on the test target.
Bitching about whether an AA filter is still needed or not (even at only
36mp) should be made after adding the D800 to the comparison tool. Moire
appears in the D800 image in exactly the same place as with the D800E
and D810, it's certainly not eliminated by the AA filter, only reduced.
I'd expect to see the same difference between the two 5DS models when
DPReview add the 5Ds to the compare tool.


http://www.dpreview.com/reviews/imag...light&attr13_0
=c
ano
n_eos5dsr&attr13_1=nikon_d810&attr13_2=pentax_645z &attr13_3=phaseone_iq18
0&
att
r15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&a ttr16_0=200&attr16_1=20
0&
att
r16_2=200&attr16_3=35&normalization=full&widget=1& x=-0.5172853599968508&y
=0
.26
069086651053863

If you check the noise in DxO you'll se that high ISO noise is lover in
1Dx than in the D810. I also downloaded the ISO 6400 files for the 5Dsr
and 1Dx from Imaging Resource and found that noise vise, at the same
imagesize they there not to tell apart...

Click on the measurements tab:

http://tinyurl.com/lng7t2c

http://www.dxomark.com/Cameras/Compa...810-versus-Nik
on-D4-versus-Canon-EOS-1Dx___963_767_753

Really?
That link to DXO shows very little difference at all in high ISO noise
between any of those 3 cameras.

That's the point. Evaluation of early 5Dsr files indicates that it will
produce better pictures at ISO 6400 than the 1Dx and D4. At the same
image seize.

The "score" - Sports (Low Light ISO)
are 2786, 2853, and 2956. If you were able to set your camera to ISO
2786 and 2956 and compare the result (image noise) do you think you'd be
able to pick the difference?

I don't care about DxO "scores". You want scores? Go to senscore.org!
scores without supported and structured data is pointless. I care about
the DxO measurements and makes my own conclusions.

Even if you could, then what practical
difference would it make anyway? File size might make a practical
difference, so might resolution.

Se the above...

I don't think we are arguing about anything.
Raw data for a selection of current/recent full frame cameras here,
based on a standard output size, charted as dynamic range at a standard
signal:noise ratio: http://i.imgur.com/BDKPNu2.png
Above ISO 1600 and above, they're basically all the same. No surprise
there, they all have similar quantum efficiency - so photon noise is
about the same.
Below ISO 1600, then things change.
Wavy lines are for some cameras which apply gain after sensor readout
for intermediate ISO settings.

I think that your conclusion that some has analog interference between
the sensor and is reasonable and correct. I also think that the
"wavyness" are to be attributed to a cruder correction than the less
wavy. That wavyness could, seriously lead to "analog" posterisation.

I don't know the source for your data or which cameras that are
measured, the above comment was made in good faith.

I don't think there's any "spin" applied to that data - as used in the
weighted scores by DXO.

I think that there is a choice on the part of DxO on how to weight data
that makes their data more interesting to me than their score.

Senscore is a bit nutty.

You don't say!

Resolution is weighted to contribute 20% of
the final score - so a sensor with only 1 pixel could score quite well -
yet the "other half" of their site - "lenscore" is obsessed with
resolution performance of lenses.
--
teleportation kills

Canon's 50mp DSLR. Higher res than D810, but more moire andnoise
 

On 11/05/2015 6:41 p.m., RichA wrote:
On Sunday, 10 May 2015 08:04:01 UTC-4, Me wrote:
On 9/05/2015 6:15 p.m., RichA wrote:


The lowest noise camera ever was the Nikon D3s. The D4 and NO Canon ever matched it.

Not correct on both counts.

Which Canon matched it? Please, no "take a 22mp image and downsize it" stuff, did any 12-16mp Canon ever match it?

1Dx matches it - very closely.
Nikon's D6*0/750/8*0 exceed it for low ISO dynamic range.