D300 versus 20D continued analysis

Started Dec 5, 2007 | Discussions
Starbuck22 Forum Member • Posts: 77
Re: D300 versus 20D continued analysis

interesting but old canon produce clearly less noisy and sharper images ...

Snappie Pants Regular Member • Posts: 240
Re: let us see the RAW files

k - I hope you don't mind. I ran the file through a simple NR step in Capture.

Adjusted the Noise Reduction properties for the NEF and then saved it as 100.

Base Adjustments ~ Detail Adjustments ~ Noise Reduction

Set the Intensity to 29% and the Sharpness to 9.

Choose "Better Quality" method.

Also check the box for Edge Noise Reduction.

mike7588 Regular Member • Posts: 250
Re: D300 versus 20D continued analysis

betterliving wrote:

It is very difficult for a 12mp sensor to compete with a smaller
sensor, such as the 8mp. Even dpreview showed that the 40d had a

There is a superior 12MP in this price range, and yet it's a full frame pro.

Have you seen 5D comparison?
The 12MP was just superb.

http://kenrockwell.com/tech/iso-comparisons/2007-11/index.htm
http://forums.dpreview.com/forums/read.asp?forum=1039&message=25847543

OP Arash Hazeghi Contributing Member • Posts: 967
Re: let us see the RAW files

You are welcomed to play with the RAW file, however I was running a comparisson not trying to get the best possible picture out of the D300 with NR, then I could equally clean up 20D image and have even a cleaner file. Of course with NR things change and that is what people are seeing,

Now regarding NR, I would actually leave it off at capture and try neat image, try this and you get better results and it runs faster than capture too. also NR faster versus high quality didn't seem to make a difference so just use faster. not sure what edge NR actually does?

Arash

Snappie Pants wrote:

k - I hope you don't mind. I ran the file through a simple NR step
in Capture.

Adjusted the Noise Reduction properties for the NEF and then saved it
as 100.

Base Adjustments ~ Detail Adjustments ~ Noise Reduction

Set the Intensity to 29% and the Sharpness to 9.

Choose "Better Quality" method.

Also check the box for Edge Noise Reduction.

Qwntm Veteran Member • Posts: 6,165
Lets start again, OK?

If you would please give me a step by step procedure for how you think a 100% crop should be made in CS2 I would appreciate it, and do my best to comply.

And yes, as I stated, the in camera noise reduction was set for "ON normal." It was a snapshot, and there is nothing I can do about it now. I read your message as if you were accusing me of applying some kind of PP noise reduction, which I did not. Sorry for any misunderstanding.

 Qwntm's gear list:Qwntm's gear list
Fujifilm X100S Pentax K-5 IIs Olympus E-M5 II Sigma 10mm F2.8 EX DC HSM Diagonal Fisheye Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) +8 more
Devendra
Devendra Veteran Member • Posts: 3,584
you seem to be a new expert here ..

BenBotha wrote:

Devendra wrote:

we all know canon's digic does internal NR, and with Nikon it has
been known that it leaves it to the photographer.

now try this with NC for high ISO shots.

go the base adjustments-> raw adjustments-> color moire reduction-> and
select moire to low- and then show us the results

try to learn on how to use the system rather than draw conclusions
from pixel peeping out of the box results. you may not have the
flexibility with canons, but nikon provides you a lot.. to play with.

btw, try printing them in 20"x30" sizes or more too and see what
holds - your 8mp vs 12mp arguments will sounds pretty bad.

This is a typical response of somebody not being ably to look
objectively at a different viewpoint and even worse, it is inaccurate!

no samples.. no comments..

 Devendra's gear list:Devendra's gear list
Sony Alpha NEX-3N Nikon D800E Nikon D200 Nikon 1 J1
Snappie Pants Regular Member • Posts: 240
so the canon had absolutely no NR applied?

No NR on the canon? Nor in the canon raw converter thing? I ask because the canon looks a bit plastic.

Further, the wall in the canon looks properly exposed. The wall in the Nikon photo looks a bit dark.

Qwntm Veteran Member • Posts: 6,165
Phil's review will tell all...

betterliving wrote:

Have you noticed how the real zealous Canon folks are fine with the
test and do not even question it's validity? Yet, when we question
it, we are labeled Fanboys. And, these are bloggers I have NEVER seen
before! Where did they come from?

This, despite the fact that three published reviews have given the
D300 top billing in noise and image quality! Amature Photographer,
What Digital Camera, and the main German mag, whose name escapes me.
As the reviews roll in, the D300 has done very well. Certianly much
better than these home photos.

Yet, a blogger comes along and posts a few odd photos in the Nikon
forum and the Canon loyalists ignore all the publications and praise
the blogger. I guess we believe what we want to believe.

You have to know that this is like chasing a ghost. At some point,
despite your hopes, you're going to hit a wall.

This will come to screeching halt when and if Phil gets his review up before Christmas ------- 2009!

 Qwntm's gear list:Qwntm's gear list
Fujifilm X100S Pentax K-5 IIs Olympus E-M5 II Sigma 10mm F2.8 EX DC HSM Diagonal Fisheye Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) +8 more
OP Arash Hazeghi Contributing Member • Posts: 967
Re: so the canon had absolutely no NR applied?

It looks plastic because there is little noise, but it is not the water-color looking from NR, usually NR causes images to look like water color painting...

exposure is the same for both, canon sensor is a bit more sensitive than Nikon, (in 40D the sensitivity is turned down to be standard like Nikon) but then leaves and everything in the highlights is blown up in the canon shot, Nikon exposure is accurate, it kept all of my highlights, if you look at the histogram you will find out that nikon shot is by all means well balances in such a high DR scene.

Snappie Pants wrote:

No NR on the canon? Nor in the canon raw converter thing? I ask
because the canon looks a bit plastic.

Further, the wall in the canon looks properly exposed. The wall in
the Nikon photo looks a bit dark.

cayzi Contributing Member • Posts: 559
Re: Very nice, thanks so much for the time put in.

Nikon have option to turn off NR if you want.

The best option on Nikon D300 is to set NR to LOW even on NEF and jpeg.

Try this and then run test again.

OP Arash Hazeghi Contributing Member • Posts: 967
Re: you seem to be a new expert here ..

I stated several times, moire setting=off for all shots. camera's AA filter is not a SW thing that you can turn off or make weaker by this setting.

nico3d
nico3d Contributing Member • Posts: 878
Re: you seem to be a new expert here ..

wow this thread is wild !! hahaahah

i m absolutely happy with my D300. i dont really care at all about your tests maybe they are right, maybe they are wrong, maybe you re biased maybe you are not. They dont give me any insecurity about my D300. this camera is by far the best, FOR ME, MYSELF, and it works perfectly to ME.

be happy both brands are awesome, i just went for one of them (dont have money for both)

 nico3d's gear list:nico3d's gear list
Nikon AF-S Nikkor 24-70mm f/2.8G ED
Snappie Pants Regular Member • Posts: 240
can canon apply NR in camera?

if not, then it looks like you have no control over the NR preferences in your pictures. The canon pictures do look like NR was run on them at some level - it seems almost that every canon shot gets NR whether the user wants it.

The canon picture is too "perfect" in terms of noise - I'm wondering what's being left out of the equation.

mike7588 Regular Member • Posts: 250
Re: Lets start again, OK?

One possibility is your original-non-100% (the second picture) was already enlarged and cropped, so that the 100% didn't look too large compared to the original-non-100%.

If this is the case just send the full unresized uncrop image 100%.
If it lost, just make another one.
Not a big problem.

But one thing for sure,
all these unbiased experiences are very helpfull for us.
Including the benefit of AF tracking etc etc.

Qwntm wrote:

If you would please give me a step by step procedure for how you
think a 100% crop should be made in CS2 I would appreciate it, and do
my best to comply.

And yes, as I stated, the in camera noise reduction was set for "ON
normal." It was a snapshot, and there is nothing I can do about it
now. I read your message as if you were accusing me of applying some
kind of PP noise reduction, which I did not. Sorry for any
misunderstanding.

subdoodle Senior Member • Posts: 1,310
So You Deliberately Underexposed Nikon?

Is that it?

subdoodle Senior Member • Posts: 1,310
Canon had NR / Nikon was underexposed

apparently. (20d has always on NR in the sensor)

OP Arash Hazeghi Contributing Member • Posts: 967
Re: can canon apply NR in camera?

I can tell you Canon has no in-camera SW NR once the sensor data is recorded. before then, it is a black box for all cameras and any NR is considered hardware NR. Actually, it seems that it is this sensor level NR that Canon has been doing so well in the past 7 years. This is not my research field but one of my friends (who now works for foveon) once told me that Canon implementes a circuit-based noise reduction before they precharge the bit lines. once the charge is shared between well cap and the bit line cap cross talk between adjacent lines cause disturbance and noise which will go directly into the gain amplifier, they use a trick at that level to subtract noise, no body knows exactly what it is except for themselves. All cameras have different on-chip hardware NR, so it is not possible to change that. Neverthless the main concern is SW NR, i.e noise reduction routines that are applied to RAW sensor data. We are talking about whatever sesnor delivers, and I believe Canon data is cleaner to begin with. Honestly even if you turn NR ON on D300, you might be able to clean it up but it will not have the detail the 20D has.

Snappie Pants wrote:

if not, then it looks like you have no control over the NR
preferences in your pictures. The canon pictures do look like NR was
run on them at some level - it seems almost that every canon shot
gets NR whether the user wants it.

The canon picture is too "perfect" in terms of noise - I'm wondering
what's being left out of the equation.

OP Arash Hazeghi Contributing Member • Posts: 967
Re: Canon had NR / Nikon was underexposed

So does D300, and it has LOTS of it, at the gain amplifier, Nikon is exposed perefectly. D300 meter used for all shots, call up histogram and evaluate exposure.

subdoodle wrote:

apparently. (20d has always on NR in the sensor)

OP Arash Hazeghi Contributing Member • Posts: 967
Re: So You Deliberately Underexposed Nikon?

this is a stupid comment

subdoodle wrote:

Is that it?

snooked123 Regular Member • Posts: 344
Re: D300 versus 20D continued analysis

Canon 20d applies NR to jpegs and you cannot turn it off. If d300's NR is off, the test is not accurate. Having said that 20d's RAW with proper exposure at iso1600 can be used for pretty large size prints without applying any sort of noise removal (except for chroma reduction applied by the raw convertor).

Additionally, it is a bit unfair to compare an 8MP sensor to a 12MP one. The 12MP one is bound to be noisier. IMHO if NR off with d300 means NR completely off, d300 is doing pretty good in the noise department.

Keyboard shortcuts:
FForum MMy threads