D300 versus 20D continued analysis

Started Dec 5, 2007 | Discussions
OP Arash Hazeghi Contributing Member • Posts: 967
Re: D300 versus 20D continued analysis

This test is RAW. not JPEG, for both cameras.

snooked123 wrote:

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.

Qwntm Veteran Member • Posts: 6,165
No response from Arash...

Guess he's not as helpful as he expects the rest of us to be...

Here's a 100% crop following this procedure:

http://www.pbase.com/wlhuber/100_crop_

I chose 1280x1024, because thats what my monitor is set at.

 Qwntm's gear list:Qwntm's gear list
Fujifilm X100S Pentax K-5 IIs Olympus E-M5 II Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) +8 more
Qwntm Veteran Member • Posts: 6,165
Re: D300 versus 20D continued analysis

Arash Hazeghi wrote:

This test is RAW. not JPEG, for both cameras.

Yea, but if you converted in DPP or EOS utility viewer, the NR was applied by Canon's software, whether you like it or not.

You are not even close to making any kind of reasonable tests here...

 Qwntm's gear list:Qwntm's gear list
Fujifilm X100S Pentax K-5 IIs Olympus E-M5 II Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) +8 more
OP Arash Hazeghi Contributing Member • Posts: 967
Re: D300 versus 20D continued analysis

ah you again, please, I don't appriciate you posting on my thread.

Qwntm wrote:

Arash Hazeghi wrote:

This test is RAW. not JPEG, for both cameras.

Yea, but if you converted in DPP or EOS utility viewer, the NR was
applied by Canon's software, whether you like it or not.

You are not even close to making any kind of reasonable tests here...

Qwntm Veteran Member • Posts: 6,165
Re: So You Deliberately Underexposed Nikon?

Arash Hazeghi wrote:
this is a stupid comment

subdoodle wrote:

Is that it?

Not really when you consider the D300 meter tends to overexpose a bit. Kinda interesting that you have the only D300 in existence that underexposes...

 Qwntm's gear list:Qwntm's gear list
Fujifilm X100S Pentax K-5 IIs Olympus E-M5 II Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) +8 more
attilacs Regular Member • Posts: 169
Re: D300 versus 20D continued analysis

D300 has more noise for sure.
Sharpness and detail is also higher on the D300 IMHO.

It is a trade-off, you know what you want, make your decision.

-- hide signature --

aTTilacS

D70s; Tokina 12-24 F/4; Nikkor 35 F/2; 50 F/1.4; 60 F/2.8 Micro; 85 F/1.4; 55-200 DX

Some of the keepers:
http://www.flickr.com/photos/attilacs/

 attilacs's gear list:attilacs's gear list
Ricoh GR Nikon D3400 Nikon D850 Nikon AF-S DX Nikkor 35mm F1.8G Nikon AF Nikkor 85mm f/1.4D +9 more
snooked123 Regular Member • Posts: 344
Re: D300 versus 20D continued analysis

Since I am not a nikon user, I cannot comment on the validity of the above mentioned claim but I have heard lots of people say that Nikon's RAW should be converted with Nikon's raw convertor as it has custom noise removal algorithms which are very very effective in removing noise that cannot be removed by other raw convertors.

Qwntm Veteran Member • Posts: 6,165
Re: D300 versus 20D continued analysis

Arash Hazeghi wrote:

ah you again, please, I don't appriciate you posting on my thread.

So I comment that Canon's software adds the noise reduction for you in the conversion, and this is your reply?

You selectively respond just as you selectively choose your parameters for your tests...

I tried to make amends, and have posted a new 100% crop after looking up an accepted procedure myself, no thanks to you.

The faster your threads get to 150 the better for all of us.

 Qwntm's gear list:Qwntm's gear list
Fujifilm X100S Pentax K-5 IIs Olympus E-M5 II Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) +8 more
Qwntm Veteran Member • Posts: 6,165
Re: D300 versus 20D continued analysis

snooked123 wrote:

Since I am not a nikon user, I cannot comment on the validity of the
above mentioned claim but I have heard lots of people say that
Nikon's RAW should be converted with Nikon's raw convertor as it has
custom noise removal algorithms which are very very effective in
removing noise that cannot be removed by other raw convertors.

I'm not sure about the noise removal, but it is generally accepted that the Nikon software, like Capture, just does an overall superior job in conversion. This is particularly true regarding the white balance and colors.

 Qwntm's gear list:Qwntm's gear list
Fujifilm X100S Pentax K-5 IIs Olympus E-M5 II Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) +8 more
jp Senior Member • Posts: 1,151
Arash: I do not agree on "per pixel sharpness" at all

I do not agree with your "per pixel sharpness" argument at all

Look for example at the details on the rainpipe: the mounting screws and the bending folds are almost invisible on the 20S pctures.

I do agree with the noise difference, that was also my observation.

But better hve some noise in the shadows than blown highlights of important image parts.

If you expose for the same brightness of your pictures, taking into account that the ISO values of the D300 are around 1/3 stop too optimistic , you will see that the noise advantage of the 40D becomes smalller.

If you would expose both pictures for the highlights then the overal picture of the D300 would be brighter than that on the 20D however, but noise would again become less and on par between both cameras after resizing.

I agree that nikon does not beat canon in noise, the more ore less closed the gap but they did not take the lead.

That said, noise is one thing but as you said yourself, there is more to it than only noise to evaluate the quality of a DSLR.

BTW: did you try the 3D tracking with CF-A1 set to focus priority?

-- hide signature --

cameras don't shoot people
People shoot people.

OP Arash Hazeghi Contributing Member • Posts: 967
Re: Arash: I do not agree on "per pixel sharpness" at all

jp wrote:

I do not agree with your "per pixel sharpness" argument at all
Look for example at the details on the rainpipe: the mounting screws
and the bending folds are almost invisible on the 20S pctures.

I still hold to my oponin looking at sharp branches and edge defination, as well as texture. (note there is 4 million pixel difference between the cameras, D300 file is more than twice as large as 20D file)

I do agree with the noise difference, that was also my observation.
But better hve some noise in the shadows than blown highlights of
important image parts.

I did mention this several times in my report.

If you expose for the same brightness of your pictures, taking into
account that the ISO values of the D300 are around 1/3 stop too
optimistic , you will see that the noise advantage of the 40D becomes
smalller.
If you would expose both pictures for the highlights then the overal
picture of the D300 would be brighter than that on the 20D however,
but noise would again become less and on par between both cameras
after resizing.

Whatever I do there are people who will say why did you expose for this, and why didn't you expose for that. I used an accurate exposure for the scene I intended to photograph and show the fall colors and contrast, based on what I know about histograms and both cameras had the same exposure, why would I force D300 to blow up? I didn't have the 40D so I wouldn't know what the reading on that camera would be.

I agree that nikon does not beat canon in noise, the more ore less
closed the gap but they did not take the lead.
That said, noise is one thing but as you said yourself, there is more
to it than only noise to evaluate the quality of a DSLR.

of course.

BTW: did you try the 3D tracking with CF-A1 set to focus priority?

Not yet, but I think dynamic 9-point is the best mode for me for tracking birds.

-- hide signature --

cameras don't shoot people
People shoot people.

Lee H Senior Member • Posts: 2,013
Problems with your tests.....

I’m no expert and correct me if I’m wrong, but didn’t you say the outside house shots were ISO 400?

I simply downloaded the two comparison shots of yours, viewing in fastStone image viewer, your test seems to have discrepancies such as, (See screen captures)
Time of day or Processed
Exposure Mode
Focal length
And ISO

Perhaps someone with greater knowledge and software could examine all the shots with greater accuracy than myself, also it should be noted that Faststone may have problems reading D300 EXIF?

I’m as intrigued as anyone with what seems like unlikely results

-- hide signature --

Regards,
Lee

OP Arash Hazeghi Contributing Member • Posts: 967
EOT

I think this thread has reached the end of its usefulness, I think whoever wanted to see such comparisson, and draw whatever conclusion has already done so, I thank you all for your attention and also the helpful comments, and I don't take any of the not-so-helpful comments personally. I will keep the samples alive for a while in case other people want to see. I hope you can enjoy your new camera and take good shots.

Good night.

Arash

jp Senior Member • Posts: 1,151
Re: Arash: I do not agree on "per pixel sharpness" at all

Whatever I do there are people who will say why did you expose for
this, and why didn't you expose for that.

True, I was only mentioning it, it was not meant as a criticism at all

I used an accurate exposure
for the scene I intended to photograph and show the fall colors and
contrast, based on what I know about histograms and both cameras had
the same exposure, why would I force D300 to blow up?

I only mean, if you would expose on both cameras to just the point of saturation, the difference would again become smaller.

On the indoor scene for example, you have allowed the 20D to blow clearly more highlights than the D300. If you would allow the D300 the same amount of blown highlights, the image would become brigter and noise would have been less. Again, not criticizing your ablilities, just saying that in order to compare noise in certain circumstances, you may want to take into account several factors.

I didn't have
the 40D so I wouldn't know what the reading on that camera would be.

I meant 20D, it was a typo

BTW: I still agree with you that noise on the canons from 20D till 40D is still slightly lower than on the D300. But the difference may just be less than what your examples illustrate.

Thanks for all your work and effort
much appreciated.

OP Arash Hazeghi Contributing Member • Posts: 967
Re: Problems with your tests.....

Lee H wrote:

I’m no expert and correct me if I’m wrong, but didn’t you say the
outside house shots were ISO 400?
I simply downloaded the two comparison shots of yours, viewing in
fastStone image viewer, your test seems to have discrepancies such
as, (See screen captures)
Time of day or Processed

shot at ~ 2PM (PST)

Exposure Mode

exposure mode doesn't matter as long as exposure is the same

Focal length

20D crop factor is 1.6, D300 is 1.5

And ISO

look at the iso for crops, the resized image iso was 100 or 200 i think

Perhaps someone with greater knowledge and software could examine all
the shots with greater accuracy than myself, also it should be noted
that Faststone may have problems reading D300 EXIF?

yes not all sw can read D300 ISO, especially ISO 100, because it is not "ISO 100"

I’m as intrigued as anyone with what seems like unlikely results

-- hide signature --

Regards,
Lee

OP Arash Hazeghi Contributing Member • Posts: 967
Re: Problems with your tests.....

Arash Hazeghi wrote:

Lee H wrote:

I’m no expert and correct me if I’m wrong, but didn’t you say the
outside house shots were ISO 400?
I simply downloaded the two comparison shots of yours, viewing in
fastStone image viewer, your test seems to have discrepancies such
as, (See screen captures)
Time of day or Processed

shot at ~ 2PM (PST)

oh one of the cameras is on daylight saving time! hehe looks like you tried REALLY hard to find something wrong....

Exposure Mode

exposure mode doesn't matter as long as exposure is the same

Focal length

20D crop factor is 1.6, D300 is 1.5

And ISO

look at the iso for crops, the resized image iso was 100 or 200 i think

Perhaps someone with greater knowledge and software could examine all
the shots with greater accuracy than myself, also it should be noted
that Faststone may have problems reading D300 EXIF?

yes not all sw can read D300 ISO, especially ISO 100, because it is
not "ISO 100"

I’m as intrigued as anyone with what seems like unlikely results

-- hide signature --

Regards,
Lee

OP Arash Hazeghi Contributing Member • Posts: 967
thank you as well

n/t

jp wrote:

Whatever I do there are people who will say why did you expose for
this, and why didn't you expose for that.

True, I was only mentioning it, it was not meant as a criticism at all

I used an accurate exposure
for the scene I intended to photograph and show the fall colors and
contrast, based on what I know about histograms and both cameras had
the same exposure, why would I force D300 to blow up?

I only mean, if you would expose on both cameras to just the point of
saturation, the difference would again become smaller.
On the indoor scene for example, you have allowed the 20D to blow
clearly more highlights than the D300. If you would allow the D300
the same amount of blown highlights, the image would become brigter
and noise would have been less. Again, not criticizing your
ablilities, just saying that in order to compare noise in certain
circumstances, you may want to take into account several factors.

I didn't have
the 40D so I wouldn't know what the reading on that camera would be.

I meant 20D, it was a typo

BTW: I still agree with you that noise on the canons from 20D till
40D is still slightly lower than on the D300. But the difference may
just be less than what your examples illustrate.

Thanks for all your work and effort
much appreciated.

Erciro Senior Member • Posts: 1,708
Re: Sorry

Arash Hazeghi wrote:

Wrong, what I was talking about was 3D tracking of a car at 5mph, please read the post carefully. no high speed moving subject comment because I haven't tried.

But you also forgot to add that the van stopped at the speed bump when the sequence started. I don't know for sure but that might cause problems.

-- hide signature --

Ahmet
---------
'All that is necessary for the triumph of evil is that good men do nothing.'
Edmund Burke

 Erciro's gear list:Erciro's gear list
Nikon Coolpix AW120 Nikon D300 Fujifilm X-T1 Nikon AF-S Nikkor 70-200mm f/2.8G ED VR Nikon AF DX Fisheye-Nikkor 10.5mm f/2.8G ED +9 more
jp Senior Member • Posts: 1,151
No discrepancies, but he made an error

Arash explained most things

And the difference in brightness comes from the fact that ISO values on the 20D are incorrect.
ISO100 is actually ISO125 ISO200 is actually ISO250 etc

That said, the canon metering in simple scenes is about 1/3 stop conservative (always been) and this compensates the higher sensitivity.

But if you use the metering results of another camera or an external one on the 20D/30D, then you get incorrect exposures. (40D is correct now)

That said: I do not believe that the 20D meter would actually have done better in these high contrast circumstances, the matrix meter is clearly superior with this kind of scenes.

-- hide signature --

cameras don't shoot people
People shoot people.

Flashlight Veteran Member • Posts: 7,895
Re: EOT

Arash Hazeghi wrote:

I think
whoever wanted to see such comparisson, and draw whatever conclusion
has already done so,

Well, I just woke up, while you posted these when I went to bed, so...

The only thing I want to remark, again, is that your D300 shots look much noisier than mine with NR off. Wonder how that can be.

-- hide signature --

Philip

Keyboard shortcuts:
FForum MMy threads