D70 / 10D ISO Comparisons - DPChallenge Forums

Sigma 24-70mm f/2.8 EX Aspherical DF for Canon

05/12/2004 09:50:36 AM · #1

I performed this test to satisfy some of my own curiousities about things I read regardging CCD vs CMOS sensors.

The test:

I setup my 10d and my dad's d70 side by side on tripods. I composed, as closely as possible, the same night time scene. I then resized each image to 8x12" dimensions and then cropped out a 4x6" segment to attempt to simulate noise at print size.

Canon Lens: 28-105mm f/3.5-4.5 @ 28mm
Nikkor Lens: 28-105mm f/3.5-4.5 @ 28mm

You can view these photos and draw your own conclusions...

//www.pbase.com/jmsetzler/iso_tests

magnetic9999

Canon EOS-20D

05/12/2004 10:23:30 AM · #2

very interesting conclusions .. about what i had suspected seeing the shots from both cameras around ..

thanks john :D

garlic

Canon EOS-10D

Sigma 15-30mm f/3.5-4.5 EX Aspherical DG DF for Canon

05/12/2004 10:46:15 AM · #3

whats you conclutuion jmsetzler?

pitsaman

Canon EOS-30D

Canon EF 50mm f/1.8 II

05/12/2004 10:49:47 AM · #4

Canon looks 3 times cleaner :-)
Which makes sense to me,cmos devices are slower but less noisier than regular transistors!

Message edited by author 2004-05-12 10:51:54.

Canon EF 70-200mm f/2.8L USM

05/12/2004 10:58:05 AM · #5

very little difference that I can see. I haven't done comparisons at night yet, but now I don't have to. It's actually remarkable how similar the images are, given the different technologies.

Pitsaman, you sure you have both lenses in your glasses? ;)

PaulMdx

Canon EOS-1D

05/12/2004 11:06:22 AM · #6

I think the biggest thing I notice is colour reproduction rather than ISO.. I really need to look at home on my decent monitor later.

DJLuba

Nikon D2Xs

Nikon AF DC Nikkor 135mm f/2.0D

05/12/2004 11:11:18 AM · #7

John -- Did you have the internal settings the same?

Sharpening - off
Auto Contrast - none

etc...

I see a very tangible difference within the photos. Great experiment.

pitsaman

Canon EOS-30D

Canon EF 50mm f/1.8 II

05/12/2004 11:12:41 AM · #8

Originally posted by Pedro:

Take that thing back and get a Rebel,don't be typical Canadian :-)

05/12/2004 11:13:42 AM · #9

Originally posted by DJLuba:

John -- Did you have the internal settings the same?

Sharpening - off
Auto Contrast - none

etc...

I see a very tangible difference within the photos. Great experiment.

Yes... both cameras were at default factory settings on all internal adjustments.

EddyG

Canon EF 70-200mm f/2.8L IS USM

05/12/2004 11:16:03 AM · #10

Looking at the 100% crops at ISO 1600, I agree with pitsaman.

These are the images I am looking at: 10D / D70

Specific areas I notice: there is a plastic temporary sign (to the left of the one that says "Outdoor Supply). Look at the wall area around that sign, and in the alcove underneath that temporary sign: a lot less noise on the 10D. Now go straight up from there and look at the noise in the dark areas above the signage and in the canopies over the windows...

Considering how much newer the D70 is than the 10D, and seeing what Canon was able to achieve in terms of high ISO on the 1D Mark II, I'd say Canon's CMOS sensor technology rocks!

Message edited by author 2004-05-12 11:17:51.

Canon EF-S 10-22mm f/3.5-4.5 USM

05/12/2004 11:23:36 AM · #11

Looking at these images, I do see that the Nikon's auto white balance seems to handle the light in this scene better than the Canon.

undieyatch

Olympus E-10

05/12/2004 11:34:56 AM · #12

......the canon has a redish tint and the nikon, has a cooler colour, is more toward the blue..... the nikon apprears to me to be a more realistic colour rendering.

Fujifilm XF 55-200mm f/3.5-4.8 R LM OIS

05/12/2004 11:50:44 AM · #13

Originally posted by pitsaman:

Originally posted by Pedro:
very little difference that I can see. I haven't done comparisons at night yet, but now I don't have to. It's actually remarkable how similar the images are, given the different technologies.

Pitsaman, you sure you have both lenses in your glasses? ;)

Take that thing back and get a Rebel,don't be typical Canadian :-)

I have both ;)

ursula

Fujifilm X-T2

05/12/2004 12:01:48 PM · #14

There is a difference!! I've been reading that there is very little difference, but, at least from John's pictures, there is a difference. Just from a "liking" point of view (or which look more pleasing to my eyes), the Canon pictures are better.

Now, would there be a difference if it were the Rebel rather than the 10D, or would that be pretty much exactly the same?

Ursula

EddyG

Canon EF 70-200mm f/2.8L IS USM

05/12/2004 12:20:44 PM · #15

Because of the reference I made earlier to the high-ISO performance of Canon's latest incarnation of their CMOS sensor (used in the 1D Mark II), I found an ISO 1600 shot I took at a wedding I attended last weekend to attempt to show this.

I took these with the Canon 16-35/2.8L wide-open at Æ�/2.8, ISO 1600, 1/30th of a second. No flash. It was a very dimly lit country club reception hall, with some colored "DJ lights" to camera-left of the bride.

I did no post-processing of these images. (No noise reduction, no sharpening, etc.) The first is the frame cropped to 8x10 aspect ratio, and then just resized. The second is a 100% crop from the same frame, showing the actual 1D Mark II pixels. (There is some motion blur evident since she is throwing the bouquet, but I was more than impressed with the image quality at ISO 1600.)

Message edited by author 2004-05-12 12:24:28.

garlic

Canon EOS-10D

Sigma 15-30mm f/3.5-4.5 EX Aspherical DG DF for Canon

05/12/2004 12:21:08 PM · #16

Originally posted by ursula:

Now, would there be a difference if it were the Rebel rather than the 10D, or would that be pretty much exactly the same?

Would supose pretty much the same results. Same senosor (or censor ?) but pentamirror instead of pentaprism. Not sure how much difference that would make under thouse circumstances.

Fujifilm XF 55-200mm f/3.5-4.8 R LM OIS

05/12/2004 12:36:42 PM · #17

Originally posted by garlic:

Originally posted by ursula:

Now, would there be a difference if it were the Rebel rather than the 10D, or would that be pretty much exactly the same?

Would supose pretty much the same results. Same senosor (or censor ?) but pentamirror instead of pentaprism. Not sure how much difference that would make under thouse circumstances.

It would not affect the image. That viewfinder system in the rebel just makes the image in the viewfinder a little dimmer.

ursula

Fujifilm X-T2

05/12/2004 03:36:26 PM · #18

Thank you!

Does anyone know if Canon is coming out with a "new" or "updated" Rebel in the Fall?

05/12/2004 03:59:13 PM · #19

this is a little excerpt from the dpreview comparison study:

"The first thing you notice is the difference in color balance between the two cameras, in this respect the D70 is more accurate, the EOS 300D's image looking warmer than it did in real life, the D70's greys are almost perfectly grey. The D70's colors also look slightly more saturated than the 300D. For absolute resolution it's very close, although there are definitely some areas of the image where the D70 has resolved more detail (and appears sharper) than the EOS 300D. To counter that we do have a little moirÃ© appearing in the D70 image (although really nothing that spoils the appearance of the image).

One notable difference is that the D70's sharpening algorithm appears to be better than that of the EOS 300D, there are almost no visible sharpening halos in the D70 image, there are some in the EOS 300D image.

Lastly there does appear to be some noise visible in the sky of the D70 image (approx. 1.75 std dev), less so in the EOS 300D image (approx. 1.10 std dev) this is consistent with our ISO noise measurements presented earlier in this review. Remember that the EOS 300D does offer ISO 100 which will deliver even cleaner images with virtually no visible noise."

Azrifel

Fujifilm X100VI

05/12/2004 04:05:44 PM · #20

Originally posted by EddyG:

I did no post-processing of these images. (No noise reduction, no sharpening, etc.) The first is the frame cropped to 8x10 aspect ratio, and then just resized. The second is a 100% crop from the same frame, showing the actual 1D Mark II pixels. (There is some motion blur evident since she is throwing the bouquet, but I was more than impressed with the image quality at ISO 1600.)

Impressive, there is hardly any chroma noise visible, the grain is something I could live with. The fullsize holds its color and DR quality very well.

Did you take the photo with a slow shutter and no flash on purpose? I think that the motion blur of the bouquet helps to tell the story of the moment. If the girls in the background were diving towards it screaming and kickin' it would be perfect. :)

Azrifel

Fujifilm X100VI

05/12/2004 04:10:02 PM · #21

Originally posted by ursula:

Thank you!

Does anyone know if Canon is coming out with a "new" or "updated" Rebel in the Fall?

In Japan they introduced a black one a month ago. :)

I think they will not introduce a new Rebel this year. The 10D and 1Ds are the next candidates for an update. Perhaps there will be new official firmware, because there are some hacks available that unlock some 10D functions for the 300D. If you see what the D70 offers in terms of control -> a whole lot more.

Message edited by author 2004-05-12 16:10:43.

paganini

Canon EF 70-200mm f/4.0L USM

05/12/2004 04:16:11 PM · #22

You know, for a $4500, you would've thought they'd do the autowhite balance right for a change....

Originally posted by EddyG:

pitsaman

Canon EOS-30D

Canon EF 50mm f/1.8 II

05/12/2004 04:47:25 PM · #23

Originally posted by Pedro:

How much you trust DPReview,after all they are for profit organisation.
They used to push Olympus 5050 so hard,I bought one (wasted my money)now they do the same thing with Oly 8080 and D70...

EddyG

Canon EF 70-200mm f/2.8L IS USM

05/12/2004 05:58:11 PM · #24

Originally posted by paganini:

You know, for a $4500, you would've thought they'd do the autowhite balance right for a change....

Given that there are flashing, multi-colored DJ lights (which I mentioned in my previous post) shining on the ceiling above the DJ area and on the bride (directly to camera-left of the bride, a couple feet off the floor), what do you propose would be a more accurate "autowhite balance"?

I feel the captured image is pretty darn close to how the scene actually looked that evening. Sure I could probably remove the "tint" caused by the DJ's lights to make her dress "snow white", but that wasn't really how it looked...