Wait till you see D300 high ISO images: WOW.

[John Sheehy]

I just wonder how you can say that the NR removes detail from these.
If its the resolution charts you look at, its obvious that the D300
retains detail a lot better than the D200.

Test charts have nearly 100% contrast. How do you expect noise to compete against that?

--
John

 

[John Sheehy]

Feature rich but below par in picture quality is nothing new from
Nikon. What are they going to do differently relative to the
competition his time? Many old film cameras have richer AF features,
many point & shoots cams have features that leaves dSLR standing.
Nikon merely adopting more of these features for the sake of IQ on
their dSLR. The whole points of dSLR is superior IQ and speed and
guess who is ahead and who lagging?

Well, if you're not shooting in low light with shadows of interest, all DSRLs are capable of much better IQ than small-sensor P&S cameras, due to a much higher photon collection (which does not vary anywhere as much as read nosie does, between DSLRs).

If I'm walking around during the daytime with my Canon DSLR, with ample light, I don't see people with Nikons and think, "Their IQ is going to be inferior". I only think that when it starts getting dark, or slower lenses or lens+TC combos are being used, hand-held. Small-sensor cameras have either chroma noise or NR artifacts, even at ISO 100, when viewed at 100% or printed large.

--
John

 

[PerL]

I just wonder how you can say that the NR removes detail from these.
If its the resolution charts you look at, its obvious that the D300
retains detail a lot better than the D200.

Test charts have nearly 100% contrast. How do you expect noise to
compete against that?

I dont know if you have seen the thread the poster refered to. The comparision showed only gray patches and resolution charts. Thats why I questioned how he could draw conclusions about how much a D300 retains detail compared to a D200.

--
John

--
Small D200 gallery:
http://www.pbase.com/interactive/d200_12
Small D40 gallery:
http://www.pbase.com/interactive/d40_12
Small Nikon P5000 gallery:
http://www.pbase.com/interactive/nikonp5000_12
http://www.pbase.com/interactive

 

[thw]

In a D300 ISO 1600 RAW I have here, with some near-black areas, the

read noise is about 0.4% to 0.5% of saturation... The D200 is about 0.64%, and the D2X about
1.47%. The worst Canon (400D) is about 0.18%; the best Canon (1Dmk3)
about 0.08%.

What is the read noise for D3 at ISO 1600 as a fraction of the saturation level? Heard you've already got some D3 RAW files.

-------------------------------------------
See the colors of my world in:
thw.smugmug.com

 

[John Sheehy]

Even from the 350D to the 40D, d200, d80 d40x etc , no one could tell
the difference from 8x10 prints of properly exposed and processed
images from these cameras.

This is a given. I don't know why people still bring this up. The differences in noise performance between different digital cameras with the same size sensors is mainly in the deep shadows. Your "properly exposed" phrase excludes images/ranges affected by the very real differences between cameras.

As long as people can't distinguish between read noise and shot noise, a lot of typing will be wasted discussing differences that fail to be differentiated.

The talk of iso noise is , I beleive , somewhat moot right now,

Yes, in your well-exposed, deep-shadow-less images.

the
D300/40D seem to be very close, certainly both cams can deliver iso
400 and above better than anything my film cameras can do , so , to
me at least, would suggest they are 'good enough'.

The elements of film noise are randomly sized and placed, though, so even though they may be quite visible, they are more aesthetically pleasing to me (pound for pound). Digital noise is gridded, and otherwise-patterned, and is more artificial-looking and distracting, pound-for-pound. Film grain noise looks like a texture of the recording medium, and so does photon shot noise in digitals (which is actually almost the same thing, except the digital is gridded), but digital read noise looks like a blanket of noise added above the image in the deep shadows, as it is not correlated to signal.

I've done some preliminary simulations of what high-ISO would be like with shot noise only (no read noise; just the camera accurately counting photons) and pound-for-pound, shot noise in the deep shadows is much less distracting and obscuring of subject detail than read noise, and is much more aesthetic. The notion that "current cameras are shot noise limited" as one of the most-quoted people on digital image quality concludes, is incorrect, and based on looking at numbers in flat, out-of focus areas, and not at the visual qualities of read and shot noises, especially in how they affect detailed subjects.

Think of a B&W resolution test, for example. Say the white parts average 3 electrons per pixel, and the black parts 0.1 electrons per pixel. With read noise as the dominant noise, half the pixels in the black parts will clip at black, and the other half range from black through grey, even clipping at white. The white areas will have half their pixels clipped at white, and the rest ranging down to black, and even clipping at black. There is a blanket of uniform noise over the image, except as altered by the clipping. With pure shot noise, the white areas will be similar to the read noise of the same statistical strength, but the black areas of the test chart will be almost black, with an occasional photon being captuired in a well-isolated pixel (the chance of a neighboring pixel also registering a photon being very low). You can remove these isolated photons or blur them away so that the image is still as accurate at a lower resolution, so that the contrast doesn't hit you in the face. Whether or not you deal with those isolated pixels, the black areas will still be far more distinct than they would with pure read noise, which would have the same strength in the black areas as in the white areas.

So, I say that reducing read-related noises to zero is a worthwhile goal, as it would

increase dynamic range tremendously, and make for far more usable high ISOs, at least as far as the shadows are concerned.

--
John

 

[Dejan Malikovski]

In a D300 ISO 1600 RAW I have here, with some near-black areas, the

read noise is about 0.4% to 0.5% of saturation... The D200 is about 0.64%, and the D2X about
1.47%. The worst Canon (400D) is about 0.18%; the best Canon (1Dmk3)
about 0.08%.

What is the read noise for D3 at ISO 1600 as a fraction of the
saturation level? Heard you've already got some D3 RAW files.

I presume if Nikon labels the D3 as a Canon camera the read noise will suddenly drop by at least 5 times. From recent studies the read noise has nothing to do with the sensor, but is very closely related to the front label of the camera. The letters in the name of Canon, on the front label, reduce the gel-assumption of the sensor and make the read noise a lot smaller (5-times to be precise).

--



2007 Digital Camera Satisfaction Study:
http://www.jdpower.com/electronics/ratings/digital_camera/dslr

 

[John Sheehy]

I dont know if you have seen the thread the poster refered to. The
comparision showed only gray patches and resolution charts. Thats why
I questioned how he could draw conclusions about how much a D300
retains detail compared to a D200.

Sorry, I read that fast and thought it was against the 40D that the comparison was being made, suggesting that the D300 had low noise because it had better resolution.

--
John

 

[John Sheehy]

...to some actual images where this read noise matters? Its very hard
to get a grip over what it means in practical use.

Any image where you need to utilize the deep shadows. The easiest way to do this is to take the same image with two cameras, set one to -2 or -3 EC, set the other to the same Av and Tv values in manual mode at a high ISO, and make generic, homogenous conversions in DCRAW or IRIS, or some other program that allows exactly the same conversion parameters for different cameras (except asnecessary to get the same tonal levels, saturation levels, and WB in the output). This will show what each camera is capable of in really low light.

So far, I can only garner statistics from unrelated exposures. I've seen all kinds of people doing comparisons allegedly "in RAW", but only showing conversions. Conversions are not RAW. They are from RAW, just like in-camera JPEGs are.

--
John

 

[John Sheehy]

In a D300 ISO 1600 RAW I have here, with some near-black areas, the

read noise is about 0.4% to 0.5% of saturation... The D200 is about 0.64%, and the D2X about
1.47%. The worst Canon (400D) is about 0.18%; the best Canon (1Dmk3)
about 0.08%.

What is the read noise for D3 at ISO 1600 as a fraction of the
saturation level? Heard you've already got some D3 RAW files.

The thumbnail of the D3 images I have look promising; they are in a gymnasium that is poorly illuminated in some areas, but the only program I have that claims to parse D3 RAWs is DCRAW, and the TIFFs and PGMs it outputs from the D3 are only 2 bits (at least in the programs I have tried to open them in), so it isn't quite working yet.

However, the full-res JPEGs I've seen have enough NR artifacts in the shadows that there may be nothing break-through here in terms of read noise, either, but the D3 starts at ISO 200, and if this is due to high quantum efficiency, then the D3 may very well be the cleanest DSLR at high ISOs in the midtones and highlights, and won't have significant noise until you go deep into the shadows (rather abruptly, somewhere down there).

If the ISO 200 is due to small photosites, then The D3's only benefits over the D300 in midtones and highlights will be the optical benefit of not being as demanding on the lenses, the ability to fully use wide lenses, and get lower DOF, etc.

--
John

 

[John Sheehy]

I presume if Nikon labels the D3 as a Canon camera the read noise
will suddenly drop by at least 5 times. From recent studies the read
noise has nothing to do with the sensor, but is very closely related
to the front label of the camera. The letters in the name of Canon,
on the front label, reduce the gel-assumption of the sensor and make
the read noise a lot smaller (5-times to be precise).

You forgot to say one thing; "Honk! Honk!"

--
John

 

[Taikonaut]

Feature rich but below par in picture quality is nothing new from
Nikon. What are they going to do differently relative to the
competition his time? Many old film cameras have richer AF features,
many point & shoots cams have features that leaves dSLR standing.
Nikon merely adopting more of these features for the sake of IQ on
their dSLR. The whole points of dSLR is superior IQ and speed and
guess who is ahead and who lagging?

Well, if you're not shooting in low light with shadows of interest,
all DSRLs are capable of much better IQ than small-sensor P&S
cameras, due to a much higher photon collection (which does not vary
anywhere as much as read nosie does, between DSLRs).

If I'm walking around during the daytime with my Canon DSLR, with
ample light, I don't see people with Nikons and think, "Their IQ is
going to be inferior". I only think that when it starts getting
dark, or slower lenses or lens+TC combos are being used, hand-held.
Small-sensor cameras have either chroma noise or NR artifacts, even
at ISO 100, when viewed at 100% or printed large.

--
John

And Nikon dSLR is closer to the P&S feature wise and image wise.

 

[Gary0001]

A bunch of us bought 40Ds for approximately $1100 with the IS lens.
An equivalent D300 would be $2400ish, more than twice the cost ! For
that it should be better, a lot better !

In reality, I bet its IQ won't measure up to the 40D. I think Nikon
made a mistake moving the D300 from 10MP to 12MP. They should have
stayed at 10MP and lowered the high ISO noise. Before you call me a
Canon lover, my last camera was a D50.

I could have bought a D300. I bought a 40D with a 28-135 IS,
50/f1.8, 430EX flash and 85 f/1.8 for less than $1700. Nikon needs
to lower the price of the D300 and improve the availability to
compete.

Really I'm in the opposite boat, I think they should raise the price and flexibility of the 40d so it's equal to the D300. Personally I need a camera that shoots 6-8fps, not 6fps above 1/4000th of a second. And I'm not sure where you get the 1250 more then the d300 figure.

d300 $1800 vs. 40d $1100 body only and my math says $700 more.

--

 

[thw]

And Nikon dSLR is closer to the P&S feature wise and image wise.

That is so true! I suppose by image, you mean the increased saturation, contrast and noise reduction. Also, Canon compact digicams are laden with features that surprisingly are not included in their DSLRs. Weird.

-------------------------------------------
See the colors of my world in:
thw.smugmug.com

 

[thomas2279]

Exactly it is not even out and people are making judgements how good the image quality is better than cameras that are out and in large no of users hands.

I could pick up several 40d's in the shop on day one as Canon have delivered large amount of stock into the shops in a short period after announcement; ulike Nikon where it will be hard to find one between 21/11/07 & Xmas.

Although the D300 will be a great camera it is not even out; is £500 dearer than 40d now which is a very good camera and with that difference you can plonk this on some lens & flash. Plus if you want more out of a Canoncamera you can always upgrade later to 5d replacement; 1d/s Mk 3, etc.

I have only seen a modest improvement of noise over D200 by 1 stop typically but other features are quite welcome on it,, although I may sell my D200 to get optinum price now and pick D300 later next year but in meantime I will shoot with excellent 40d & Nikon d40x cameras.
--



My Pictures & Web Site: http://www.dltp.co.uk

 

[saynomore]

You should get a 5D and start hanging out at the 5D/1D forum. Or just go over there and say you have one. I don't have a 20D anymore, and it seems better discussions are being done in this forum than in the other one.

 

[Dejan Malikovski]

And Nikon dSLR is closer to the P&S feature wise and image wise.

That is so true! I suppose by image, you mean the increased
saturation, contrast and noise reduction. Also, Canon compact
digicams are laden with features that surprisingly are not included
in their DSLRs. Weird.

He didn't mean that. He meant to say that Nikon images are closer to P&S in terms of ISO performance. Between the lines...

I, as a matter of fact, don't share that particular thought. And I feel offended (again). Should I be? No. Why? "Because everyone that has made an actual test, is saying that they(Canon and Nikon) are the same in terms of IQ/ISO."
Munch away...
--



2007 Digital Camera Satisfaction Study:
http://www.jdpower.com/electronics/ratings/digital_camera/dslr

 

[Dejan Malikovski]

You should get a 5D and start hanging out at the 5D/1D forum. Or just
go over there and say you have one. I don't have a 20D anymore, and
it seems better discussions are being done in this forum than in the
other one.

My opinion exactly. Especially the part with "say that you have one" is very good. "Say that you have raw files" - rings a bell.
--



2007 Digital Camera Satisfaction Study:
http://www.jdpower.com/electronics/ratings/digital_camera/dslr

 

[saynomore]

You should get a 5D and start hanging out at the 5D/1D forum. Or just
go over there and say you have one. I don't have a 20D anymore, and
it seems better discussions are being done in this forum than in the
other one.

My opinion exactly. Especially the part with "say that you have one"
is very good. "Say that you have raw files" - rings a bell.
--

I'm pretty sure he does have RAW files. I was just saying he should say he has a 5D anyway, but clearly I was just kidding. I don't understand you, you criticize him, and then agree with me. Explain, please?

 

[svx94]

A nice list about the differences! But I do think they worth $500. And in fact, the street price will show much less difference than $500. Just think about the price difference between the 1Ds and 5D...

I will add one more point to your list, "linking the spot AE to all AF points" - very important to me.

But to Rahul's point, lens is still on the Canon side, mainly because of the 17-55 IS and both 70-200 F4L versions, may be the 10-22 as well. They cover the most used focal range. For that reason, if I have to pick between 40D and D300 today, I may stay with Canon.

Nikon will make up the lens gap sooner or later; Canon will put more goodies on the next 5D. I think the picture goods great for buyers!

Cheers!

Congrats to Nikon for FINALLY catching up in the area of high ISO,
but it really is interesting to see Nikon shooters so gleeful> > >

As a former 30D owner (and regular Nikon user), I know that the D300
samples shown show about the same noise...perhaps a bit worse/better,
but nothing very different from the 30D. ISO 3200 prints a beautiful
8.5X100 print and so does the 30D and 40D if you've exposed properly
and avoided cropping. End of THAT story.
The D300's advantages for FIVE HUNDRED DOLLARS MORE are as follows:

1) Totally professional Auto Focus. Same module used in the D3.
2) Faster frame rate potential
3) The ONLY pro LCD on the market. That's right, PRO LCD. Once you
see it on a D3 or D300 you'll realize that the rear LCD has been a
missing tool for some time.
4) Higher MP
5) FULL 100% coverage viewfinder
6) HDMI Out
7) Weather Sealing
8) Custom focus adjustment for lens issues
9) Functional AF system while Live View is activated
10) On demand grid lines
11) 100 shot buffer (Not sure about this one yet)
12) Higher rated shutter life (By 50'000)
13) Advanced Wireless commander function built in

I'm sure a few things are off on my list, but that's the jist of it.
Is all of that worth 500 bucks? That's up to you. Yapping about ISO
differences is fun, but in real world shooting I've done fine with
the D200, 30D, and D80. They were all quite close when I did my job
right. So those features above are more of what the D300 is all about
and why it costs more. NOT the ISO range. For the price, the 40D is
utterly fantastic. But the D300 is a higher end camera and will do
more for SOME people. If that's not "you" then you saved 500 bucks.
Nikon and Canon did not build the 40D and D300 to compete head to
head. The Canon beats everything in it's class, such as the aging D80
and D200. Nikon slotted the D300 between the 40D and 5D. It has no
real battle to win at it's price point. And the 5D, 40D and D300 are
all VERY viable cameras....at their respective price points. For MY
usage, the D300 beats them both. That's the great thing here. We all
get to pick out what suits us.

Cheers,

Thomas

 

[svx94]

about a year ago, I made a long wishlist for 40D with 13 some items. I was expecting Canon to fulfill a few points. It turned out Canon only missed a few points. That was as big improvement as you can get for a camera !!! Compare your car with a 50-year old model, how much really changed there?

For the next 5D, I am sure Canon will put all the improvement from 40D in it. But I'd like to see two additional changes:
1. Link the spot AE to all AF point;

2. Modify the AF sensor for FF, not just use the module from APS-C. This will provide better AF area coverage.

I certainly like to see more Nikon features in Canon body put don't put any hope on it, namely:
1. on-demand grid in the VF;
2. build-in wireless flash command

With that said, hope the next 5D IQ won't suffer for more "useless pixels" !

Cheers!

Thank you for sharing what I don't have... Thank you for making me
jealous of you Nikon D300 & D3 owners.. Thank you for rubbing it in
my face, because now I want to cry..

Canon, I hate you, but I'm stuck with your mediocrity lately, until
the next generation of cameras,...(I presume!) I am not really
looking forward to the 5D upgrade if it is going to be like the
40D..(waste of money)...with a fatter but not any cleaner
sensor.....that performs like a FAT lazy SLOB!!

;-)

JP
--

I'm no 'John Holmes', but I do have a rather 'Large' lens..

http://www.Myspace.com/JPphotographer