Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: Rob's review is now up on 14n  (Read 5570 times)
Ray
Sr. Member
****
Offline Offline

Posts: 8878


« on: April 15, 2003, 06:43:40 PM »
ReplyReply

Good review, but no surprises in the noise department. It seems the general perception that the 14n is a noisy camera still holds true. The point that noise is often worse than it need be due to the user's unfamiliarity with Photodesk and other noise reducing software, is really a separate issue. All photos can be improved in all sorts of ways if the user has the skill, the patience and the time, and any post processing noise reduction techniques applied to the 14n can also be applied to any digital image to produce at least some improvement.

What I find the most disturbing is the implication that noise in the shadows could still be unacceptable at ISO 80 in certain high contrast or low light situations.

Once again there's no comparison between the 14n and 1Ds with both cameras set at ISO 100 and 200.
Logged
Paul Caldwell
Newbie
*
Offline Offline

Posts: 25


« Reply #1 on: April 16, 2003, 09:00:42 AM »
ReplyReply

Actually based on what I have seen on both Michael's review and Rob's more so Rob's, the 14n and 1ds in regards to noise are closer than I had originally thought.

My 1ds is far from clean, in fact even at 100 ISO I can still pull noise, especially in blue skies.  Enough noise that I often need to use Neat Image.  

Others seem to have much better 1ds cameras in regards to the higher ISO's.  In Rob's review, you can look at a ISO 400 shot he took of a sky line at evening.  My camera couldn't do that, that cleanly.  In fact my ISO 200 would be very close to that, but not my 400.

The issue I don't like on the 14n, besides some of the body design issues, is that you can't turn off noise reduction.  From Rob's review you can clearly see the effect of the noise reduction, the abstract look, or brushed look like I have commented on before.  

The 1ds has many more ISO options, but if you can't use them in everyday work, what good are they.  

Also I haven't sent in my camera to Canon as I am already sure the answer will be "within specs".  Common.

However there have been plenty of other posts in regards to noise in the images.

Paul Caldwell
Logged
Guest
« Reply #2 on: April 16, 2003, 03:14:10 PM »
ReplyReply

I can't comment on the problem that Paul may be having with his Canon 1Ds but I have extensively used 2 different bodies; the one I tested in September and October, and the one that I've owned since November 2002, and both have been exceedingly noise free at ISO 100 - 400. Essentially noise is nonexistant at these speeds. At 800 and 1000 it's there, but not serious.

I also have seen prints made with at least 5 other 1Ds bodies owned by friends and folks that have come on my workshops, and all had very similar characteristics.

The only camera I've ever seen with better hgih ISO noise characteristics is the Canon 10D.

Michael
Logged
sergio
Sr. Member
****
Offline Offline

Posts: 661


WWW
« Reply #3 on: April 16, 2003, 09:41:58 PM »
ReplyReply

I understand outside temperatures can affect performance of the sensor. Is it hot whet you are?
Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8878


« Reply #4 on: April 21, 2003, 08:48:43 PM »
ReplyReply

Samir,
Congratulations on an excellent web site  for those interested in astronomy! Very informative. But I'm puzzled by the D30/1Ds comparison of shots of Trapezium. You describe the D30 as being 'hopelessly-too-noisy-for-faint-fuzzies'. And indeed, the 1Ds shot looks a hundred times better than the D30 crop.

Either you're pulling rabbits out of the hat or I'm missing some crucial point. (Both perhaps). There have been reports on this web site of D30/14n comparisons at the pixel level. It seems a D30 pixel (after in-camera processing of course) is less noisy than a 14n pixel. BLJ has pointed out that perceived noise is dependent upon the final enlargement and that images with more pixels, even though they be comprised of noisier pixels, might actually look less noisy because of the random nature of noise. (At least that's what I think he meant, but I've yet to see a practical, side by side comparison, so I can't get a measure of this.)

However, the Canon stable of DSLRs appear to be all equally noise-free, with subtle differences at certain ISO settings. The 1Ds , for example, has lower noise at the higher ISO settings as well as lower noise at long exposures of several seconds, but I've got the impression these improvements, whilst noticeable, are still relatively marginal.

Your Trapezium comparison of the D30 and 1Ds shows a chasm of a difference, yet the centre of interest is the same size. The over all image of the 1Ds is much larger, as you would expect, but the detail is the same size. I can't make sense of this comparison. Can you elaborate, please.
Logged
samirkharusi
Full Member
***
Offline Offline

Posts: 196


WWW
« Reply #5 on: April 21, 2003, 11:32:34 PM »
ReplyReply

Ray, both the D30 and the 1Ds images are 1:1 crops, nowhere near full-frame. The 1Ds image shows more of the nebula simply because it had a longer effective exposure. Unless I have cocked up somewhere the detail should look roughly similar is size (D30 pixels are 9.5 microns square, 1Ds 8.8 microns). Noise behaviour in these cameras is very different depending on exposure lengths. For astrophotography one needs both high ISO and loooong exposures. For landscapes we are talking about exposures that are well under a second long. For landscapes thermal noise should be almost a non-issue. Though if you shoot at 45 to 50 degree C (I do, it's hot here!) you do see a lot of noise even on landscapes. Tests at room temperature yield the same dark-frame noise histogram (20% along the x-axis at the back-of-camera histogram) at ISO 800 for the D30, D60, 10D, 1Ds with exposure lengths of 0.25 minutes, 2 minutes, 12 minutes and 27 minutes respectively. You can therefore see that in loooooong exposures the D30 is hopelessly noisy for shooting nebulae and galaxies. These require exposures up to hours long (somebody recently posted a shot with an exposure stack of 45 1Ds frames each 5 minutes long!). The D60 was the first, truly viable DSLR for such stuff, capable of giving superb results with the right equipment in the right hands. Canon has truly revolutionised long-exposure noise handling between the D30 and D60. What they do seems to be a closely-guarded secret. So far I have not managed to see even a semi-plausible explanation of how they do it. None of the usual camera-test forums publish data that are very pertinent to astrophotographically long exposures and I had to beg for data on these Canon DSLRs all over the world to get them. I finally compiled the above data from Florida, Hong Kong, Colorado and Oman (where I am located). No idea how the 14n would behave. Anyone who owns one, put on lenscap, set ISO to 800 and shoot exposures of 2 seconds, 4 sec, 8, 16, 32, 64,... doubling each time until the histogram toe gets to one-fifth up the x-axis. Tell us what the exposure length you got to. Perhaps it's an undiscovered gem for astrophotography? I suspect it to be more like a D30 (15 seconds)...
Logged

Bored? Peruse my website: Samir's Home
Paul Caldwell
Newbie
*
Offline Offline

Posts: 25


« Reply #6 on: April 24, 2003, 10:13:16 AM »
ReplyReply

I  most definately would not say I am not dependent on "automatic everything".

I work all my images as individual creations, and the process I use is very selective.

My point on the 1ds noise, is that out of the camera, my images shot considerable more noise than say my S2.  I am talking noise shown right after the raw conversion.  I fully understand that is you sharpern certain parts of a image, say the sky you can pull the existing noise out even more.  I have written several articles on selective sharpening for outbackphoto.com on this subject.

For me, the 1ds at 400 can produce a good image, but the shooting situation can have a  very large effect on this image.  If you are even slightly underexposed the amount of noise that is created is much larger.  Also the software used for the raw conversion is very very important.  The Canon software will produce considerably more noise in most situations than the Capture One software.  However Capture One is always using some noise reduction and I can see why.  

The 1ds also I find produces much more noise in the background areas of close up shots.  This is a simple fix with Neat image, but I am getting this sometimes even at ISO 100.

Paul Caldwell
Logged
Paul Caldwell
Newbie
*
Offline Offline

Posts: 25


« Reply #7 on: April 15, 2003, 02:59:32 PM »
ReplyReply

If you haven't already seen it, Rob Galbraith's review is up on the 14n.

http://www.robgalbraith.com/bins/multi_pag...cid=7-6131-6139


Good review and should be of interest to many.

Paul Caldwell
Logged
Andrew Richards
Guest
« Reply #8 on: April 16, 2003, 08:57:50 AM »
ReplyReply

'But until the 14n can produce vibrant, pleasing, printable colour in a variety of shooting situations, and until  noise processing is less harmful to detail in low contrast areas, this is not a camera we would be prepared to shoot with to get our work done. Even at the low ISO settings this camera will be best used at.'
Does'nt sound good does it?
Logged
sergio
Sr. Member
****
Offline Offline

Posts: 661


WWW
« Reply #9 on: April 16, 2003, 01:03:49 PM »
ReplyReply

I thought the 1Ds was noise free in the low ISO range, and very clean up to 800 ISO. At least according to Michael's review. Does this differ from camera to camera?
Logged

Paul Caldwell
Newbie
*
Offline Offline

Posts: 25


« Reply #10 on: April 16, 2003, 08:52:58 PM »
ReplyReply

Hello Michael,

I would have to say that based on your review and images, that yes you have a better ISO range than my camera.  My body past 400 is really not very useable unless its being used in very good light.  

There have been other posts to the same effect as mine, so its apparent that Canon's putting out cameras with a range of high ISO quality.  

As I mainly use ISO 50 or 100, its not a major issue for me, but there are times I need the higher shutter speeds oft'd by a higher ISO setting.

I also feel that the Capture One software, (Phase One) is taking more noise out of the image than the Canon code so if you are using the Capture One software, you are getting a image with considerably less noise.  Canon's own software especially at the ISO 250 and up is very noisy IMO.  Capture One is always using noise reduction, as even with the slider all the way to the left, you are set to L.

As I have said before, Thank God for Capture One.

I have now made many ISO 400 tests, and still not that pleased, but still learning the camera and have hope I will get better here.  The exposure setting at the higher ISO's is very important, and this is not as true with the ISO 100 or 160.  If you are off as much as 0.5 with my camera then the noise is considerably worse.  

I have considered sending it to Canon, but before I do, I am waiting till my local dealer gets in another 1ds and then will shoot it against mine.  If I see a really big diff. I will then consider sending it to Canon.  But can't do that till after the Spring is over.

Paul Caldwell
Logged
Guest
« Reply #11 on: April 17, 2003, 05:37:05 AM »
ReplyReply

Temperature can be an issue with sensor noise, but I haven't seen it make a huge difference with the 1Ds. I've used the camera at -32F in Quebec this past winter and +100F in Costa Rica and seen little variation at high ISO settings.

Michael
Logged
samirkharusi
Full Member
***
Offline Offline

Posts: 196


WWW
« Reply #12 on: April 21, 2003, 12:04:42 AM »
ReplyReply

Just wondering if people are comparing similar stuff. Noise is most vsible in uniform areas (like a pale blue sky or a uniform area in shade). It'll become progressively more visible with contrast stretching. One way to alleviate this is to use a higher threshold in unsharp masking. It's obvious that if one insists on using a threshold of zero (because that looks best in the main parts of a landscape) then the sky may start showing noise. It's also obvious that any extreme curves manipulation often needs to be done selectively (with masking). Perhaps people are just becoming too dependent on automating everything. Nobody says that you cannot deliberately gaussian blur a sky and simultaneously sharpen everything else. Anybody who is fussy enough about his high ISO efforts needs to put in the time to coax out maximum quality. While not exactly relevant to discussions on landscapes, I do have a couple of shots of the Trapezium in Orion (astronomical object), both 1:1 crops, one from a D30 at ISO1600, another from a 1Ds at ISO1250:
http://www.geocities.com/samirkharusi/m42.html
The 1Ds version is much more smooth, but both required my best efforts at combining blurring/sharpening/masking. For major noise just look at the D30 version...
Logged

Bored? Peruse my website: Samir's Home
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #13 on: April 21, 2003, 10:59:46 PM »
ReplyReply

Keep in mind that the D30 shot was taken at ISO 1600, and the 1Ds was taken at ISO 1250.
Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8878


« Reply #14 on: April 22, 2003, 10:32:40 AM »
ReplyReply

Quote
No idea how the 14n would behave. Anyone who owns one, put on lenscap, set ISO to 800 and shoot exposures of 2 seconds, 4 sec, 8, 16, 32, 64,... doubling each time until the histogram toe gets to one-fifth up the x-axis. Tell us what the exposure length you got to. Perhaps it's an undiscovered gem for astrophotography? I suspect it to be more like a D30 (15 seconds)...
Samir,
Not sure if you're being ironical but something tells me the 14n is not going to be an ideal choice for astrophotography.

I recall reading that the D60 had improved noise over the D30 at long exposures. Didn't realise the difference is so marked. Thanks for the explanation.
Logged
Guest
« Reply #15 on: April 24, 2003, 01:36:34 PM »
ReplyReply

Paul,

I see that you've been making the point here, as well as on other forums around the Net, that the 1Ds is somewhat noisy at 400 and below.

This is completely at odds with my experience with two cameras (a test sample and my own) and after having shot some 7,000 frames over 5 months. It also is not what I'm hearing from any other 1Ds owner.

My suggestion is that you return your camera to a Canon service center to have it checked. You should not be satisfied with anything less than essentially noise free images at up to ISO 400.

Michael
Logged
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad