High ISO Pictures from the Canon 1D Mark IV

[baronfizzy]

Although I'm very much against the over use of test charts and people posting pictures of bookshelves etc to "test" a camera's capabilities, here are some links to some images I shot with the new 1D Mark IV while sitting at my desk, as well as one that I took earlier today while walking through the park. Even though the reason for the walk was to get the camera shop to look at the centre focus point, which seemed to be too far to the left, when I looked at these images later, I was absolutely overjoyed at the high ISO performance. I haven't done anything to the shots, really, apart from convert them to JPEG with the software provided. Here are the links:





http://fightshots.com/DPP_0001.JPG


http://fightshots.com/DPP_0003.JPG


http://fightshots.com/DPP_0005.JPG


http://fightshots.com/DPP_0006.JPG


http://fightshots.com/aucklanddomaintree.JPG





Check the exif on the first ones - I had to double check when I saw them too ...

[Dave Johnston]

Very Impressive noise level at 12800. Thanks for sharing.

[freelanceshots]

<meta http-equiv="CONTENT-TYPE" content="text/html; charset=utf-8" />
<title></title>
<meta name="GENERATOR" content="OpenOffice.org 3.1 (Win32)" />
<style type="text/css"]<!--
&lt;!
@page { margin: 0.79in }
P { margin-bottom: 0.08in }
&gt;
--></style>

<p style="margin-bottom: 0in;"]I was checking out your images and by
the way thanks for posting them but I don't think these shots under
those conditions can accurately show how much improved the ISO noise
levels are on the new 1D series Mark IV. Digital noise does not
really effect image quality when there is a decent amount of light
available or when there is a higher quality of light available. To
really show how image quality is effected at higher ISO levels one
needs to take the camera body into a dark gym with really poor
tungsten lighting or a high profile event held in a barley lit event
hall or church, just some examples. To show this point I just went
out and shot an image at ISO 12800 with the 5DII and the image looks
pretty respectable at this setting because theres more then enough
light before the sensor starts to struggle. It is a pretty
gloomy day here where I shot this first image as the sun was going
down but there was still a decent amount of light left but dark and in the shadows. Now if
I where at ISO 12800 in manual mode where I was at an aperture of f/4
or 5.6 and I was squeaking out a hand holdable 1/50 of a second or
slower exposure just to reach the point of getting a workable image
then there will be a lot more noise present and less detail. The info
on the first image with the rust and cracks is Tv 1/4000, Av 5.6, ISO
12800. The second image with the sport bike is Tv 1/10, Av 5.6, ISO
12800 with much less light to work with hence more noise present in the image.
<p style="margin-bottom: 0in;"]
<p style="margin-bottom: 0in;"]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/test-image-12800-ISO.JPG
<p style="margin-bottom: 0in;"]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/sportbike-ISO-12800.JPG

[baronfizzy]

About your statement: NO SHIT. These images were taken under EXTREMELY low light conditions. You sound like one of the APS-H conditions and to lecture me on signal to noise issues is insulting. I pretty much constantly shoot boxing matches under terrible lighting conditions. There was a lot of hatred from die hard 5D Mark II users when the 7D came out - even though the noise on those was more pleasing and random than any DSLR before it. Are we going to see the same thing again now that this groundbreaking camera is out? Your statement reads like complete nonsense to me.

[baronfizzy]

For some reason, some of my reply was lost. Your image of the motorbike was very well lit compared to any of mine, which were taken at night. By your own admission, you kind of cancel out your argument. But I guess, that's the way it goes when people are so devoted to a camera as people are to their full frame DSLRs.

[freelanceshots]

I did not mean to get you upset or say that you where not a good photographer. I could care less if I was shooting with a rebel XTI, 40D, 7D or the Canon EOS 1D series camera as long as I am getting the best results. I don't care to pay those games with whos got the better camera as I got this 5D II bought for me to do the job this company wanted. You can get all hot and bothered as I can't control that but all I'm really interested in is seeing the image quality the 1D IV can produce and if it way beter then the Mark III series. I was just making a point as your images that you posted did not look really any better then the 5D's maybe because of the light that was available. If I'm going to spring a lot more money and get a 1D IV in the near future I want to see a substantial gain in the image quality.

[baronfizzy]

I didn't take your statement as meaning that I wasn't a good photographer - it's just irritating when people point out the obvious. I'm also not concerned about "who has a better camera", as I'm sure there are people out there who now have full frame and a crop sensor. The main point of my original post is this: The fact that a crop sensor camera can now match (and from what I've seen from the motorbike picture, exceed) the high ISO quality of a full frame DSLR is pretty exciting for anybody who required the speed and auto focus accuracy of a 1 series body, but had held off buying one because of the incredible image quality of the 5D Mark ii. I was going to buy a 5D Mark ii but was put off by the terrible auto focus and slowness of that camera. Now there's an option for people who require both these things in the one camera.

[freelanceshots]

With my motorbike picture it was not lit very well when compared to the outside image that I took. With that image I had all the blinds in my home shut where the sun had set and the TV was on. I just shot that image becasue it was right infront of me as I wrote my post. You should be able to tell by my camera setting that there was not a lot of light as I was at ISO 12800 with the f stop set at 5.6. and the shutter speed at 1/10. Cameras can take dark pictures and make them appear brighter and that is when the noise increases. This post was not about which camera is better as a 1D series camera is clearly the winner with the overall package but not everybody can spend 4 to 5 grand on a camera body.

[baronfizzy]

No of course not - there normally has to be some kind of commercial use to justify the purchase. I'm sure Canon doesn't care about our opinions, since both the 1 and 5 series cameras are going to sell regardless of what we think. Whether or not they care that the centre focus point on mine is off centre is something that I'm going to find out when their customer support comes back after the Christmas break &gt;:(

[freelanceshots]

<meta http-equiv="CONTENT-TYPE" content="text/html; charset=utf-8" />
<title></title>
<meta name="GENERATOR" content="OpenOffice.org 3.1 (Win32)" />
<style type="text/css"]
&lt;!--
@page { margin: 0.79in }
P { margin-bottom: 0.08in }
--&gt;
</style>
<p style="margin-bottom: 0in;"]First off the 5D's auto focus works
great for the area of photography that is was designed for. I would
never think that I was going to try and capture the front page image
for a mag or newspaper from an NFL game with a 5D2. That's not
what the 5D2 was designed to do in the first place. As far as using
the 5D2's auto focus and frame rate to capture a award wining photo
at a sports related event or motorsports related event it can easily
be done. Most of the time its the photographer that makes the
pictures and not just the tool. The overall point that I understand
is why not use the tool that can best assist you in your job and in
turn makes your job easier which of course would be a 1D series body
or even the new 7D in that instance. When it comes to taking
advantage of Canon's wider L series lenses the full frame sensor is
the only option and that is where the 5D2 shines. In the image that I
took of the motorcycle I did nothing to make that picture come out
good. I just took a simple, quick shot to show the differences in the
lighting vs noise issue. There were plenty of simple thing that I
could have done to make that particular image have less noise at even
12800 ISO. As for shooting with different cameras I shot with a 40D
for two years where that is still an excellent camera body. I
continue to use it for the shots that I need where the 5D's
performance falls short.

[Sinh Nhut Nguyen]

Next time you guys do a noise level comparision, may I suggest shooting the scene at ISO 100 and then shootthe samescene againat ISO 3200, 6400 and 12800

[greggf]

hey Baron, just out of curiosity....where are you located? I've only seen the MarkIV released in Australia, and Norway(DPReview users). The posts on there seem to correspond with what you're saying also. One gentleman posted several shots using AI Servo of his dogs running towards him, with several bursts, including a 32 shot burst!! He said only 2 out of that series were OOF, and blamed them on trying to maintain focus on the dogs head with center point, while the dog was running towards him. Here is the link to his photos...gallery.mac.com/panfinnse (http://gallery.mac.com/panfinnsen]http://<span class="domain)


I have the MarkIV on order through Amazon, and haven't heard a word yet??!! I preordered on Oct 19 at 10pm PST, so I am patiently anticipating its arrival!!

[freelanceshots]

I'm up for a more controlled side by side sensor comparrison. It might be helpful to see real world sensor comparisons at the selected ISO settings. The images would just need to have comparable subjects, the same camera settings with similar light where tripods would need to be used of course. In the long run I would hope that a 5,000 dollar investment would return higher quality results then the 2,700 investment.

[baronfizzy]

Regarding the different ISO tests: I'm also interested in seeing shots at different ISO's, since there seems to be a lot of emphasis placed on high ISO. I'm guilty of this now, of course, but I only posted them because I was so surprised. I can post some lower ISO stuff (not at 100, since nobody really wants to shoot a DSLR at that level) when I get the chance, since it's unfair to not show this.





To the above person: I'm in New Zealand, so like Australia, we're part of the Asia/Pacific region that would have gotten the camera first.

[freelanceshots]

<meta http-equiv="CONTENT-TYPE" content="text/html; charset=utf-8" />
<title></title>
<meta name="GENERATOR" content="OpenOffice.org 3.1 (Win32)" />
<style type="text/css"]
&lt;!--
@page { margin: 0.79in }
P { margin-bottom: 0.08in }
--&gt;
</style>
<p style="margin-bottom: 0in;"]Another issue that would have to be the
same would be the upload size and quality number. My sample images
where uploaded at 72dpi and where approx 1000 x 664 with a quality
setting of 9 was used. The original images uploaded by baronfizzy
where pretty slow to open so I can only assume that they where saved
at much higher quality settings. Images would need to be
uploaded to the digital picture website for review. I would say that
72dpi and 1000 px longest side with a save quality number for the web
at 70 percent using photoshop, DPP or Lightroom ect would be the
standard. Sample shots would need to be of something with high
detail but not in the direct sunlight. Something out of the direct
sunlight if possible. Maybe use 70-200mm at 100mm where the image
would have approx 2 foot x 3 foot rectangle. I'm thinking an outdoor
images in the early evening time would be the easiest. Aperture set
at f/8, ISO 800, 1600, 3200, 6400 and then 12800.

[freelanceshots]

Lets also throw a 25600 ISO into to mix as well for comparison.

[Todd Ovick]

Next time you guys do a noise level comparision, may I suggest shooting the scene at ISO 100 and then shootthe samescene againat ISO 3200, 6400 and 12800
<div style="clear: both;"]</div>





hehehehehehehe,...I will have to 2nd that!!!!!!

[freelanceshots]

<meta http-equiv="CONTENT-TYPE" content="text/html; charset=utf-8" />
<title></title>
<meta name="GENERATOR" content="OpenOffice.org 3.1 (Win32)" />
<style type="text/css"]
&lt;!--
@page { margin: 0.79in }
P { margin-bottom: 0.08in }
--&gt;
</style>
<p style="margin-bottom: 0in;"]
<meta http-equiv="CONTENT-TYPE" content="text/html; charset=utf-8" />
<title></title>
<meta name="GENERATOR" content="OpenOffice.org 3.1 (Win32)" />
<style type="text/css"]
&lt;!--
@page { margin: 0.79in }
P { margin-bottom: 0.08in }
--&gt;
</style>
<p style="margin-bottom: 0in;"]The real test when it comes to noise
production and noise control is when the sensor has to in a sense
create light by over-extending the exposure. A perfect example that I
can contest to is shooting an interior room for commercial use that's
not well lit. In this case you just simply can't put in brighter
bulbs in the lamps or use strobes as this will blow out certain areas
of a medium to large room plus cast blown out highlights on the walls
and ceiling and create shadows. Even if you shoot multiple exposures
and blend them together theres only so far that you can go. I
see images shot on a daily basis at ISO 200 that have huge amounts of
noise/loss of detail even with the full frame sensor on the 5D2. The
reason for the higher noise content even while shooting at ISO 200 is
because of the other settings that are required. ISO 200, f/11 to
f/16 which sometimes requires a 15 to 25 sec shutter speeds breeds
noise. This combination of settings with low light and long exposure
is the true test for any sensor. A Canon 40D, 50D, and even a
7D in this sense would have even more noise where it would just not
be at an acceptable level plus you'd have to use a 10-12mm lens where
you'd end up cropping more then half the image away after you done
correcting it's perspective to meet publication standards. I want to
see true, "no holds barred" noise test used on professional
equipment. The 1D series of camera is strictly a professional camera
with a professional price tag so it needs to see harsh conditions
which has nothing to do with weather sealing.

[freelanceshots]

Image of my dog at ISO 25600, 35mm, Tv 1/100, Av 2.8, entire room lit by one 40 watt energy saver bulb in a small lamp in the living room. He had a food allergy reaction where he developed bumps all over his back today so that does not count as noise.


/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/My_2D00_dog_2D00_ISO_2D00_25600.jpg.

[Fast Glass]

Could you post a 100% crop at 102400 iso?


By the way you'd be the firstperson to do that[;)]

[freelanceshots]

I took some sample ISO images using the 5D Mark II today where the details are as follows:


150mm vocal length, Sharpness 5, High ISO speed noise reduction LOW, aperture Priority AE, TV variable, AV 8.0 for all images, Auto white balance, manual focus


Images taken right after the sun had set where there is a wide range from the shadows to highlights


ISO Levels 100, 800, 1600, 6400, 12800, 25600


--In the lower fairing on the bike what looks like noise is the reflection of the concrete in the beginning--


/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/ISO_2D00_100.jpg


ISO 100 above


/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/ISO_2D00_800.jpg


ISO 800 above


/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/ISO_2D00_1600.jpg


ISO 1600 above


/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/ISO_2D00_6400.jpg


ISO 6400 above


/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/ISO_2D00_12800.jpg


ISO 12800 above


/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/ISO_2D00_25600.jpg


ISO 25600 above

[freelanceshots]

Forgot to add:


Images uploaded at 72 dpi, 800px X 533px, save for web at 70 percent using Photoshop CS4


-No additional changes made to images-

[freelanceshots]

I used the black bike for the ISO test above as black objects in lower light with higher ISO's always produce more noise. Here is the
same image as above but with better lighting where a more pleasing white balance was
used but I still shot it at ISO 12800./cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.32.66/sportbike_2D00_web.jpg