40D vs 50D at high iso

[Steve Foss]

Ken and I were goofing around this afternoon. I put my 70-200 f2.8L on the 50D, and he put his 70-200 f2.8L IS on the 40D. We duplicated as closely as possible metering, focal length, framing and aperture to compare the high iso performance of both cameras.

Note: The in-camera noise reduction feature on the 50D was set on strong.

You won't see banding because banding rarely shows up unless there are larger expanses of fairly consistent color/texture.

These were shot large jpeg. There was no pp done. They were simply resized to 750 pixels across and put online. What we didn't think about was in-camera picture style settings, so if there are some differences, it's possible the 40D and 50D had different picture style parameters selected. Both cameras were on automatic white balance.

40D iso800

50D iso800

40D at iso1600

50D at iso1600

40D at iso3200

50D at iso3200

50D at iso6400

50D at iso12800

[finnbay]

Parameters the same, Steve.

[finnbay]

My first general impression is that I sure like the 50D. Some saturation differences and contrast differences. Lenses were the same except that mine has IS. Technically speaking, I don't know a lot about sensors, but could they cause these differences, or is it more the processor?

[MN Shutterbug]

The thing I noticed right away is the shallower dof with the 40D.

[polarsusd81]

That is what I noticed as well Mike.

[Steve Foss]

The images from both cameras also have different color casts.

[polarsusd81]

Yeah, they do, but it is not near as noticeable as the DOF differences.

[finnbay]

I believe Dan had mentioned somewhere, either on the forum in a post, or possibly in an email to me, that the temperature of his 50D seemed to be a little on the cool side.

[Steve Foss]

The 50D images were at f4.5, the 40D at f4. Both at 200mm and the same distance from the subject. Not enough aperture difference (1/3 stop) to cause that much difference in apparent DOF.

Mmmmmmmm, a mystery!

[Dbl]

Yeah I did mention that somewhere Ken. Definite blue cast to the 50D. I don't like how AWB handles shadows or shade.

[Steve Foss]

Dan, I wonder if the firmware upgrade will tackle the cool temp on the 50D on AWB. Have you shot it on any of the other WB presets or set WB manually yet? If so, does the cool temp prevail there too?

If it's cool across the board, that'll mean one more step in pp for every image I take if I buy the body since I like things a bit warmer.

[buzzsaw]

The images from both cameras also have different color casts.

So auto WB is different between the two, maybe set them both to 5500K and fire away.

[MARINERMAGNUM]

Interesting. Except for the high iso perf. and 15 mp,the 40d looks just as good if not better in some shots.

I guess for me it boils down to doing noise reduction in PP or doing it in camera [50d]

Albeit they are supposed to be smaller/better pixels,but I'm just wondering if Canon pushed it too far with 15 mp on this small sensor?

I saw some bird images posted on the site with a d&p in the name-from the 50d- at iso 3200,and they looked pretty noisey.

I wish I wasn't so in love with the 50d's LCD!

From what I can gather,the 50d seems to be an attempt at doubling the mp,without increasing the noise-which is GOOD,but if you don't need the extra mp????

I just hope Canon isn't pushing the point and shoot nonsense on the SLRs,trying to gain market by going with the most pixels. I would like to hear other input.

[Steve Foss]

Buzz, that was my thinking for next time. The subjects also don't show noise very easily because of all the detail and texture.

MM, with the bird image you mentioned, a lot depends on whether the in-camera noise reduction was on and what it was set on, and whether any further NR was done in post processing.

The 50D seems to produce images with roughly the same degree of noise as the 5D, and since I've used the 5D with great results at iso3200, I like the 50D because of how affordable it is and because of the tremendous number of other added options on the 50D not found on a used 5D.

So it's pretty certain I'll be picking up a 50D instead of the used 5D later this year.

Now, that's just my thinking on it, and different priorities dictate different needs.

[Dbl]

Steve, the cool temp I'm noticing is ONLY on AWB in the shade. Otherwise the camera seems to do well with AWB. I tried shooting with the WB set to shade or cloudy and that warmed the shots up enough.

Mike I do shoot K temps a lot, but generally only indoors. Outdoors I've found you don't get consistent results unless you are using flash with K temp. A constant K temp is tough to nail down, I know the 5500K is daylight balanced but I've found you need to go up or down from that under different lighting conditions to get good results. You would be surprised how the color of the light changes from one spot to another, color casts from surrounding grass, trees, etc.

The Mark IIN handles AWB generally better than a custom white balance for a lot of indoor situations. I don't always care for it outdoors, just like the 50D it can cool shots down in the shade.

I did do a custom white balance with the 50D a week ago and it gave me consistent results, so I don't think there are any worries there.

I have not had a chance to do any shooting after the firmware upgrade, I did do a shoot just before it in the shade but used a strobe so I didn't notice much cooling there.

For me shooting from sun to shade all in a few feet I will make a few small tweaks to warm up the shot, just like the Mark IIN. If you are shooting in the shade with people or groups it would be no problem to make the appropriate adjustments to warm up your images.

If I get a chance latter I will try and find a shot or two that shows the difference.

[Steve Foss]

Thanks, Dan.

It should be noted that the shots in this comparison posted above were in the shade.

[Dbl]

MM there is a definite difference in the files from the 50D and I mean in a good way. I have a friend with a 40D and I sent him some files to play with and he could not believe the difference. Plus I'm not a mega pixel guy but having all those pixels to play with is unbelievable. I shot some large group photos last week with the 50D and what a difference from the 30D and Mark IIN. I was stunned and that was at the smallest RAW setting. I love having all that information available if I want to use it.

Here is the deal, the site with the d and the p has people spending way too much time pixel peeping. I will be honest this is an amazing camera for what you are paying. Think about it $1400 for this camera and its capabilities and I understand there are some price drops out there at a few places.

I've looked at many of those shots and honestly I don't put a lot of faith in how they were shot. I've put 2,000 shots through the camera in a variety of conditions. I'm impressed and it takes a lot to do that.

Is it perfect, nope but I've not shot that camera yet. Is it a step up from a 40D, yeah I think it is and my results prove it will provide very good images at 6400 and even 12800 in a pinch. I guess that is the bottom line, I'm happy and it takes a lot to make me happy with equipment.

[finnbay]

Thanks, Dan! All info that can help tweak shooter performance as well as camera performance. Haven't tried to upgrade yet. Hopefully I'll have some time this weekend.

[Dbl]

Ken I had the firmware sitting on my CF card but didn't do it before a shoot I had Wednesday evening. Something about bricking my camera just before a gig didn't give me a warm fuzzy feeling.

I've never had a problem with a firmware upgrade but I was thinking with my luck this would be the first time. I am anxious to see what if any difference they made to noise profiles.

[Steve Foss]

Dan, what exactly about the image quality of the native 50D file is better than previous images in the 20D/30D/40D sequence? Is it a matter of sharpness, saturation, contrast, vividness? How would you characterize it?

[MARINERMAGNUM]

DP's review says it has more noise than a 40D?

[Steve Foss]

Is that with the in-camera noise reduction on the 50D turned on or off? Using Ken's 50D on iso3200 with in-camera noise reduction on "strong," the images are quite a bit less noisy than my 30D at the same iso.

[Steve Foss]

OK, I went in and read the review. They were comparing the two bodies with no in-camera noise reduction turned on in the 50D.

They called that as close as they could get to a "level playing field," and if a direct comparison of sensor noise upon capture is the goal, it is level.

However, the 50D produces its excellent noise performance through that in-camera NR, and to remove that option from the comparison means it's not a real-world comparison, because out here in the real world (with all due respect to DP Review's pixel peepers) we use the features the camera offers.

[MARINERMAGNUM]

Steve, I agree wholeheartily. It kinda confirms my suspicions though,I just think 15mp on a crop is beyond the limit unless some kind of NR is applied-whether in-camera or PP. I'm afraid Canon is falling into the trap of the MP marketing race-bigtime. It's hard to bend the laws of physics-something has to give. Reminds me of my kids making Kool-Aid:2 cups of sugar is good-let's dump in 4 cups! it outta be awesome! eek!yuk!. Without the in camera NR, I suspect at iso 3200 it would be a mess of noise. The images they posted in the test looked like-to me anyway- that the 40D had better detail and clarity. Alot of folks say that it's hard to get all that any good slr can deliver unless you use good quality glass-a good argument. I think 10-12mp is about the limit,with 10 being the ideal number you can squeeze on a 1.6 and not have to run "nitrous oxide" to keep the noise tamed down. Just like the 70-200 2.8 lenses are a "sweet spot" in the sharp zooms,I think 10mp is the "sweet spot" for a 1.6 sensor.I'm not bashing the 50D,I'd love to try one for a day,but it seems the ability to crop deeper would just intensify the noise or the loss of detail that the NR induces.

Cropping into an image really sets off the noise IMO. I remember seeing some deer images that jimalm posted on here with his 30d that were at iso 1600,and by not cropping and just sizing down,the noise was invisible-I was amazed.

The LCD on the 50d will certainly take "chimping" to a higher level!

After seeing what kind of work from you and others can be produced with the 20D, it's hard to see a "must upgrade" scenario in my future. If I do upgrade,it will prolly be a 40D. Seems to be a well proven camera that no one has a bad word to say about,and seeing them new for $700 on evilbaay,is very tempting!

I sure love talking about the gear,and value everyone's opinions and truly take them to heart.

[Steve Foss]

Well, the real world IQ test the matters most is how they print, and based on what I've seen so far (I shot Ken's 50D down to iso800 with NR on "strong", the 50D has better high iso IQ on the screen than my 30D, and I'm willing to bet it'll print better, too.

My main point, I guess, is that it doesn't do any good to compare the 50D (with in-camera NR off) with any other camera because Canon put the 15 Mp on that sensor calculating that the feature would be used. I don't disagree that the sweet spot for that size sensor would be 10-12 Mp with no in-camera NR, but I don't think the distinction has much value in the real photographic world. Like putting an old pickup next to a new one with optional power steering and leaving the power steering off on the new pickup when comparing how they both steer. You can do it, but why would you?

What I'd like to do next time I'm out with Ken is shoot some with his 50D at iso100-400 with NR on strong and identical shots with my 30D with the same iso and lens and see which looks best, both on screen and printed. I mean, all the talk about its sweet high iso performance (again with NR on) is good stuff, and that's a great feature to have, but in the woods and at outdoor weddings and portrait shoots, I'll still be at iso100 to 400 almost all the time, and that's where I want to make the comparisons. I've also got quite a lot of work done with the 5D at those iso settings, so I'll be comparing the 50D to those as well. The stuff I've got from the 50D at iso3200 with the NR on strong is slightly less noisy than what I shot with the 5D.

Lots of fun to run things through the wringer and see what comes out the other end!