5D vs 5DM3 High ISO

Home · Register · Search · Software · Hosting · Software · Join Upload & Sell

FM Forum Rules
Canon SLRs, primes, and zooms lenses reviews

FM Forums \| Canon-mount SLRs \| Join Upload & Sell
1 2 3 4 5 end

5D vs 5DM3 High ISO

vscd
Offline
•
Upload & Sell: Off

p.5 #1 · 5D vs 5DM3 High ISO

Of course you're right. I just thought 0.2EV on T-Stop, 0.33EV on ISO and maybe another quarter Step on the physical part of the low resolution. This all together is an interesting enhancement... and maybe some marketing-fanboys won't tell you

The cam is >7 years old by now and still a great tool. The prices are around 700-800 Euro at the moment, I guess (used ones). So if someone really needs faster ISO than ISO1600 (and I know there are) the 5D is out, but for the Rest there is no need to upgrade. Just look at those examples: ISO Comparision

For example compare Nikon D3 vs 5Dc @ISO800. And we speak about the almighty D3!

Edit: I didn't even know that there is a difference between the Canon Pro and Non-Pro line regarding real 1/3 stops on iso-implementation:
- [...]Lower end Canon models do not perform analog amplification for the intermediate ISO's, rather the intermediate ISO's are implemented by a multiplication of the raw data in software after quantization, and there is only a single stage amplification in hardware; strictly speaking, they do not have intermediate ISO amplification. [...] Source: LINK

Jan 07, 2013 at 01:06 AM

vscd
Offline
•
Upload & Sell: Off

p.5 #2 · 5D vs 5DM3 High ISO

I just tried to push the 5Dc to the highest ISO Mode it could produce out of the box to be curious about the results: ISO12800. You can't do this on the 5D Mark1? You can emulate...it's a little bit awkward but easy, I just choosed ISO3200 (H) with manual -2EV. The results were of course darker than in real but got pushed by +2EV afterwards (in ACDSee 5 Pro). The results were quite useable for the really dark scene I choosed.

As 3rd step I uses NeatImage to reduce the noise with the defaultscheme... I didn't try it in good light at the moment but the results in bright light would be at least better than in this nightscene

If the shutter is fast enough.

I uploaded the results at LINK and compared it with a picture from the same scene with no manipulation. I think the results of ISO12800 can be used in some circumstances. I've seen worse.

Jan 09, 2013 at 02:01 AM

15Bit
Offline
• • • •
Upload & Sell: Off

p.5 #3 · 5D vs 5DM3 High ISO

I'd say that the 5DII at ISO6400 holds up well against the 5D at ISO 1600 in those shots. A touch more noise perhaps, but all the detail is still there to my eyes. I agree with Lars though - downsizing the 5DII image to fit the 5D imparts an "incorrect" fairness. You should upsize the 5D shot.

May 20, 2013 at 05:15 PM

snapsy
Online
• • • •
Upload & Sell: On

p.5 #4 · 5D vs 5DM3 High ISO

15Bit wrote:
I'd say that the 5DII at ISO6400 holds up well against the 5D at ISO 1600 in those shots. A touch more noise perhaps, but all the detail is still there to my eyes. I agree with Lars though - downsizing the 5DII image to fit the 5D imparts an "incorrect" fairness. You should upsize the 5D shot.

IMO downsampling is the preferred way to compare High ISO because most don't print/view such images larger than 8x10.

May 20, 2013 at 07:09 PM

Pixel Perfect
Offline
• • • • • •
Upload & Sell: On

p.5 #5 · 5D vs 5DM3 High ISO

Looks to me to be around 1.5 stops improvement.

May 20, 2013 at 11:57 PM

dehowie
Offline
• • •
Upload & Sell: Off

p.5 #6 · 5D vs 5DM3 High ISO

Why are they downsized to 8mp not uprezzed to 21?
This topic of down rezzing images has been done to death and they should be uprezzed to maintain the detail advantage of the higher rez sensor.
Any conclusions based off down rezzing are inaccurate at best.
That being said three stops difference is optimistic at best.

May 21, 2013 at 12:59 AM

lucas lumiere
Offline
• •
Upload & Sell: Off

p.5 #7 · 5D vs 5DM3 High ISO

I prefer to judge the difference in ISO performance by what I'm actually willing to use. I never liked going higher than 1600 on the classic, but with the III I will shoot 6400 willingly and even 16,000 in a pinch. A nice upgrade.

May 21, 2013 at 01:37 AM

snapsy
Online
• • • •
Upload & Sell: On

p.5 #8 · 5D vs 5DM3 High ISO

dehowie wrote:
Why are they downsized to 8mp not uprezzed to 21?
This topic of down rezzing images has been done to death and they should be uprezzed to maintain the detail advantage of the higher rez sensor.
Any conclusions based off down rezzing are inaccurate at best.
That being said three stops difference is optimistic at best.

Because High ISO images aren't usually printed huge. And the detail advantage of the higher density sensor is not lost in downsampling.

May 21, 2013 at 01:40 AM

Beni
Offline
• • • • •
Upload & Sell: On

p.5 #9 · 5D vs 5DM3 High ISO

Not sure I'd agree with that, for us wedding photographers we are printing high iso images on large album spreads all the time. Last album I did had a ring shot from the ceremony, shot at iso 3200, printed across a double spread, 28" wide in total (was a 10X14" album, not the largest album size by far). Viewed far closer than a similar sized print would be on a wall. This is standard never mind the wall prints, etc.

In any case, now that I have a 5D3 and shooting it alongside my 5Dc, processing in ACR, 1.5 stop advantage at the luminance noise level, 3 stop at the colour noise level but more than that, iso 6400 (highest I've played with) is as malleable as iso 1600 on the 5Dc. You can push and play with the files and the colour, tonality and DR is still fine. You can't say that about iso 3200 on the 5D which falls apart if you look at it the wrong way and whose colours are all wrong.

May 21, 2013 at 09:21 AM

morganb4
Offline
• • • • •
Upload & Sell: On

p.5 #10 · 5D vs 5DM3 High ISO

3200 on the 5Dc was (to me at least) a cow. I tried it once and thought 'your kidding, why is that setting even available' and never went back.

Perhaps I should have just persevered... 1600 seemed ok to me, still does.

May 21, 2013 at 09:53 AM

snapsy
Online
• • • •
Upload & Sell: On

p.5 #11 · 5D vs 5DM3 High ISO

Beni wrote:
Not sure I'd agree with that, for us wedding photographers we are printing high iso images on large album spreads all the time. Last album I did had a ring shot from the ceremony, shot at iso 3200, printed across a double spread, 28" wide in total (was a 10X14" album, not the largest album size by far). Viewed far closer than a similar sized print would be on a wall. This is standard never mind the wall prints, etc.

In any case, now that I have a 5D3 and shooting it alongside my 5Dc, processing in ACR, 1.5 stop advantage at the luminance noise level, 3 stop at the colour noise level but more than that, iso 6400 (highest I've played with) is as malleable as iso 1600 on the 5Dc. You can push and play with the files and the colour, tonality and DR is still fine. You can't say that about iso 3200 on the 5D which falls apart if you look at it the wrong way and whose colours are all wrong.

What percentage of your ISO 3200+ shots are printed that large? And more specifically, what percentage where the subject doesn't fill the frame like a ring shot, where the extra resolution isn't needed.

May 21, 2013 at 12:37 PM

rrxjon
Offline
• •
Upload & Sell: On

p.5 #12 · 5D vs 5DM3 High ISO

Beni wrote:
Not sure I'd agree with that, for us wedding photographers we are printing high iso images on large album spreads all the time. Last album I did had a ring shot from the ceremony, shot at iso 3200, printed across a double spread, 28" wide in total (was a 10X14" album, not the largest album size by far). Viewed far closer than a similar sized print would be on a wall. This is standard never mind the wall prints, etc.

In any case, now that I have a 5D3 and shooting it alongside my 5Dc, processing in ACR, 1.5 stop advantage at the luminance noise level, 3 stop at the colour noise level but more than that, iso 6400 (highest I've played with) is as malleable as iso 1600 on the 5Dc. You can push and play with the files and the colour, tonality and DR is still fine. You can't say that about iso 3200 on the 5D which falls apart if you look at it the wrong way and whose colours are all wrong.

This. And I don't hesitate to go higher than 6400iso.

May 21, 2013 at 02:47 PM

goosemang
Offline
• • •
Upload & Sell: Off

p.5 #13 · 5D vs 5DM3 High ISO

snapsy wrote:

Because High ISO images aren't usually printed huge.

why?

May 21, 2013 at 02:54 PM

Beni
Offline
• • • • •
Upload & Sell: On

p.5 #14 · 5D vs 5DM3 High ISO

snapsy wrote:
Beni wrote:
Not sure I'd agree with that, for us wedding photographers we are printing high iso images on large album spreads all the time. Last album I did had a ring shot from the ceremony, shot at iso 3200, printed across a double spread, 28" wide in total (was a 10X14" album, not the largest album size by far). Viewed far closer than a similar sized print would be on a wall. This is standard never mind the wall prints, etc.

In any case, now that I have a 5D3 and shooting it alongside my 5Dc, processing in ACR, 1.5 stop advantage at the luminance noise level, 3 stop at the colour noise level but more than that, iso 6400 (highest I've played with) is as malleable as iso 1600 on the 5Dc. You can push and play with the files and the colour, tonality and DR is still fine. You can't say that about iso 3200 on the 5D which falls apart if you look at it the wrong way and whose colours are all wrong.

What percentage of your ISO 3200+ shots are printed that large? And more specifically, what percentage where the subject doesn't fill the frame like a ring shot, where the extra resolution isn't needed.

Even half of that size, i.e. a single spread is still 14X10" printed at 300dpi. I would say that 50% plus of the pictures in an album are iso 1600 or higher. As for resolution, I believe that with this as well as the size thing you are trying to force your personal needs on the wider photographic community. Express it as an opinion not a fact.

May 21, 2013 at 03:21 PM

snapsy
Online
• • • •
Upload & Sell: On

p.5 #15 · 5D vs 5DM3 High ISO

goosemang wrote:
snapsy wrote:

Because High ISO images aren't usually printed huge.

why?

Noise, and more specifically, subject-to-noise ratio, with a few exceptions. If you have a subject that fills a large percentage of the frame (30% or more), then that image can tolerate more NR while still yielding acceptable detail and so printing large is workable. Or in cases where the aesthetic of the image allows for noise (or NR) then that is workable as well. But for images where subject details are somewhat important and only represent a small percentage of the frame then the SNR at native pixel sizes even on the best sensors is still lacking.

And important point not considered by most is that High ISO is almost always used in shutter-speed limited situations, where to achieve an acceptable level of sharpness for a given DOF the photographer has to increase the ISO to achieve an acceptable level of output brightness. However to benefit from a higher-density sensor you must use a higher shutter speed relative to a lower-density sensor, otherwise you're just oversampling motion blur. It's partly for this reason that sensor-area SNR and downsampling are the right ways to compare High ISO noise performance between sensors.

May 21, 2013 at 03:34 PM

snapsy
Online
• • • •
Upload & Sell: On

p.5 #16 · 5D vs 5DM3 High ISO

Beni wrote:
Even half of that size, i.e. a single spread is still 14X10" printed at 300dpi. I would say that 50% plus of the pictures in an album are iso 1600 or higher. As for resolution, I believe that with this as well as the size thing you are trying to force your personal needs on the wider photographic community. Express it as an opinion not a fact.

I'd agree that one's tolerance for noise in a given image is subjective but objectively there is more noise in lower-exposure situations that require High ISO and thus fewer opportunities where a large print would be acceptable, again depending on individual tastes but nevertheless the needle moves.

May 21, 2013 at 03:37 PM

goosemang
Offline
• • •
Upload & Sell: Off

p.5 #17 · 5D vs 5DM3 High ISO

snapsy wrote:
goosemang wrote:
snapsy wrote:

Because High ISO images aren't usually printed huge.

why?

Noise, and more specifically, subject-to-noise ratio, with a few exceptions. If you have a subject that fills a large percentage of the frame (30% or more), then that image can tolerate more NR while still yielding acceptable detail and so printing large is workable. Or in cases where the aesthetic of the image allows for noise (or NR) then that is workable as well. But for images where subject details are somewhat important and only represent a small percentage of the frame then the SNR at native pixel sizes even on the best sensors is still lacking.

And important point not considered by most is that High ISO is almost always used in shutter-speed limited situations, where to achieve an acceptable level of sharpness for a given DOF the photographer has to increase the ISO to achieve an acceptable level of output brightness. However to benefit from a higher-density sensor you must use a higher shutter speed relative to a lower-density sensor, otherwise you're just oversampling motion blur. It's partly for this reason that sensor-area SNR and downsampling are the right ways to compare High ISO noise performance between sensors.

what

May 21, 2013 at 03:47 PM

goosemang
Offline
• • •
Upload & Sell: Off

p.5 #18 · 5D vs 5DM3 High ISO

i mean sorry, i'm being kind of flippant

but what i want to point out here is how different individual workflows are. for instance mine and yours, based on what you've written there, are from completely different mindsets. and that's cool. but i never think of any of that stuff when i shoot. percentage of the frame the subject fills and how that will impact noise reduction? i don't even touch noise reduction until i hit like 6400 because i personally don't think it's necessary except in very specific scenarios that yield weird results. i'm way less concerned with sharpness, noise, "oversampling motion blur" than i am with getting an image that communicates well.

so i don't mean to kind of marginalize what you're saying - which i admittedly was doing - but generalizations on this kind of stuff are dangerous no matter where you're coming from because everybody approaches things differently and have wildly varying tolerances for things like noise, etc.

May 21, 2013 at 03:54 PM

snapsy
Online
• • • •
Upload & Sell: On

p.5 #19 · 5D vs 5DM3 High ISO

I agree I'm making a generalization but I think it's a valid one - the more noise an image has, the less chance that a full-size print of the image will be acceptable. Tastes and noise tolerances vary but the point is that an exposure requiring ISO 6400 yields fewer large-print candidates than an exposure requiring ISO 100. We could debate where that needle of acceptability sits, and I gave some examples of how particular subject matter affects the needle, but IMO the generalization holds. It's akin to the debate about how large two base ISO images can be printed based on the capture resolution.

May 21, 2013 at 04:01 PM

1 2 3 4 5 end
FM Forums \| Canon-mount SLRs \| Join Upload & Sell 1 2 3 4 5 end