Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

Started Nov 30, 2010 | Discussions
_GUI_ Regular Member • Posts: 386
Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

I did my own noise measurements over uniform patches to calculate the Signal to Noise Ratio (SNR) for several cameras: Canon 5D, 5D2 and 7D, and Pentax K5.

I used the lowest real ISO (ISO100 in Canon and ISO80 for the Pentax), and ISO1600, which is the highest useful ISO value for any RAW shooter on any camera (pushing ISO furthers means no improvement in SNR).

These plots were obtained:

Since those cameras have different pixel counts, a fair comparision requires to normalise all SNR values for a given output resolution. I used the Canon 5D as a reference, so the other cameras (with higher pixel count than the 5D), obtain improved SNR plots:

Looking at this noise response, the following conclusions are clear:

  • At ISO1600 the K5 performs like any other modern APS-C in the whole range

  • At ISO80, the SNR curve of the K5 is excellent, yielding a very low presecne of noise for low exposure values

This means the K5 at ISO80 is a very high dynamic range camera, capable of recording detail in the deep shadows where other cameras fail. Another advantage of this behaviour is the ability to 'recover' detail when the captures was erroneously underexposed.

Calculating the Dynamic Range over the last set of plots, using a 12dB SNR threshold criteria, the following comparable DR figures are obtained:

At ISO80, the K5 has almost 2 stops more of DR than a FF camera like the Canon 5D Mark II. When new FF sensors appear with such a performance, we can start to dream about forgetting the need to shoot several times in HDR scenes.

Complete article (Spanish): http://www.guillermoluijk.com/tutorial/noisedr/

Regards

-- hide signature --
Philip Goh Regular Member • Posts: 447
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

At ISO 1600 the K 5 performs like any other modern APS-C in the whole range

That's not very accurate. It produces a similar SNR across the ranges covered by other APS-C cameras, but the the dynamic range at 1600 is at least 2 stops better than competing APS-C cameras. This is surely nothing to sniff at and should be mentioned. The summary statement about ISO 1600 performance misleads the reader into thinking that the K-5 has no advantage over the 7D.

OP _GUI_ Regular Member • Posts: 386
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

Philip Goh wrote:

At ISO 1600 the K 5 performs like any other modern APS-C in the whole range

That's not very accurate. It produces a similar SNR across the ranges covered by other APS-C cameras, but the the dynamic range at 1600 is at least 2 stops better than competing APS-C cameras.

If you look at the plots, at ISO1600 the DR of the K5 is not higher than 7D's DR, and is lower than 5D2's DR. The high DR performance is just for low ISO values. Just look at the SNR plots, nothing else is needed.

Regards

-- hide signature --
ManuH
ManuH Veteran Member • Posts: 3,898
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

GUI wrote:

Philip Goh wrote:

At ISO 1600 the K 5 performs like any other modern APS-C in the whole range

That's not very accurate. It produces a similar SNR across the ranges covered by other APS-C cameras, but the the dynamic range at 1600 is at least 2 stops better than competing APS-C cameras.

If you look at the plots, at ISO1600 the DR of the K5 is not higher than 7D's DR, and is lower than 5D2's DR. The high DR performance is just for low ISO values. Just look at the SNR plots, nothing else is needed.

Your measurement is close to DXO but not quite. At low ISO, up to ISO 400, the K-5 has 2 stops more DR than 7D. Starting at ISO 400, the 7D lags only by a bit less than 1 stop.

OP _GUI_ Regular Member • Posts: 386
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

Right, the problem here is that DxO have the bad habit to use a 0dB threshold SNR criteria, which is totally useless for a photographer (an image with a SNR of 0dB is basically noise and can't be recognised or printed). So DR provided by DxO are not comparable with the ones I calculated here.

If you look at the plots, the K5 tends to have a lower slope in its SNR curves (in fact this is its main advantage for DR). This means having the same DR as the 7D with a 12dB criteria, is compatible with having a higher DR than the 7D with a less demanding criteria. But I insist, the DxO criteria (0dB) is useless for someone who wants to take a picture of a given DR.

DxO SNR curves can be used to recalculate the DR for a different criteria than 0dB. Surely if we do this, the K5 and the 7D would provide the same DR for high ISO values.

Regards

-- hide signature --
DrugaRunda Senior Member • Posts: 2,741
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

nice to read some independant confirmation of great DR that K-5 has at base ISO, so 2 stops advantage seems to be correct in "real-life" as well.
--
common sense is anything but common

ManuH
ManuH Veteran Member • Posts: 3,898
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

GUI wrote:

But I insist, the DxO criteria (0dB) is useless for someone who wants to take a picture of a given DR.

I certainly do not have the competence to contradict you It would be nice to see a "real world" comparison of what a higher DR can do for us.

MAGMATCICO62 Regular Member • Posts: 199
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

Thank you Guillermo, I will read the Spanish link
--
http://www.pentaxphotogallery.com/miguelangelgarciamartin

 MAGMATCICO62's gear list:MAGMATCICO62's gear list
Pentax K-1
dosdan Contributing Member • Posts: 540
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

Choosing a different lower limit for DR is fine. That produces about a 3 stops difference in DR figures for all cameras compared to the DxO figures, but the same low ISO DR trend that DxO shows has been replicated. Well done.

Dan

 dosdan's gear list:dosdan's gear list
Pentax K100D Super Pentax K20D Pentax K-5 Pentax K-01 Pentax K-3 +1 more
The Davinator
The Davinator Forum Pro • Posts: 24,493
Wow

It seems that a lot of the benefits of fullframe just evaporated with this camera. Huge DR, low noise, in body IS. Pentax seems the one to beat. I find the 16mp enough for me....but it's the huge DR that is great. Lesser chance of blown highlights....more natural look.

Great camera, great glass, great system!

 The Davinator's gear list:The Davinator's gear list
Canon EOS D30 Canon EOS 10D Nikon D2X Fujifilm X-Pro1 Nikon Z7 +16 more
SOB1 Contributing Member • Posts: 551
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

well, thanks for sharing but if you did not test at least 5 copies of each camera , I must say the test is flawed as with all other sensor tests published.

GUI wrote:

I did my own noise measurements over uniform patches to calculate the Signal to Noise Ratio (SNR) for several cameras: Canon 5D, 5D2 and 7D, and Pentax K5.

how many samples for each body?

if not more than 5 samples , it is useless test.

I used the lowest real ISO (ISO100 in Canon and ISO80 for the Pentax), and ISO1600, which is the highest useful ISO value for any RAW shooter on any camera (pushing ISO furthers means no improvement in SNR).

the ISO 80 of the Pentax is not real ISO 80.

These plots were obtained:

Since those cameras have different pixel counts, a fair comparision requires to normalise all SNR values for a given output resolution. I used the Canon 5D as a reference, so the other cameras (with higher pixel count than the 5D), obtain improved SNR plots:

Looking at this noise response, the following conclusions are clear:

  • At ISO1600 the K5 performs like any other modern APS-C in the whole range

  • At ISO80, the SNR curve of the K5 is excellent, yielding a very low presecne of noise for low exposure values

This means the K5 at ISO80 is a very high dynamic range camera, capable of recording detail in the deep shadows where other cameras fail. Another advantage of this behaviour is the ability to 'recover' detail when the captures was erroneously underexposed.

this is true.

Calculating the Dynamic Range over the last set of plots, using a 12dB SNR threshold criteria, the following comparable DR figures are obtained:

At ISO80, the K5 has almost 2 stops more of DR than a FF camera like the Canon 5D Mark II. When new FF sensors appear with such a performance, we can start to dream about forgetting the need to shoot several times in HDR scenes.

in fact , this is why FF is already doomed and now.

and dont forget after the 645D introduction MFB is getting cheaper and cheaper.

so most of serious landscapers will go MF eventually and others will find the latest gen APS-C completely acceptable for their needs.

IMO, FF is really doomed now, espcially a low mp FF like the D700.

Complete article (Spanish): http://www.guillermoluijk.com/tutorial/noisedr/

Regards

-- hide signature --
-- hide signature --

Ken.

OP _GUI_ Regular Member • Posts: 386
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

SOB1 wrote:

well, thanks for sharing but if you did not test at least 5 copies of each camera , I must say the test is flawed as with all other sensor tests published.

No problem if the test is a bit flawed as long as it is 100% representative. There are no huge differences in sensor noise between camera units.

how many samples for each body?
if not more than 5 samples , it is useless test.

Depends on body, from 12 to 27. The source data is here: http://www.guillermoluijk.com/download/ruidoyrangodinamico.xls

the ISO 80 of the Pentax is not real ISO 80.

I think it is 100% real (Canon's ISO50 is not). RAW histograms are full of levels and vary according to shutter:

ISO speed: 80
Shutter: 1/13.0 sec
Aperture: f/2.8

ISO speed: 100
Shutter: 1/15.0 sec
Aperture: f/2.8

Cheers!

-- hide signature --
kadarpik Forum Member • Posts: 96
Some DR tests

Single shot no HDR not special tweaking, Lightroom fill light is only PP. ISO 160, just took for testing when I was driving back home today.

Philip Goh Regular Member • Posts: 447
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

If you look at the plots, at ISO1600 the DR of the K5 is not higher than 7D's DR, and is lower than 5D2's DR. The high DR performance is just for low ISO values. Just look at the SNR plots, nothing else is needed.

Unless I'm misunderstanding the plots, at ISO 1600 the K-5 starts from -0.7 EV to -7.5 EV, which gives it 6.8 stops of DR. The 7D starts from -2.5 EV and goes to -7 EV, for 4.5 stops of DR. From -2.5 EV to -7 EV, the K-5 and 7D have the same SNR, but from the graph it looks like the K-5 has higher DR even at ISO 1600 compared to all the other cameras.

kadarpik Forum Member • Posts: 96
Re: Some DR tests

this image has fill light at the maximum to retain details at the truck, without fill light the image is here.

OP _GUI_ Regular Member • Posts: 386
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

Philip Goh wrote:

Unless I'm misunderstanding the plots, at ISO 1600 the K-5 starts from -0.7 EV to -7.5 EV, which gives it 6.8 stops of DR. The 7D starts from -2.5 EV and goes to -7 EV, for 4.5 stops of DR. From -2.5 EV to -7 EV, the K-5 and 7D have the same SNR, but from the graph it looks like the K-5 has higher DR even at ISO 1600 compared to all the other cameras.

DR is measured from 0EV (sensor saturation). The plots don't reach that limit simply because I didn't care to have samples so close to saturation, since they are irrelevant. The only important thing to precisely know is where each sensor saturates and make that point the reference for 0EV.

These are the sat points I used for the test:

CAMERA ISO100/80 ISO1600
-------------------------------------------
5D: 3692 3692
5D2: 15758 15763
7D: 13582 15299
K5: 16383 16383

Regards

-- hide signature --
OP _GUI_ Regular Member • Posts: 386
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

ManuH wrote:

GUI wrote:

But I insist, the DxO criteria (0dB) is useless for someone who wants to take a picture of a given DR.

I certainly do not have the competence to contradict you It would be nice to see a "real world" comparison of what a higher DR can do for us.

Basically two things:

  • Being able to capture detail both in the highlights and shadows of high dynamic range scenes in a single shot. Just to make an idea, any interior room with a window open outside to a sunny day is at least 12 stops. In my calculations the K5 captures 11,2 stops.

  • Being able to make mistakes in exposure (typically strong underexposure), and still have decent images.

Examples:

Typical 12 stops DR scene:

Image 6 stops underexposed K5 at ISO80 (left original JPEG, right processed RAW):

-- hide signature --
tigrebleu Senior Member • Posts: 2,060
Agreed.

GUI wrote:

At ISO80, the K5 has almost 2 stops more of DR than a FF camera like the Canon 5D Mark II. When new FF sensors appear with such a performance, we can start to dream about forgetting the need to shoot several times in HDR scenes.

Regards

The K-5 is kind of spectacular when it comes to dynamic range. However measuring this DSLR againts a camera like the 5D Mk II, which is two years old, is almost unfair, LOL!

Today's APS-C sensors will usually beat two-year old FF sensors on all aspects except noise at high ISO. Sensor age is now as much important (if not more) than sensor size when it's time to measure image quality, since technology evolves so fast.

Just look at the D700 vs. the D7000 on the DxO Marks website.

With the exception of high ISO, the D7000 almost catches up with the D700 on all other levels. Same goes for Canon: the 7D has almost as much DR than the 5DMkII, despite having photosites twice as small!

What is surprising is to see the 16 Mpix Sony sensor actually beating the 18 Mpix Canon sensor by such a wide margin. Sony has clearly taken the lead in APS-C sensor performance, at least for now.

The new FF sensors coming this year will probably have DR in the range of 15-16 stops, with even better high ISO performance and color depth... and all this with even more megapixels than before! Unless my sources are wrong, expect two BIG — and I mean BIG — announcements in February or March, first Nikon then Canon.

The full frame is far from being dead.

 tigrebleu's gear list:tigrebleu's gear list
Olympus E-M1 Olympus M.Zuiko Digital ED 60mm F2.8 Macro Olympus 12-40mm F2.8 Pro Phase One Capture One Pro HDRsoft Photomatix Pro +3 more
moving_comfort
moving_comfort Veteran Member • Posts: 8,242
Re: Some DR tests

kadarpik wrote:

this image has fill light at the maximum to retain details at the truck, without fill light the image is here.

.

That's really impressive - any other PP there after the fill light, NR?

.

 moving_comfort's gear list:moving_comfort's gear list
Pentax K20D Nikon D800 Nikon AF-S DX Nikkor 35mm F1.8G Nikon AF-S Nikkor 300mm f/4D ED-IF Nikon AF Nikkor 20mm f/2.8D +10 more
photogerald Senior Member • Posts: 1,798
Re: Canon 5D, 5D2, 7D vs Pentax K5 noise and DR comparision

SOB1 wrote:

the ISO 80 of the Pentax is not real ISO 80.

The "real" or native ISO of the K-5 may not be quite as low as 80, but it has been shown to be less than ISO100 (therefore there is some IQ benefit to shooting at ISO80).

Keyboard shortcuts:
FForum MMy threads