PetaPixel

Canon 6D and 5DMk3 Noise Comparison for High-ISO Long Exposures

Astrophotography enthusiast Don Marcotte wanted to find out whether the Canon 6D or Canon 5D Mark III was more suitable for his area of photography, so he pitted the two cameras against one another in a few noise tests at his local camera store. He simply shot long exposures without any light (the cap was on) in order to see how much noise would show up in the frame.

As you’re looking at the following images showing the results, keep in mind that this test was for a very specific purpose and might not be relevant to other types of photography. Marcotte tells us,

I was evaluating the 6D vs Mk III for a specific photographic application: nightscapes with the sky in the background. Even with super wide angle lenses, the stars will trail at 26 secs. But you need a lot of photons to capture the sky. So I wanted to find a very low noise camera that would give me what I needed at 15 secs and ISO 3200.

For this first comparison, Marcotte shot 15-second exposures with the two cameras in RAW mode with noise reduction turned off, converted the images to 16-bit TIFFs in Canon Digital Photo Professional, created 400x400px crops, and then boosted the gamma up to 3.0 in Photoshop CS5 in order to see how much noise is present:

Here’s the same test done, except the exposure time was increased to 30 seconds:

For a final test, Marcotte decided to pit the two newer cameras against his personal Canon 5D Mark II, which is modified for astrophotography. He used the same process as the two tests above and a 15-second exposure time:

Here’s the conclusion Marcotte came to, which he shared over at planet5D (they’ve published a lengthier writeup of this test):

The 6D is the clear winner at 15 and 30 seconds for ISO 3200 and 6400 [...] I would expect that no one could detect a noise difference between the 6D and 5D3 for the vast majority of daylight photography. Only in low light conditions would a 6D deliver observably better results.

It’s interesting that the “entry-level” full-frame DSLR outperforms the higher-end one, but what’s crazier is how much the sensor in both cameras have improved since the 5D Mark II.


Update: DxOMark’s latest sensor tests of the three cameras seem to confirm Marcotte’s findings:


Image credits: Photographs by Don Marcotte and used with permission


 
  • http://twitter.com/abezabih Abraham Zabih

    What is going on here?

  • MD

    Canon has always touted their philosophy of giving every camera the best sensor they have at the moment, regardless of its spot in the lineup. This has P.O.’d a lot of early adopters, namely with the 7D/60D and 1DIII/5DII crowds, but Canon certainly can’t be faulted for not moving forward in at least one regard.

    That said, the samples I’ve seen from the 6D have been pretty underwhelming (noise aside), but I’m holding out hope that the tester was just incompetent… :D

  • Gary C

    To my untrained, the 5DMKIII sees more stars ;-)

  • http://twitter.com/mugurm Mugur Marculescu

    Hahaha, at first glance I thought the same thing.

  • http://twitter.com/Mch1978 Mike

    Who does need a long exposure at higher ISO?

  • nate parker

    no way- come on now!

  • GtrQc

    Stars at night !

  • GtrQc

    6D vs 1DX ?

  • Mikko Löppönen

    Did he check that the exposure is also the same? I mean…the 5d mark II at ISO 1600 is a different exposure level than the 5d mark III. A pretty big difference. Same with D800. You can’t just compare an ISO number.

  • brandon

    I haven’t read the article on planet 5d, perhaps it may, possibly be legit, but i doubt it. not that i think it’s fake, but i think there is some much going on in dslr’s these days it’s very hard to say for sure what is really happening from a long exposure with a lens cap off. I can only imagine all the image manipulation going on behind the curtain, and I can imagine a great many tricks. give me some real life side by side images.

  • Trey Mortensen

    I am actually going out tonight to shoot stars and with my T2i, I struggle with the balance of noise and enough light to get the exposures I want. That is one of the main reasons I am saving up to get a full-frame camera.

  • http://www.facebook.com/leoabreuphoto Leonardo Abreu

    The 6D is unbelievable…

  • sayithere

    damn.. I thought the 5DMKII could successfully shot the reddish milky way :|

  • sayithere

    To my surprise, I just did the same test to my $600 Pentax K-r at ISO6400 and 30 seconds exposure. And the result is even clearer than of the 6D.. no noise at all..

  • RmnTz

    Did you do it “with noise reduction turned off, converted the images to 16-bit TIFFs in Canon Digital Photo Professional, created 400x400px crops, and then boosted the gamma up to 3.0 in Photoshop CS5 in order to see how much noise is present” ??

    If not, you’re comment is irrelevant …

  • RmnTz

    Did you do it “with noise reduction turned off, converted the images to 16-bit TIFFs in Canon Digital Photo Professional, created 400x400px crops, and then boosted the gamma up to 3.0 in Photoshop CS5 in order to see how much noise is present” ??

    If not, you’re comment is irrelevant …

  • Gus

    That’s a wrong statement.

    ISO stands for “International Organization for Standardisation”. It is the regulatory body for all things comparable. The whole purpose of it is to make numbers fully comparable, irregardless of when, where and who produced something which is regulated by ISO.

    So, yes, you can compare any ISO number, always. You must compare ISO numbers. That’s the whole point of it being “ISO”

    If a company cheats and does not comply with the ISO standard, it’s their loss. It’s not up to the consumer to be making correction factors and such

  • Tim

    When i tried the 5D2 i wasn’t that impressed with the noise handling. At iso 3200 i could definitely see some noise, and that was in daylight. I’m not saying it was bad, just not as good as i was hoping it would be.

  • http://www.facebook.com/people/Logan-Luckey/100001905246310 Logan Luckey

    just did the same test with my T3i, should have forked out the little extra for the 6D… good test for future reference though, thanks for the info!

  • francisbasco

    curse canon for coming out with the 6d. shouldve waited for this instead of the MKIII, but CF > SD.

  • http://twitter.com/hassan_chamoun Hassan Chamoun

    lucky bastard

  • ozi

    this is an illusion, just giant pixels 6d. please a little above refer to your monitor. 5dm3 better me to.(i shoot raw)

  • http://www.facebook.com/people/Joey-Duncan/1111692326 Joey Duncan

    I saw all kinds of dust and bits…. then I realized my screen is just dirty.

  • http://www.facebook.com/patryk.kizny Patryk Kizny

    One important side note – to make this test really accurate and objective you’d need to turn off all kinds of noise correction done in the camera software. That would tell you about the sensor and electronics itself. If you aim for really high quality night sky pictures you should also consider using a BIAS/FLAT/DARK frame calibration while processing your images.

  • Shm

    I’m waiting for this test for a way too long.. No one is doing this iso comparation :(

  • Bob

    Digital ISO is only an approximation of the film speeds. There is no consistency between manufacturers, nor does ISO regulate it.

  • Daniel Lowe

    I do a ton of night photography; and I’m regretting not getting a Canon 6D from the moment it came out.

    What you want at night.. is a big sensor, but not a lot of pixels. More pixels (“megapixels”) in a dark frame equals more noise.

    The 6D is a full frame sensor with a lower number of megapixels relative to the Canon 5DmkIII… this is better for night photography where you’re pushing your camera’s light sensitivity to the maximum.

    This was the consensus opinion from Babak Tafreshi and a panel of night sky photographers at the Astro La Palma 2013 Fall night sky workshop. They recommended the Canon 6D + Samyang 24mm f/1.4 lens as the best “night sky” package available at that time.

  • KiraHaxellepa

    before I looked at the check of $8543 , I accept …that…my
    neighbour woz like they say truley earning money parttime on their apple
    labtop. . there sisters neighbour has done this 4 only 19 months and by now
    cleared the debts on their house and bourt a gorgeous Ford . visit this site C­a­s­h­d­u­t­i­e­s­.­C­O­M­