Sigma 50-500 OS vs Canon 100-400 IS

[Update: There is a second test in the comments section in this article]

Alright, let me put it like this, today I was going to test how the Canon 100-400 4,5-5,6 IS compared to Sigma 50-500 4,5-6,3 OS and I was a bit afraid of testing these two cameras against each other because the 5D has 21MP and I got 4.6MP. There’s a slight difference. :)

First image has been shot using ISO200 on both cameras, Sigma at F7,1 and Canon at F6,3 .

I used Lightroom 3 beta 2 for Canon and Sigma Photo Pro 4 for the Sigma.

SD14 vs 5D. 50-500 OS vs 100-400 IS
Original Canon 100-400 IS has a VERY sweet bokeh! Yummy! (right shot)

SD14 vs 5D. 50-500 OS vs 100-400 IS

SD14 vs 5D. 50-500 OS vs 100-400 IS
Original (here I pushed sharpening on the Canon version, and some to the Sigma shot as well. +0.3) You can see how much more detail the Canon has – but my Sigma is not too afraid of the fight.. :)  Looking at this shot I get the feeling that Canon’s 100-400 has the edge in terms of sharpness but that said, it’s not easy to see without those extra pixels. mmm

Second shot: Sigma at ISo50, Canon at ISO100. I tried to make WB as similar as I could. Wasn’t that easy. Canons bigger sensor gives a nice blur to the background.

SD14 vs 5D. 50-500 OS vs 100-400 IS

SD14 vs 5D. 50-500 OS vs 100-400 IS
Original I took quite a few shots and none of them from the Canon came out very sharp. I’ll redo this another day. Or is this how it’s supposed to be?

SD14 vs 5D. 50-500 OS vs 100-400 IS
Original Difference here is quite small. In order to take the Sigma shot, I walked away 5-8m to make background blur more. Didn’t help much.

So, don’t make hasty conclusions about this test. Having better optics on both in controlled situations the 5D would probably win by a wider margin. I think.

Part 2 (before in the comment section but due to problems with WordPress I put it here so that you can see these pictures)
Next test! Now I switched lens to 120-300, left Sigma, right Canon. Sigma shot at f5,6 @ 120mm. Canon 100-400 IS shot at F9 @ 100mm. I will have to do a third test. :) And I found Canon 24-70 2.8L lens isn’t sharp enough for these tests.. I wonder if the 24mm TS is better? Anyhow, the 100-400 lens from Canon looks very good, all the way from 100-400mm – it degrades slightly at 400mm but the Sigma 50-500 degrades much more, especially at corners. Will come back to this.

This one shows how sharp the 50mm 1.4 Sigma is. I tried to use it for above comparison against 24-70mm 2.8 L but failed to frame it correctly.

About Carl Rytterfalk

Welcome to my blog! I'm Carl Rytterfalk, a swedish photographer who loves everything that is interesting in the world of photography. In 2002 I fell in love with the three layered Foveon sensor and has since then been an addicted user of Sigma cameras. Though I use Canon and Nikon as well. :)
This entry was posted in 50-500 OS, Adobe Lightroom, Canon, Review, Sigma Photo Pro, Sigma SD14 and tagged , , , , , , . Bookmark the permalink.

34 Responses to Sigma 50-500 OS vs Canon 100-400 IS

  1. Sharpness and detail is one thing, but the Sigma shots seems to have rendered more nuance. I’m interested in how the SD15 will fare.

  2. Jeff says:

    Thanks Carl,

    I agree with Scott in that although the Canon is sharper (as in the bird nest shot), the Sigma has better nuance (look at the tree moss above the box).

    Would you mind doing a B&W comparison sometime? I’m interested in comparing B&W shots with Foveon noise vs. Bayer noise. Obviously there will be a LOT less noise on the full frame 5D MkII, but perhaps if you cranked the ISO up such that there’s equal luminescence noise on each we could see the differences in the sensors.


  3. Luis Cotero says:

    The Foveon with “just” 4.6MP has really shown a great sharpness, but the most outstanding characteristic is the Dynamic Range. In every single shot you can see that Foveon sensor give us a lot of detail and information in the highlights and in the shadows, where the Canon doesn’t.

    So, with this beautiful sensor you got sharpenes and a superb dynamic range. What else could you wish for? A full-frame Foveon with 10MP (or 30MP if you count the three layers!) maybe? Or just a real APS-C size with crop factor 1.5x, instead of 1.7x that Sigma have, with more MP but remaining the pixel size (for dynamic range issues)?

  4. This is an answer to a private mail, it’s about why we feel the way we do about these foveon files.. And how it compares to bayern (5D mark 2 and the like) and film. I thought it would be interesting for more to read, and please reply if you have your own thoughts. (I often get questions like these so an open answer might be a good thing)

    Hi Gus,

    These things are all a bit tricky. On one hand, yes, you get a very film-like grain, and detail. What that means is that nothing in the picture comes as a surprise. Nothing is guesses, everything is predictable. Bayern uses a very different approach, and has to guess so much that different areas of the photo will have a different feeling, some nice, some bad.

    Film is predictable, and therefore more easy on the brain. More “real” as single dirt etc has it’s realness applied to it. No bayern guesses has lost info it in the process. For me, a lot of this real feeling can be seen in the out of focus area (or the area on the way to become out of focus), especially on shots that has very low iso, ISO50 or 100. I’ve had the Canon for about a month at home, with lots of nice lenses and how much have I used it? Very little. I thought I would use it MUCH more. but it’s something in these Sigma RAW files that makes you come back – open them a little, peak around – change a little, and .. they’re just so nice.. the dynamics are just very different than any of those Canon raws – Canon raw is more like a fight for me. Dealing with them in Lightroom, Capture One, Aperture – I feel like I can sit forever and I’m never really satisfied. My eyes has been damaged by Sigma..

    Now, 4,6MP’s isn’t much. I just did a quick test – as you might have noticed and I was very surprised (yes, I was) to see how well those 4,6MP’s stood against 21MP of bayern – in the last shot it’s not even a clear winner even in the sharpest areas. (Where Canon should be king)..

    SD15 will have much lower noise – better AF (it’s not bad today, but better is nice) many more frames before buffer is full (21 shots) and a better LCD and body. I’m told it’s built like a tank with nice rubbery plastics. But I haven’t touched it since the pre-production days so I can’t confirm anything. It has a slightly different body since then too.

    BTW, I have shot seagulls 6×6 quite a lot back in China, and I love the qualities with Kodaks 160 professional film.. It’s different. But closer to foveon than to bayern. That’s for sure.

    About lens choice. Wait for the 17-50 2.8 OS. I got the “old” none version and it’s more “pro” than the 17-70.. More predictable sharpness over the entire zoom.

  5. Scott, did you see your name there in the twitter flow? :D (first screen shot)

  6. Scott Greiff says:

    I’m famous now, Carl! Thanks! :)

  7. uri says:

    hello everybody

    fisrt off all use dpp. it’s much better.

    i think you need to shoot with Sraw2 witch give the max sharpness and foveon feel and in this mode the resolusion is excacly like the sigma.

    i spoke to canon and thay sad that in this mode each pixel has more info.

    or you can downsampale to the sigma res.and one more thing the canon dpp does a better job extracting detail and i do the sharping with “genuine fractals” action.

    you cant do this test why dont you upsamples the foveon file to 21mp and take the test.

    thank you for all

  8. I really think it is a function the resolving power of the lens. Look at most lens test and resolution per picture height. Very few crop sensors exceed the x3s 1760 pixels. So high pixel count Bayer sensors really bring not much more real resolution to the table.

  9. Matthias says:

    Hi Carl,

    What Camera Profile did you use with Lightroom? I always use ‘Camera Neutral’. The ‘Adobe Standard’ profile looks a bit artificial and blows the highlights a bit more, IMHO…

    Regarding highlights: I mostly underexpose the 5D2 by -1/3 when there is a contrasty scene.

  10. Daniel Larsson says:

    “In order to take the Sigma shot, I walked away 5-8m to make background blur more. Didn’t help much.”

    Try the opposite and walk closer! That will give you more blur of the background.

  11. Daniel, :D If you want to maintain the frame, you better step back and zoom in. That way you will eventually get more blur, and more possible shake but that’s why we have OS. :)

  12. Matthias, I will try that for sure! I made a new test, and it’s very interesting results and I will also try other developers to make sure I get the most out of the Canon. But for highlights, I will try your way!


  13. Uri, even if it looks like you have more detail on RAW2 mode images, it really isn’t (just smaller). You have less. Less pixels and also a pre-processed shot, like a tiff. It’s not bad like a jpeg, but it’s not “real” RAW either.

  14. Next test! Now I switched lens to 120-300, left Sigma, right Canon. Sigma shot at f5,6 @ 120mm. Canon 100-400 IS shot at F9 @ 100mm. I will have to do a third test. :) And I found Canon 24-70 2.8L lens isn’t sharp enough for these tests.. I wonder if the 24mm TS is better? Anyhow, the 100-400 lens from Canon looks very good, all the way from 100-400mm – it degrades slightly at 400mm but the Sigma 50-500 degrades much more, especially at corners. Will come back to this.

    [img] [/img]

    This one shows how sharp the 50mm 1.4 Sigma is. I tried to use it for above comparison against 24-70mm 2.8 L but failed to frame it correctly.

  15. uri says:

    hello carl

    it look’s like this tests is sold out because the test are nod done properly.

    i spoke to canon and thay sad that the Sraw2 in a real raw file and not tiff.

    i think that if you want to do this test right you need to shoot in sraw2 as well.

    i am sending you a link to “ken rockwell site. look what he has to say.



    As hoped, lower resolution files get sharper because Bayer Interpolation is no longer needed.

    Unlike Nikons, which don’t get any sharper at 100% when set to smaller resolution files, M and L files form the 5D mark II are super, duper sharp.

    I don’t need 21MP. I usually shoot set down to 11MP and get images much sharper than from 12MP (native) cameras.”

    ” Resolution

    No one needs 21MP. All it does is slow everything and clog your hard drive.

    Try shooting your 5D Mark II at it’s M (11MP) or S (5MP) settings. If you look at your images at 100%, you’ll see that the lower resolution shots are sharper!

    Why? Because they use less, or no, Bayer interpolation. No digital camera really resolves its rated resolution; they cheat and interpolate up, so at 100% at its rated resolution, no digital camera image is as sharp as a true scan from film.

    At the 5MP setting, you have 100% R, G and B pixels, exactly as if you were using a Sigma Foveon sensor. If Sigma was selling this, they’d sell the 5MP (S) setting as if it were 15MP (also a lie).

    What this means is that the lower resolution settings actually pack away lot more detail than you think. The S (5MP) setting of the 5D Mark II is a lot sharper than any 5MP camera.

    When I’m photographing family and friends, I shot at S!

    Likewise, there may be very little extra real definition gains by shooting at L, as opposed to M. M many be all anyone needs for any size enlargement.

    If you’re testing lenses, sure, shoot at L, but for everything else, try the settings for yourself, You’ll probably get what you need at the smaller settings. For instance, the 11MP setting of the 5D Mark II has way more detail than any of the 12MP (native) Nikon cameras.

    The resolution advantage of the 5D Mark II is obvious, even at lower settings. Try them. ”

    thank you again

  16. Scott Greiff says:

    You’re getting very good results from not-top-of-the-line lenses on both sides of the camps.

    While I understand Uri’s quest for Canon-perfection, I would have to argue that the point of this comparison is not for one to win and the other to lose. It’s hard to imagine that any Canon shooter, given Carl’s Canon files, would be all that disappointed with the results.

    It seems Carl is just demonstrating the Sigma / Foveon ability to hold its own and to show its potential. Carl’s work is an inspiration for sure.

  17. Uri, Ken is perhaps not the best source..

    Use google and you will find plenty of tests on this subject. It even looks like noise reduction has been applied already in the sRAW file (hence being like a tiff, and also why noise seams to be less compared to resized full size version).

    When you make something bigger smaller, you loose something. Even if it isn’t much it’s not “better” or “sharper”. It’s worse but takes less space and you might not need whatever was thrown away. Even Canon claims this.

    Here’s some people talking on Fred Miranda about it, complaining about lost sharpness.

    Uri, you can try this yourself easily. And you will be surprised! Put your camera on a tripod, take a shot in RAW, sRAW2 and sRAW1. Bring them inside the computer and resize all of them in photoshop to RAW size. (21MP). Compare.

  18. uri says:

    hello carl

    i did a test i shoot a raw file and then downsample it to the Sraw2 and it came less sharp.

    one more thing the sraw2 look less sharp but when you sharp it you dont get grain or noise at all.

    how you can compare the two cams ? why dont you upsample the sigmas file to the canon and compare it. and more when dont live in a world when you shoot at iso 50-100 all the time we are not in studio. most of the time you need 200-400 to get speed with long lenses. canon has alot of advantages the 5dmk2 have weather sealing and my lenses too. it has excellent optics and IS has speed you you can shoot at high iso and we can crop witch you cant because you dont have larg files.

    thank you

  19. Uri, I’m sorry, but you’re way to fast! :) Take a deep breath and take a closer look at my “test”.

    The SD14 file has in ALL cases been saved as “double size” output from Sigma Photo Pro that gives you a 18MP file that compares well with the 21MP file from Canon. I never tried to downsample as I don’t see the point of doing so.

    ISO on the first shot (with the bird house) is ISO200 in both Canon and Sigma. Pretty much what you asked for. ISO50 a sunny day is very “normal” for any camera. In that case I used ISO100 which is the “best” you get with the Canon. Hence logical choice.

    The 5D isn’t weather sealed. It can take a little of rain (10mm in 3 minutes) but it’s not sealed. From Dpreview: “The 5D Mark II has the same pixel count and the same sensor size as the flagship 1Ds Mark III. While it lacks some of the more robust professional features from the 1Ds, such as weather sealing and the more advanced AF system”

    I’m happy you like your camera! You should! :) It’s not a bad camera even if I love my Sigma and if nothing else, I have more than 8 years of experience with three layered photography.

  20. Anonymous says:

    hello carl

    i love sigma and foveon i think there is something special about it.
    rhat why i am thinking to buy sd15 !!.

    i am just thinking about the total package! like lenses.

    i have alot of L lenses and the iso preformance is good.

    is there any camera you think is better than the sigmas

    thank you

  21. Martin Carlsson says:

    Ooooh, so many comments, thoughts and feelings :)

    Either way I think it’s just amazing that you even can compare a 4.6 mpx with 21 mpx. If anything, this test does really show that the megapixel-race that is going on right now is just ridiculous, specially on compact cameras.

    The reason why I don’t use Sigma is mostly the noise at high ISO and that it’s so slow. But I’m really looking forward to the SD15 and will probably buy one if it’s not to expansive. Then I’ll have to borrow all your lenses Carl :)

    Oh, and please Sigma, fix SPP! Or Apple, fix support for Sigma in Aperture 3!

    Thanks for the test! It was very fun to see!! See you when I get home! Bye for now!

  22. Martin Carlsson says:

    By the way, about the 24-70mm… I think I probably need to calibrate that one actually. Sometimes the AF isn’t focusing 100% where it should be.

    Oh, and have you tried f 3.2 and 3.6? I actually find it most sharp there!

  23. Anonymous, Today, most cameras out there is really good. Nikon, Canon, Pentax, Leica, Phase One, Mamiya, Hasselblad etc. I tried most of them, liked most of them – but it’s like choosing film.. I like the Foveon film better, but I can really enjoy others as well but I feel they need much more attention from the photographer in post, where the Sigma (as for now) needs a bit more attention in pre. (FPS, AF speed, ISO etc etc)..

    SD15 will help me a lot on my assignments and I might not need to lend Nikons and Canons as much in the future. :)

  24. Benji says:

    Thanks for these test shots. It’s very informative for someone like me who is considering getting a 5DM2 when I already have a SD14. I first decided to buy a SD14 based on many hours of pouring over your blog, along with flickr!
    I love the foveon images here, and I agree about your comment relating DSLR brands are like different films :)
    If anything, I mostly crave a 5DM2 for it’s low-light abilities (I work a lot of night shifts) and great looking narrow DOF. But I still prefer the Foveon tone and details personally. I can’t wait for the SD15 (and hopefully Sigma has something beyond that planned for the not too distant future).

    Anyway, just wanting to say thanks for the great blog and time and effort you put into it :)

  25. Jeff says:

    Hey Carl, I know this is in response to an old comment, but why do you think the SD15 will have better AF? I haven’t seen that listed anywhere. I don’t doubt you, but I’d like to know your source. :)

  26. NewbieSD9 says:

    Hey Carl

    Ive just recently got the SD9, and having had Fuji S2 pro, Olymous E330, Sony R1, everytime i open a raw file from the SD9, its wow.
    Ive taken pics of my local park, woods for a year using the above bayen cams, SD9s foveon gives that nuance, realism, subtlety, which makes all the three baen i sued kind of clinically sterile.

    I wonder, some foveon users have said, sd14, 15, dont quite have the magic of SD9, how do you feel about it please.

    Thank you.

  27. Hi NewbieSD9!

    You have the magic camera! It’s more tricky to use as it’s a bit more noisy at higher ISO and you need to feed it with batteries – but if you expose carefully and deal with those images in the best of ways – then you will have a result that is really hard to beat – even for the latest of Foveon equipped Sigmas.

    Please share your best shots as comment. :)

  28. NewbieSD9 says:

    Hey Carl

    These i’ve downloaded from my photobucket, as i thought they’d take up less space on your site :-))

    All from SD9

  29. NewbieSD9 says:



  30. NewbieSD9 says:




  31. NewbieSD9 says:




    i think these are all without any unsharp mask. Using standard kit lens 18-55 f3.5-5.6.

    Hope you like them.

  32. NewbieSD9, just upload. My server has close to unlimited space. I can’t see your photobucket links. If you still want to link, please use [img] http:// and end with [/ img]..

  33. Pingback: SIGMA SD1! « CoteroPro

  34. Hi Carl,

    I just got my second Foveon camera, a SD15 (I already have a DP2) with a 17-50 2.8 EX DC Macro, and waiting for my 70/2.8 EX DC Macro to arrive. It was you who told me to buy this lens some months ago hehe.

    Well, I never was a big fan of any Sigma lens, so I decided to use my old M42 Fujinons Super EBC on it. It’s amazing how better the pictures are, compared with the zoom that came in my “kit”. Much much sharper, believe me. I ordered a custom made focus screen from Mr. Haoda two weeks ago and he asked me for 3 weeks to make it. Not bad !

    I have a friend that ordered at the same time, in the same order, a SD15 but he decided to convert it to Leica-R mount. I was completely shocked with the samples he sent me.

    Please take a look at:

    Sorry, my website is still in portuguese =)

    I’m not that sure if the 5D wins for a large margin if you use a really decent lens on your SD15. I have a 5D with some good L glass and this little SD15 is a beast. With quirks, but still a beast.

    I agree with Scott. The nuances are far better. Scott, Jeff and Seng have stunning results with our beloved Sigma. Well we are maybe addicts (including you hehe). I’m glad to be part of the family now :P

    Please take a look at my SD15 gallery:

    Cheers !


Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>


Optionally add an image (JPEG only)