How high can your camera fly? We present our 2014 High ISO 8 x 10 print showdown - Updated 7/16/14
posted Sunday, July 13, 2014 at 5:02 PM EDT
UPDATE 7/16/14: Seeing all these cameras lined up like this, IR Tech editor Zig Weidelich pointed out that some of the cameras listed weren't necessarily shooting at the ISOs they were set to. See NOTE 1 below for an explanation of this, and the changes we made to this story as a result.
Here at IR we're a bit old-fashioned in one sense, in that many of us still like to print our photos for displaying on walls or sharing with family and friends. We also share another love, and that's the ability to capture good images in low light without needing to rely on a flash.
We learned many years back that print quality doesn't necessarily follow in lockstep with an image's on-screen appearance, viewed at 100%. Sometimes, noise in one camera's images can actually look worse on-screen than a second camera's, yet the second can produce better-looking prints. We've also found that, at any given ISO level, there's a fairly distinct maximum print size you can manage before the image quality drops visibly.
This is why we make dozens of prints for most cameras we test: to report on just how the cameras stack up against each other when it comes to print size from default JPEGs. (See note 1 below, and thanks to Canon for their support of these efforts, providing printers like the excellent PIXMA Pro 100 and the liters of ink we go through in a typical year.) As far as we know, we're the only camera review site that performs this service for readers on a consistent basis, and we're proud to deliver this unique content in most of our reviews.
Another great reason to perform this test is that it really doesn't matter how high your camera says it can go on the ISO scale. After all, we see some cameras that can print a good 5 x 7 at their highest ISO, and yet others who can't even print a reasonable 4 x 6 at their top two ISOs! Obviously there is a good bit of descrepancy from camera to camera, and some manufacturers tend to be more realistic in their claims than others.
to determine how large you can expect to achieve good prints at each ISO level.
Recently, while discussing print sizes for a camera under review, we thought it might be interesting to post a quick comparison of print sizes for some of the key cameras we've tested in the last year or so. Our goal is to compare the maximum ISO each model can reach and still deliver a solid 8 x 10 inch print. We chose 8 x 10 since it's a common print size, but still larger than a typical snapshot. And when we say a "solid print" we mean one with good sharpness, contrast, detail, color and very little visible high-ISO noise. In other words, a print that we'd be proud to show to a friend or colleague.
So, just how high can the current models fly? Below are some examples of the most popular cameras selling today and how they fare in this competition. The listed ISO is the highest sensitivity at which you can expect to achieve a good 8 x 10 inch print - crisp and full-colored with little to no obvious noise reduction artifacts.
We've grouped them by their ISO ranking, starting with a few popular premium compacts that can print a good 8 x 10 at ISO 800, and then moving steadily onto the higher flyers.
NOTE 1: Most cameras "cheat" a little on their ISO rating, in that the actual sensitivity at the sensor level is lower than the camera's setting would indicate. Most of the time, the fudging is minor, on the order of a third of an f-stop or less. Other times it can be quite a bit more. To take this into account, we've updated the ranking based on ISO sensitivity numbers as measured by DxO and available on their DxO Mark website. The numbers in parenthesis below show the camera setting and the actual ISO as determined by DxO's evaluation. Accordingly, some cameras will appear in a lower ISO category than they were set to when the shot was captured. We've set the threshold for "demoting" a camera to a lower category when it undershoots its rated value by 30%, which is half an f-stop below the rated value. This will mean that cameras moved to a lower category in this way will generally be operating at a higher actual ISO than others in the new category, plus or minus a half of a stop is the "rounding error" anyway. We've color-coded the listings to highlight some of the variations. Actual ISO levels noticeably above the camera's setting are shown in bold green, those more than 20% below the category value are shown in red, cameras shooting more than a half-stop below their rated value have the ISO setting shown in bold red and appear in the next-lower category. Many of the demoted cameras will then be shooting somewhat above the category level, so their measured ISO value will appear in bold green as well. (We know it's a little confusing, but think it'll become clear when you read through the data.)
NOTE 2: We use a consistent, internal standard for evaluating print quality. It's proven consistent between our different raters, and is fairly stringent in terms of the amount of visible noise we allow and the sharpness of rendered detail. If we've learned anything over the years, though, it's that different people have different standards for what constitutes acceptable image noise and detail rendering. That's why we've from the very beginning provided access to all of our test images, exactly as they came from the camera. Starting several years ago, we also routinely include a very extensive set of RAW files, for you to download (for non-commercial use), convert and print using your own software and settings. What ultimately matters is what you think of the photos your camera captures. So use the results below as a jumping-off point for your own evaluations.
ISO 800 (recreational aviation)
Canon G16 (ISO measured at 998)
Olympus Stylus 1 (ISO measured at 718)
Panasonic LF1 (ISO measured at 823)
These worthy compact cameras have a 1/1.7" sensor, allowing them to produce good 8 x 10 inch prints at ISO 800. Most compacts with 1/2.3" sensors can only go as high as ISO 400 for this purpose, so this is one good way to distinguish between the two as a general rule of thumb. Note that the Canon G16 actually gives you an ISO bonus boost to almost 1000 according to DxO's measured level.
ISO 1600 (private aviation)
Panasonic GH4* (ISO set to 3200 measured at 1860)
Sony RX10* (ISO measured at 1164)
The GH4 is the first camera in this list that failed to make the grade for its measured ISO, so while it did indeed print a good 8 x 10 at a setting of ISO 3200, it turns out that DxO measured the GH4's "ISO 3200" as only yielding an actual working ISO of 1860, thereby placing it here at ISO 1600 to remain in fairness to the other cameras.
We were surprised the RX10 didn't print as large as the Sony RX100 II, but taking a look at the Still Life image at ISO 3200 in our Comparometer reveals the reason: too much noise in certain areas, at least at default noise reduction settings. A great camera, but to our eyes, it didn't do as well as its cousin the RX100 II. It is important to note however that here at ISO 1600 we rated it good at 11 x 14 inches, and thus the asterisk that you'll see here and beside a few more entrants below. (In other words, we never ranked it officially "good" at 8 x 10, as it fell from a good 11 x 14 here to a good 5 x 7 at ISO 3200.) It also just barely made the ISO 1600 grade for its DxO measured level, roughly 27% below ISO 1600.
ISO 3200 (general aviation)
Canon EOS M (ISO measured at 2646)
Canon G1X II (ISO measured at 3026)
Canon T5i (ISO measured at 2627)
Olympus E-M1 (ISO set to 6400 measured at 3870)
Olympus E-M10 (ISO set to 6400 measured at 3394)
Olympus EPM2 (ISO set to 6400 measured at 3678)
Olympus EPL5 (ISO set to 6400 measured at 3704)
Olympus EP5 (ISO set to 6400 measured at 3153)
Panasonic GM1 (ISO measured at 2287)
Panasonic GX7 (ISO measured at 2425)
Pentax K3 (ISO measured at 2933)
Pentax K5II (ISO measured at 2482)
Ricoh GR (ISO measured at 2419)
Sony A58 (ISO set to 6400 measured at 3787)
Sony RX100 II (ISO measured at 2248)
At ISO 3200 we enter new territory, as this setting allows for non-flash shooting in fairly low light while still achieving a reasonably fast shutter speed for capturing movement. All models here have Micro Four Thirds or APS-C sized sensors, except for the RX100 II with it's smaller 1" type sensor. (Great performance for a smaller sensor, but its listed ISO only made the cut right on the cut-line, as it measures a full 30% lower than actual ISO 3200.) Only the three Canons and the Pentax K3 turned in measured numbers less than 20% below the line, while five popular Olympus cameras were brought down to ISO 3200, as their measured ISO 6400 figures fell below our acceptable standard of within a half stop (and the EP5 actually fell more than 50%). The Sony A58 was also demoted from ISO 6400 status, as its measured results were roughly 40% below actual ISO 6400.
ISO 6400 (cruising altitude)
Canon 70D (ISO measured at 5254)
Fujifilm X-T1 (ISO not measured by DxO)
Fujifilm X100S** (ISO set to 12,800 measured at 6200 (by IR))
Nikon D7100 (ISO measured at 4516)
Sony A6000 (ISO measured at 5181)
Sony A7 (ISO measured at 4860)
Sony A77 II (ISO measured at 4610)
Sony RX1R* (ISO measured at 4521)
Ah... now we're getting into really interesting territory. ISO 6400 represents a great deal of shooting freedom in low light! Where a shot at ISO 3200 might require a 1/15s exposure, resulting in too much motion blur, ISO 6400 would only need 1/30s in the same situation and likely do the trick (this is just one example to illustrate the point; the same could also be the case in a sports shooting situation where 1/250s is too slow, but 1/500s captures the moment). All cameras here are Micro Four Thirds or APS-C, except for the full frame Sony A7 and RX1R. The Fujifilm X100S printed a good 8 x 10 at ISO 12,800, which was a real eye opener here in our lab until we tested it ourselves and found that it only yields a level of 6200 when set to ISO 12,800. (DxO doesn't measure X-Trans sensors in this way, so we will test it once it returns to our lab.) Also interesting is that the two cameras that yielded results close to their claims (the Canon 70D and the Sony A6000) are both APS-C cameras, while 2 full frame offerings (the Sony A7 and RX1R) just barely made the cut here based on their actual measured ISO.
ISO 12,800 (rarefied air)
Canon 1DX (ISO measured at 9891)
Canon 5DIII (ISO measured at 10084)
Canon 6D (ISO measured at 10052)
Nikon D610 (ISO measured at 9214)
Nikon D800 (ISO set to 25,600 measured at 16,117)
Nikon Df (ISO measured at 9540)
Sony A7R (ISO measured at 9109)
Sony A7S (ISO measured at 10,587)
Only the "big three" manufacturers remaining here at ISO 12,800, and all full frame offerings. Anyone convinced that a crop sensor camera can really do what the big boys can do should take heed of these figures (and we're not even done yet!). This is an amazing setting to be assured a good 8 x 10 inch print indeed! So much freedom afforded a shooter at this setting with shutter speeds in low light. The D800 actually printed a good 8 x 10 at ISO 25,600, but its reported measurement didn't make the cut there, being more than 30% below the acceptable line. This is still a very high ISO though, so it's still in great company.
ISO 25,600 (into the stratosphere)
Nikon D4S (ISO measured at 19,034)
Nikon D800E (ISO measured at 17,845)
And now to the point many of us have only dreamed of - the ridiculously high ISO setting of 25,600. Sure, some of us have had this setting "available" to us on our cameras, but that really has nothing to do with whether or not it's actually usable in the real world. Well, let us assure you that on these two cameras, this setting will absolutely work in the real world, at least if a nice printed 8 x 10 will meet your needs. (You'll need way less than that amount of resolution for even a large on-screen image.) Note however that neither camera actually reaches above ISO 20,000 in DxO measured testing, though they do still make the cut for remaining at this hallowed level, measuring far higher than most cameras listed here at ISO 12,800.
And the winners are clear
Professional level Nikons seem to rule supreme in the land of high ISO shooting, at least as of this writing. Canon and Sony get worthy runners-up status at ISO 12,800, along with three terrific Nikons. These three companies are joined by the Fujifilm X100S and X-T1 at ISO 6400, as well as additional offerings from each of them.
Of course, these are only a sampling of the most popular cameras from the past year. For anyone interested in a camera not listed above, you can go to that camera's review on our site and can see where it stands in the ranks of an 8 x 10 inch print at high ISOs, if we've done a print quality test for it. If there is an "exposure" tab at the top of the review, click that and scroll down until you find "print quality". For other reviews, the print quality results may also be found on the review's main landing page. [Note that our print quality results do not currently reflect these updated measurements from DxO.]
The Future
We can't wait until someone breaks the 51,200 mark for a good 8 x 10" print! That will indeed be news in and of itself. Will it be the Nikon D810? The Pentax 645Z? Something from Canon or someone else out of Photokina in September? This should be a big year for Photokina product announcements; based on the rumor mill, we're expecting to see at least one new higher end camera from Canon, and are confident there will be more coming from many of the other manufacturers as well.
Stay tuned, low light shooters who enjoy your prints. Check back often, as we're the only site to consistently provide "high altitude" print-quality results, and you can count on their timely delivery when they hit the tarmac in our test lab. And, as always, we welcome your opinions in the comments section below!
[Written with Dave Etchells]
Additional Notes
1) Our print quality tests are conducted using straight-from-camera JPEGS at default noise reduction settings, in order to have a common standard on which to base these comparisons. Your standards for what constitutes an acceptable print might be different from ours; we based our evaluations on a fairly stringent standard, that has proven surprisingly consistent between three of us at IRHQ over the last couple of years, despite it being a fundamentally subjective visual criterion. Those of you converting from RAW files may find significantly different results, particularly depending on the noise-reduction and sharpening processing you apply; with no NR, you'll probably find smaller top print sizes, with large amounts of NR, you'll likely find bigger - Although over-aggressive NR processing reduces resolution and smears detail, so in our evaluations, too much is as bad as too little. (Also, the noise reduction capabilities of a program like DxO Optics Pro will go waaay beyond what any camera's internal JPEG processing is capable of.) If you'd like to experiment with your own software and settings, RAW files of our test shots are available for most cameras under each review's "Samples" tab. When there's a RAW version available, you'll find a second link present under the image thumbnail, that will let you download the RAW file.
2) The Pentax 645D has a maximum ISO of 1600, but prints a gorgeous 16 x 20 at that setting, so we're eager to get the higher-ISO 645Z into our lab soon!
* For several cameras listed in this article there was not a definitive 8 x 10 inch print in the ranks. This is because for whatever reason these models produced a "good" 11 x 14 at a certain ISO, but by the next ISO up the quality had fallen to 5 x 7 inches being good. This is not uncommon, as many cameras tend to have a point at which quality can fall significantly from one ISO to the next. We wanted to mention this so that you'd be aware that those models did exceptionally well at their listed ISO.
** Unfortunately, DxO doesn't provide any data for cameras using Fujifilm's X-Trans sensors, because of their innovative color filter array pattern and the resulting non-standard RAW file formats. DxO's tests of Fuji's earlier X-100 model showed its ISO essentially flat-lining at a maximum ISO of roughly 1,000, though, even with the camera's controls set to values as high as 6,400. Given this, and what looked like really unusual print quality at ISO 12,800, we thought to check the actual ISO of the X100s at that setting. Without going through all the gory details, we compared the X100s's high-ISO exposures to those of the Nikon D7000, which DxO Mark does show data for. The result of our tests was that the X100s actually performs at an ISO level of about 6,200 when set to 12,800 via the user interface. It does produce a very nice 8x10 inch print there, but in fairness to the other cameras, it isn't at an ISO of 12,800, as you'd expect from the camera settings. Consequently, we've listed the X100s above as producing a good 8x10 at ISO 6,400, matching the ISO level we calculated for that image.
