Establishing the No Parallax Point (NPP) of a Lens

Figure 1. Mount Rainier as seen from the Moraine Trail near Paradise. Comprised of 16 vertical shots stitched together to form a large Giga-panorama. © Beau Liddell, all rights reserved. ImagesByBeaulin.com.

Figure 1. Mount Rainier as seen from the Moraine Trail near Paradise. Comprised of 16 vertical shots stitched together to form a high-resolution giga-panorama.  80mm, f/5.6, 1/400 sec., ISO 100.  Correctly capturing large panoramas requires preparation and proper execution of technique.  © Beau Liddell, all rights reserved. ImagesByBeaulin.com.

Introduction

Technology during the past 10-15 years has advanced if not revolutionized photographic creativity to previously unimaginable levels. One product of this transformation is that modern-day photographers often take multiple, overlapping images to capture a larger scene and later stitch and crop them into a final composition (Figures 1 & 2). Often the process is considered synonymous with creating panoramas, although the technique can be used to create a composition of any desired crop ratio (Figure 3).

Figure 2. Screen shot of the panorama in Figure 1 showing each overlapping frame prior to final stitching.

Figure 2. Screen shot of the panorama in Figure 1 showing each overlapping frame before final stitching.

The Milky Way shining over a campsite at Paradise Beach along Lake Superior's North Shore near Covill, Minnesota.

Figure 3. The Milky Way shining over a campsite at Paradise Beach along Lake Superior’s North Shore near Covill, Minnesota. This shot is comprised of 3 vertically stitched horizontal images to eliminate aberrations along the edges and corners of the sky, and cropped back to original size. As this image illustrates, stitched photos don’t have to be panoramas.  © Beau Liddell, all rights reserved.  ImagesByBeaulin.com.

I’ll cover the process of capturing images for stitching in a separate tutorial. But one important aspect to consistently getting good results is to avoid major parallax errors in the final, stitched image. If parallax can’t be avoided, then the software used may not be able to stitch the photos successfully, or very well. By rotating the camera-lens combination around the lens’ no-parallax-point (NPP) we can effectively eliminate parallax errors. The NPP is equivalent to the lens’ entrance pupil, which is the location of the optical image of the physical aperture as seen through the front of the lens. In this tutorial I’ll try my best to explain how to establish the NPP for a lens.

Ask the Manufacturer or Consult the User Manual

First, you might be able to get data from the manufacturer on the distance of the optical center of their lenses from the sensor plane or camera-lens interface in millimeters (mm). For example, Zeiss reports the entrance pupil distance from the focal plane in the lens’ specification sheet.  Armed with this information, attach the lens to the camera and measure the distance provided so you know where it occurs along the lens barrel (but remember that the actual NPP is in the center of the lens!).

If it’s a prime lens, I recommend marking the NPP on the barrel, or if possible apply gaffers tape to the barrel and mark the NPP position on the tape for future reference. Since the NPP will vary with magnification, a cheat sheet is needed if a zoom lens is involved (Figures 4 & 5; I like to record NPPs for focal lengths printed on the zoom ring). Then mount the camera on a pano slide (also referred to as a NPP or nodal slide) that has a distance scale marked in millimeters, making note of where along the scale the NPP occurs (Figure 5). This is the mark you’ll use to align the slide on the tripod head to control for parallax.

If you’re lucky, the tripod mounting plate or L-bracket that you purchased for your camera will have been designed to mount on the base of the camera or accessory grip so that it’s center coincides precisely with the camera’s sensor plane.  If that’s the case, just mount the camera on a pano slide and position the slide on the tripod at the point where the slide’s scale equals the NPP/entrance pupil distance provided by the manufacturer.  Then the camera will be correctly positioned on the NPP of the lens until/unless your using a ballhead and reposition vertically off level (see Axis of Rotation and Tripod Heads near the end of this tutorial for how to compensate when using a ballhead or multi-way panning head – this is where marking the NPP on the lens is handy).

Incidentally, if you own the same camera and lenses I use, you can’t necessarily use the data shown in figure 4 to establish the NPP unless you also use the same camera L-bracket I do (Really Right Stuff BGE11-LB).  You’ll also notice that the NPP with the same lens and focal length vary with the camera’s orientation.  That’s because for this particular L-bracket, the side and bottom mounting plates are not aligned in the same location relative to the camera’s sensor plane.  Finally, if you purchase a new camera and tripod mounting bracket, the NPP distances you previously used may no longer work (even though you might be using the same lenses, and regardless of how you initially established the NPP for your lenses) unless the brackets for both cameras are aligned with the sensor plane.  This is another reason to mark the NPP on your lenses so that you’ll have an easier time re-establishing it with new gear.

Figure ##. Example of a no parallax point table for lenses I frequently use to make stitched images.

Figure 4. Example of a no parallax point table for lenses I frequently use to make stitched images.  This is especially handy if you use many lenses or use zoom lenses.

Figure ##. Various styles of pano slides I use to align the NPP of my lenses with the tripod head when taking photos for making stitched images.

Figure 5. Various styles of pano slides I’ve used to align the NPP of my lenses with the tripod head when taking photos for making stitched images.  Documenting NPP and other pertinent data on post-it note tape and affixing that to a pano slide is a convenient way to ensure you always know what NPP distance to use for your lenses in case you forget or misplace your cheat sheets.

Rough Method for Determining the NPP

Unfortunately, lens manufacturers don’t often report the optical center point of their lenses, particularly at different focal lengths for a zoom lens. So now what? Although it’s not difficult to precisely determine the NPP yourself, you can roughly approximate it by viewing the front of the lens with the aperture engaged, visually estimate where you see the aperture along the barrel and measure the distance (mm) from that point back toward the sensor plane. Mark this point/plane on the lens.  This is the plane you will use when positioning the camera on a pano slide that in-turn is mounted on the tripod head.

Unless your depth perception is quite poor, it’s surprising how close to the actual NPP you can get by using this crude, visual method.

Precisely Estimating the NPP

There are many tutorials available on the Internet if you want to more precisely measure the NPP yourself.  Stitching software has gotten very good in recent years, thus it’s not critical that you be absolutely perfect.  But the closer you can position the camera relative to the lens’ true NPP the less risk there is that the software won’t be able to perform the stitch, and the less rotating, cropping or warping you’ll have to do on the final stitched result.  This is especially true if stitching images taken using wide-angle or ultra wide-angle lenses.

The process to precisely determine the NPP involves the following considerations and steps:

1)  Use a tripod and head that’s perfectly level and locked down to prevent any vertical movement. I highly recommend using a fully adjustable gimbal or pano-gimbal head (Figure 6) for doing stitched images and to determine a lens’ NPP, but a multi-way panning head (Figure 7) or ballhead (Figures 8) will work if you take added precautions noted below.

Figure 6. Camera mounted on a pano-gimbal tripod head. This is the ultimate type of head to use for creating stitched images as it enables quick and accurate setup and enables rotation around the NPP along all axes.

Figure 8. Example of a multi-way panning head. These heads enable precise positioning, but unfortunatley don't allow for rotation around a lens' NPP over all axes, requiring certain precautions to get consistent results when taking shots for stitched images.

Figure 7. Example of a multi-way panning head. These heads enable precise positioning, but unfortunately don’t allow for rotation around a lens’ NPP over all axes, requiring certain precautions to get consistent results when taking shots for stitched images.

Figure 9. Camera mounted on a pano slide and ballhead. Extra steps are needed to ensure alignment with the lens' NPP when positioning the horizon

Figure 8. Camera mounted on a pano slide and ballhead. Extra precaution is needed to ensure alignment with the lens’ NPP after positioning the horizon when using this type of tripod head.  The camera-lens combination in this photo are set to the lens’ NPP at 93mm, and will remain at the NPP distance so long the camera remains level on this type of tripod head.

2)  Mount the camera-lens combination on a pano slide marked with a distance scale (mm), and position the slide on the tripod head such that the center of the lens is roughly aligned with the horizontal axis of rotation (Figure 8).

3)  On a table or similar flat platform, place a couple of relatively thin, straight objects that you can stand on-end, positioning them in-line as viewed through the camera, but at least a couple of feet apart (Figure 9). You also don’t want the front object to completely cover up the back object as viewed through the camera, and want the objects close enough to the lens so you can view them easily. Adjust the tripod distance and/or height, or the platform the objects are on so that the top of each object can be seen through the viewfinder or on the LCD, with the top of the front object preferably near the center of the frame. You might have to re-level the tripod, or re-position the objects so that everything is level and lined up.

Figure 10.

Figure 9.  Photo showing a simple set-up and objects used to establish the no parallax point of my lenses.  I used two thin LED flashlights (shorter one closer to camera), positioned about 3 feet apart, aligned with the camera’s line of sight and positioned such that the top of the nearest flashlight was in the center of the frame.  Camera and lens were mounted on a level tripod with a gimbal head.

Another option is to position the tripod near a window, and draw a thin vertical line with a grease pen on the glass in front a well-defined background reference object outside.

4)  Once your reference objects are set and aligned with the center of the frame, slowly pan the horizontal axis of the tripod head back and forth, viewing the objects on the LCD or through the viewfinder. Unless by luck you positioned the lens on the tripod head exactly at the NPP (unlikely), you’ll notice either the rear or front object move relative to the other object as you pan. That’s parallax (Figure 10, also click the video link below showing parallax in real-time).

Figure 11.

Figure 10.  Example of parallax.  Center image shows objects aligned before beginning to pan the camera.  Left image shows the left side of the frame after panning to the right, and the right image shows the right side of the frame after panning fully to the left.  The camera was not being rotated around the lens’ no parallax point (NPP) since the two flashlights moved relative to one another while panning.  The camera was in fact being rotated in front of the lens’ NPP, and additional incremental movements backwards was required before the correct NPP distance could be determined.

5)  Next, re-center the objects in the frame, move the slide fore or aft 5-10mm, and repeat the panning process. If the apparent parallax worsens (more relative movement among the objects), stop, re-position, move the slide in the opposite direction, and continue panning the camera. But, if the parallax seems to improve (less relative movement between the objects), stop, re-position, and move the slide in the same direction using smaller increments, repeating the process until the two objects remain aligned in the same relative position to each other throughout the field of view as the camera is panned (Figure 11; or click the video link below showing what it looks like when the NPP has been attained). Once this has been achieved, record the distance on the slide’s scale that’s aligned with the center of rotation, and you’ve established the NPP for the lens at that focal length.

Figure 12

Figure 11.  After several iterations of re-positioning the pano slide on the tripod head and panning the camera to determine whether parallax was still present I was able to arrive at the NPP for the lens as shown here where the two flashlights were consistently aligned relative to one another as I panned across the field of view.  The left image shows the left side of the frame after panning to the far right, and the right image shows the right side of the frame after panning fully to the left.

Regardless of how you established the NPP of the lens, in the future all you have to do to control for parallax when shooting images for stitching is to position the camera-lens-slide combination on the tripod head at that distance. The axis of rotation is now set at the NPP for that lens and focal length, provided the platform is level. If you might be creating stitched images with multiple lenses or focal lengths (in the case of a zoom), it’s helpful to create a tabular cheat sheet, or jot the data on post-it tape and stick it on your pano slide component (Figures 4 & 5).

Axis of Rotation and Tripod Heads

One final precaution is needed before you start taking photos for stitching.  Make sure the tripod is level so that the horizon is appropriately aligned; otherwise you may get a poor stitch, or a skewed orientation that might require you rotate the image as well as crop out much of the composition (Figure 12). Once the camera/lens combination is set at the NPP on the slide and attached to an appropriately prepared tripod, you need to be able to rotate the lens precisely around that point as you pan among frames.

Milky Way at Splitrock Lighthouse State Park

Figure 12.  When making stitched images it’s important to have the horizon level.  Failure to do so such as with this 15-vertical image panorama will result in either a poor stitch or at a minimum require image rotation, potentially custom warping, and might also require some significant portions of the composition be cropped.

If you use a 3- to 5-way panning head or a ballhead (Figures 7 & 8), after positioning the horizon where you want it (which will result in a vertical adjustment), the camera will no longer be level, nor rotating around the correct axis of rotation to prevent parallax. Instead it will be rotating in front of the NPP (if you pushed the horizon toward the top of the frame) or behind the NPP (if you positioned the horizon toward the bottom of the frame). Depending on the lens, the axis of rotation could well be off the NPP by several inches if you positioned the horizon near the top or bottom edge of the frame (Figure 13).

Figure 14.

Figure 13.  Although the camera was initially positioned to rotate around the lens’ NPP in Figure 8, when using a multi-panning head or ball head as shown here, as soon as I compose the shot by positioning the horizon above or below center, the camera is no longer level, and will no longer rotate around the NPP as I pan.

As a result, you will need to adjust the pano slide fore or aft as needed to re-establish rotation of the horizontal axis at the NPP to compensate for the angle created when re-positioning the horizon (Figure 14). If the lens involved is a prime and you marked the NPP on the barrel, it should be easy to make the necessary adjustment. Assuming the horizon is still level after all this, you are now ready to capture a series of overlapping, parallax-free frames that will form a row in the final stitched image.

Figure 15.

Figure 14.  Since the camera in Figure 13 is no longer rotating around the lens’ NPP, I need to take steps to re-establish the NPP before I begin to take overlapping shots for a stitched image.  If the horizon is composed toward the extreme lower or upper edge of the frame (e.g. common when taking star photo landscapes) as simulated here, you can see that the camera was rotating several inches behind the NPP and required shifting the pano slide significantly forward.  Knowing where the NPP is on your lens barrel helps make these adjustments, and is made even easier if you mark the lens barrel (for a prime lens only).

If using a fully adjustable gimbal or pano-gimbal tripod head (Figure 6), once leveled, attach the camera-lens-pano slide combination to the head at the NPP marking, and shift the vertical riser of the head to the left or right as needed to center the lens over the horizontal axis of rotation.   Now the lens is centered along both axes at the NPP and you won’t have to make any further adjustments to the pano slide after you’ve positioned the horizon where you want it.

What’s more, if you need to capture multiple rows for the final stitched image, by using a pano-gimbal head you won’t have to worry about re-establishing the NPP point when you rotate the lens in the vertical plane (Figure 6; or click on the video below demonstrating how these heads precisely rotate a lens around the NPP). This efficiency is one of the big advantages of using these types of tripod heads over multi-way panning heads or ballheads when capturing images for stitching.

I hope you found some of this information useful for capturing images meant for stitched compositions. If you have any questions regarding the information provided in this tutorial, please leave a comment or contact me at ImagesByBeaulin@charter.net.

© Beau Liddell, ImagesByBeaulin.com, All rights reserved.

Advertisements

Canon EOS 7D Mark II Image Quality Review

The new 7D Mark II, released November 1st, 2014. This feature-packed crop sensor dSLR from Canon excels at high ISO performance compared to its predecessor and market competitors.

Figure 1.  Shot of the latest addition to my gear, the new Canon EOS 7D Mark II, released November 1st, 2014. This feature-packed crop sensor dSLR excels at high ISO performance compared to its predecessor and market competitors. This camera sets the mark for crop sensor performance, particular for those shooting action sports and wildlife subjects.

Now that the long-awaited Canon EOS 7D Mark II has been available for a few weeks and its images are supported by various raw editing applications I thought I would jump in the online fray reviewing this latest camera from Canon.

If you would prefer to view a video of this material, please visit my YouTube page.

Please understand up-front, that I don’t work for and am not being paid to promote this product, or any others mentioned in this post or on my video blogs.  I’ve used Canon dSLRs and EF lenses for years, including the original 7D since it’s release in 2009.  I’m simply sharing information and my experience to date with the new Mark II that might be of use to wildlife or action photographers who are in need for a new setup and might be contemplating purchasing this new release from Canon.

I’m not going to provide an exhaustive review of the camera and its features.  For that, visit Canon USA’s website at http://www.usa.canon.com/cusa/consumer/products/cameras/slr_cameras/eos_7d_mark_ii, visit an excellent review at The-Digital-Picture.com, or do a simple internet search for all the reviews out there on the camera’s specifications.  There are even a few reviews of images taken with the new camera, including the review just mentioned, as well as by other photographer’s who had the opportunity to test the pre-release version.  And, just recently the camera has been reviewed by DxO Labs.  After posting this blog Tony Northrup posted a very good review here of this camera and its image quality, and discusses some very relevant issues associated with DxO labs sensor ratings and what you really need to be concerned about as a photographer.

The original EOS 7D was, and still is by all accounts, a very good dSLR.  Many of the photos in my portfolio at ImagesByBeaulin.com, and most of my wildlife photography the last 5 years has been taken with that camera.  The AF system performance, overall camera responsiveness and dynamic range at 11.7 EVs is very sufficient for most circumstances.  However, with the advances the past 10 years in digital imaging technology, we have turned into pixel-peepers, and expect pixel quality that frankly we probably don’t really need.

As such the original EOS 7D was criticized shortly after its release for performing poorly (a.k.a. generating a lot of digital noise) at higher ISO settings, especially beyond 800 ISO where we often have to shoot to maintain sufficient shutter speeds with using long focal lengths and moderate to small apertures.

Suffice it to say, the EOS 7D Mark II is a true performer, with even better AF features & performance than the original version (not surprising since it inherits and builds on technologies from both the EOS 1DX  and 5D Mark III AF systems), amazing responsiveness (similar to the 1DX), improvements to Canon’s dual-pixel sensor technology originally introduced in the EOS 70D, and integration of dual Digic 6 processors.

The new camera is a great machine for shooting wildlife and sports (although it will take good photos of anything you point it at), especially considering its reasonable price point at about $1,799 USD (as of Nov. 2014).  Although to be most functional ergonomically when shooting in portrait orientation, you really should invest in the accessory battery grip & 2nd battery which will add another $420 to the sticker price.

Back of the new Canon EOS 7D Mark II

Figure 2.  Outward appearance of the back of the new Canon EOS 7D Mark II. Few changes from its predecessor, although a bit more like the 5D Mark III, including better build quality, and new to the EOS line-up includes an AF point selection lever that enhances control over the sophisticated 65-point AF system that the camera inherited from the EOS 1DX.

The 65-point AF system in the new EOS 7D Mark II is built for performance off of the EOS 1DX, but has been further enhanced, and the dual pixel technology provides amazing autofocus capabilities when shooting video as well as stillshots via LiveView.

Figure 3.  The 65-point AF system in the new EOS 7D Mark II is built for performance off of the EOS 1DX, but has been further enhanced, and the dual pixel technology provides amazing auto focus capabilities when shooting video as well as still shots via LiveView. © Beau Liddell, All rights reserved, ImagesByBeaulin.com.

Thus, it was pretty much a foregone conclusion that I would be upgrading to the new camera when it was released due to the performance enhancements alone, and my camera body arrived from Canon Direct store during the official release date on November 1st, 2014.  Both Canon and several reviewers had been touting the higher ISO performance compared to the original version, as well as to the other competitors in its class.

But, I had to test the sensor quality myself immediately after receiving it to satisfy my curiosity.  It’s really the camera’s performance at high ISO that I want to get at, which is a common need when shooting with longer focal lengths, particularly when the subject’s lighting is less than ideal.

Now, I’m not talking about pushing ISO beyond 1600, although there are times when that’s needed and might still produce acceptable results with the right lighting.  But, if I need to do that routinely for a given subject (say I need to really under-expose and recover shadows during post-processing for some reason, or more commonly take shots in very low-lighting conditions such as star photography landscapes), then I most certainly will reach for a full-frame sensor with its superior signal to noise performance across the frame at such high ISO settings.

One final caveat before reading any further, is that if you’re expecting or hoping that crop-sensor cameras should or will stack up well to full-frame sensor cameras with regards to IQ and noise at high ISO settings, then you need a reality check.  No crop sensor ever has or ever will provide as good IQ performance across the frame under high ISO settings as their contemporary full-frame competitors will provide, period.  It’s a physical limitation of the sensor designs and nature of how sensors capture light linearly.  So, there will always be a need for full-frame sensors, and they will always outperform crop-sensors in regards to pixel quality under certain lighting conditions.  I’d love to be proven wrong, but I’m comfortable making such an absolute statement on this issue.

Anyway, none of this means that a crop sensor won’t perform well and provide sufficient quality for any given output.  And, there are definitely circumstances and genre where a crop sensor will outperform a full-frame sensor (i.e. resolution of fine-edge detail, increased DOF, and magnification factor when needing long focal lengths).  I’ll address all these points below, and you’ll get the most out of this review if you need to shoot with a crop sensor for whatever reason and might benefit from improved performance with as little as 1-2 stop increment boost in ISO from what the predominant crop-sensors on the market provide as of late-2014.  For more on sensor size and impacts on image quality, depth-of-field and resolution, see the excellent review on digital camera sensor sizes at Cambridge in Colour.

Photo of the subject and controlled circumstances under which test shots were taken to compare image quality at various ISO settings.

Figure 4.  Photo of the subject and controlled circumstances under which test shots were taken to compare image quality at various ISO settings.  © Beau Liddell, All rights reserved, ImagesByBeaulin.com.

Not long after receiving my new camera, I was setting up to take test shots of the sensor’s IQ under controlled circumstances (Figure 4).  The subject was a mount of a turkey in my den, and lit by 5500K daylight fluorescent ceiling lamps.  All shots were taken on a tripod, with custom white balance set to 5100K, using the same lens (EF 70-200 f/2.8L IS USM), identical subject framing at equivalent focal lengths (100mm for the 7D, 7D Mark II, and 153mm for the 5D Mark III), using LiveView, focusing manually on the birds eye, using manual exposure mode aperture set to f/11.0, and adjusting exposure time as needed using the camera’s meter to assess exposure balance as ISO settings were adjusted.

All test shots were taken in the raw file format, and automatic in-camera adjustments such as long-exposure or high ISO noise reduction, lens aberrations and distortion corrections, auto lighting optimizer, etc. were all disabled.  Therefore, the only fundamental difference between shots were ISO settings and adjustments to shutter speed to achieve equivalent exposures among shots and among camera sensors.  I took 3 shots with each camera body at ISO settings of 100, 1600 and 6400 for comparative purposes.

Once all of the test shots were imported into Lightroom, I did not apply any sharpening, nor any exposure adjustments.  The only adjustments that were applied was to set the camera profile to Camera Standard, and to apply the lens profile correction.

Figure 5 shows two test shots taken at ISO 100, comparing the 7D (photo on the right) with the 7D Mark II (photo on the left).  The thing that jumped out at me right away after importing the files was that the photo taken with the 7D Mark II seemed to be a little brighter, with higher contrast and saturation compared to the original 7D.  And, this general exposure discrepancy carried through all other ISO and camera body comparisons, including the comparison with the full-frame 5D Mark III, even if I applied a neutral or faithful camera profile to the files.

ISO 100 test shots (raw files) from the 7D (right), and 7D Mark II (left), without any exposure or detail adjustments. Notice the slightly brighter image taken by the 7D Mark III.

Figure 5.  Raw file ISO 100 test shots from the 7D (right), and 7D Mark II (left), without any exposure or detail adjustments. Notice the slightly brighter, higher contrast in the image taken by the 7D Mark II. © Beau Liddell, All rights reserved, ImagesByBeaulin.com.

Therefore, it appears that there is something about the sensor in the 7D Mark II that does an excellent job of gathering light compared to its predecessor.  It’s almost as if I had tried using Expose-To-The-Right (ETTR) technique to slightly push the histogram to open the shadows and reduce noise.  For a thorough description of ETTR technique as it relates to image quality, I have another video tutorial about Maximizing Image Quality on YouTube that covers the basics, or read the white papers “Understanding digital raw capture“, and “Linear gamma” by the late Bruce Fraser at Adobe’s Camera Raw web page.  I’ll demonstrate ETTR a little later in this review.

When I zoom into the same area on each image you can see that pixel quality is good with both camera bodies, and would definitely be suitable for publication or production of a large format print (Figure 6).  Again, the greater brightness and better contrast and saturation is noticeable in the shot taken by the 7D Mark II.  Also, as we would expect with a higher density of photosites on the sensor, there is a slight improvement in detail with the new camera, and there also appears to be slightly less noise, but is barely noticeable at such a low ISO setting.

Figure 6. 100% magnification of the images shown in Figure 5, with the newer model on the left showing slightly better detail and less noise, although the difference is barely noticeable.

Figure 6. 100% magnification of the images shown in Figure 5, with the 7D Mark II on the left showing slightly better detail and less noise, although the difference in noise is barely noticeable.  Click the image for a larger version if it’s not displayed large enough on your monitor.

For most of the following image comparisons, I will only show the 1:1 magnification views since the full image comparisons all pretty much look the same at the fit-to-view scale.

Next let’s look at two shots taken by the same two sensors, but this time at ISO 1600 (Figure 7).  At 100%, the noise is definitely apparent, and noticeably less in the left shot taken by the 7D Mark II.  This shot also illustrates that for a properly lit subject, both of the cameras can take a reasonably good quality shot that could be salvaged easily for publication or display, but the newer camera seems to do a better job.  Thus, as noted in the DxO Labs and The-Digital-Picture.com reviews mentioned above, the new 7D Mark II does a more than respectable job from an image quality standpoint a little beyond 1000 ISO, and that’s far better than we’ve come to expect from other dSLR, PAS, and mirror-less crop-sensors on the market as of late-2014.

Thus, I have no qualms whatsoever upon testing the 7D Mark II to use it at up to 1600 ISO when the need arises, especially when I am able to get by with using an ETTR exposure approach (see more on that below).

Figure 7. 1:1 view if two test shots taken with the original EOS 7D (right) and EOS 7D Mark II (left) at 1600 ISO.

Figure 7. 1:1 view if two test shots taken with the original EOS 7D (right) and EOS 7D Mark II (left) at 1600 ISO.  Click the image for a larger version if it’s not displayed large enough on your monitor.

Now, just for the sake of it, let’s look at a comparison of two shots taken at 6400 ISO with these same two sensors (Figure 8).  As you can see there is definitely a ton of noise in these shots, with slightly less generated by the 7D Mark II.  And, it would have been even worse if the shots had been originally under-exposed, in which case some noticeable color banding and splotching would likely have been created in the darker regions of the shot.

I wouldn’t feel too comfortable in trying to reproduce these shots, and if I did would have to spend some time to further reduce the noise (although this surely does beat the graininess we experienced in the film days when shooting above 400 ISO!!).  You can also see the degradation of detail, contrast and saturation relative to the previous examples that you get when you start shooting at such high ISO settings.  I might add, that this would generally be true of full-frame sensors at such high ISO settings, although the actual amount of noise and posterization would be considerably less than that generated by a crop sensor.

Figure 8. Two shots taken at 6400 ISO by the original EOS 7D (right) and EOS 7D Mark II (left).

Figure 8. Two shots taken at 6400 ISO by the original EOS 7D (right) and EOS 7D Mark II (left), demonstrating the relatively large amount of noise that’s generated at such high settings.  Click the image for a larger version if it’s not displayed large enough on your monitor.

Figure 9 shows a shot taken with the 7D Mark II where I purposely over-exposed the shot by about +2/3 of a stop (left photo) with no image adjustments applied, paired against a virtual copy of the same image (right photo) where I’ve rebalanced the exposure using basic exposure adjustments, and also applied a slight amount of capture sharpening.  This figure demonstrates what I call “Expose-To-The-Right” or ETTR technique.

The goal here is to purposely push the histogram slightly to the right, thereby capturing more luminance levels in the dark mid-tones and shadows, while not blowing-out the whites, and then correct the exposure during post-processing to reduce the noise generated in the dark regions of the image.

Figure 9. One shot taken by the 7D Mark II. Left image is the original raw file, over-exposed by +2/3 EV, and the right image is a virtual copy with basic exposure adjustments applied.

Figure 9. One shot taken by the 7D Mark II. Left image is the original raw file, over-exposed by +2/3 EV, and the right image is a virtual copy with basic exposure adjustments applied.

This technique works even better when you have extremely dark, almost black, regions dominating the scene where you want to reveal more shadow detail without opening them up during post-processing, thereby introducing a lot of noise and color banding.  You can’t use this technique if you have highlights that are almost or already blown out that you don’t want to clip, although you can still use ETTR technique under such circumstances if you take and blend multiple shots together during post-processing.

Figure 10 shows the same images at 100% magnification.  This isn’t the best example of the benefits of ETTR technique, but if you look closely there is relatively little noise in the shot, and slightly less in the shadow regions of the corrected virtual copy.  Thus, if able to employ even slight ETTR technique, I can get even better image pixel quality when shooting at higher ISO settings.  In the past I could not get this clean output with the original 7D or other crop sensor cameras at higher ISOs without very bright lighting.

Figure 10. Same shot in Figure 9 shown at 100% magnification, demonstrating slight improvement in noise in the shadow tones.

Figure 10. Same shot in Figure 9 shown at 100% magnification, demonstrating slight improvement in noise in the shadow tones.  Click the image for a larger version if it’s not displayed large enough on your monitor.

Figures 11 and 12 show 100 ISO image comparisons among the 7D Mark II and 5D Mark III, at fit-to-view, and 100% view, respectively.  Again, the intent here isn’t to compare image pixel quality in terms of signal to noise ratio (that’s pointless), but demonstrates why there can be an advantage to using a crop sensor at equivalent full-frame focal lengths in terms of depth-of-field and resolution of detail.

Figure 11. Comparison of 100 ISO shots taken with the EOS 7D Mark II (left) and 5D Mark III (right).

Figure 11. Fit-to-view comparison of 100 ISO shots taken with the EOS 7D Mark II (left) and 5D Mark III (right).  Very comparable light capture ability of both sensors, but the 7D Mark II still seems to be capturing light slightly better.

Note in Figure 11 that the full-frame version (photo on the right) does a better job collecting light compared to the original 7D as shown in the right photo in Figure 5, and this shouldn’t be surprising.  But, what was surprising to me is that the 7D Mark II’s sensor seems to collect light a little better (or applies a slight increased gain to the exposure) compared to the 5D Mark III.  The improvement in sensor efficiency with the 7D Mark II is significant, even rivaling that of many full-frame sensors, and hint at what Canon may be offering with its future sensors.  For more on this, view the 7D Mark II review by Tony Northrup mentioned near the beginning of this blog post.

It should also be noted that the dynamic range of both cameras is essentially the same (about 11.8 EVs), although as you increase ISO the loss of dynamic range will be greater for the crop-sensor compared to the full-frame sensor.  Finally, the better light capture of the 7D Mark II was achieved with a faster shutter speed compared to the 5D Mark III, and this can be an important factor when shooting fast moving subjects….the faster shutters the better, especially when shooting with long focal lengths (assuming of course you don’t want vibration or motion blur in the shot).

Anyway, upon inspection of the 1:1 view in figure 12, both shots are essentially noise-free for all practical purposes.  There appears to be more noise in the image to the left taken with the crop-sensor, but actually that’s because there is more detail due to a greater proportion of the photo being in or near the focal plane as compared to the full-frame photo on the right.  This is caused by the difference in sensor size, with higher relative density of pixels in the crop-sensor, and it’s different angle of view of the photosites.  Since an APS-C sized dSLR has 1.5X to 1.6X more depth of field, or 50-60% less background blur than a full frame camera for any given f-stop, in this example the same lens on a Canon crop-sensor at f/11 behaves as if I was using f/17.6 on the full-frame sensor.

Of course I can get identical depth-of-field (or background blur depending on how you’re looking at it) among the two sensor sizes by adjusting the aperture size to account for the crop factor.  Although, I generally won’t need to do that unless I’m doing portrait or close-up work.

Figure 12. 1:1 View of the shots shown in Figure 11, showing excellent IQ of both sensors, as well as increased DOF and resolution of detail in the crop-sensor (left photo).

Figure 12. 1:1 View of the shots shown in Figure 11, showing excellent IQ of both sensors, as well as increased DOF and resolution of detail in the crop-sensor (left photo).  Click the image for a larger version if it’s not displayed large enough on your monitor.

In general, when I’m shooting wildlife, the depth of field is often quite shallow and a 50-60% increase in apparent depth (resolution of detail) is often a good thing.  This can be seen in Figure 12 where the greater detail is particularly apparent on the back of the neck of the turkey mount, where it’s out of the focal plane in the full-frame sensor, and in focus in the crop-sensor.  This is a universal trade-off involved with full-frame versus crop-sensors, and for more on this, view the digital camera sensor size review at Cambridge in Colour.  You might get better IQ from the full-frame in terms of signal to noise ratio, especially at high ISO, but it will come at the expense of detail and depth-of-field unless you stop down the aperture.  In the example shown, if I needed the entire neck of the bird in perfect focus, then the shot taken by the full-frame sensor would be unacceptable at the equivalent focal length I needed unless I would have decreased the aperture size, and doing so would have required a compensatory shift in ISO (higher, and hence more noise) or slower shutter speed (and slower shutter speeds, which could introduce more motion or vibration blur).

So, for one final image comparison let’s check out two shots taken by these same two camera bodies at 1600 ISO (Figure 13).  The same trade-offs hold at the higher ISO.  Yes, the full-frame performs better in terms of pixel IQ, as we would expect because it’s gathering a little more than twice the total light across the sensor as compared to the crop sensor.  But, the crop-sensor image on the left is of more than sufficient quality for reproduction, and more importantly, crucial parts of the subject are in the focal plane.  The noise at 1600 ISO in the crop-sensor is slight enough where some creative masking of noise reduction in photoshop could easily make the final result very clean if I wanted close to no noise showing whatsoever.

Figure 13. 100% view shots at 1600 ISO, with EOS 7D Mark II photo on left, and 5D Mark III photo on the right. Demonstrates difference in noise, as well as DOF and detail trade-offs among the two sensors.

Figure 13. 100% view shots at 1600 ISO, with EOS 7D Mark II photo on left, and 5D Mark III photo on the right. Demonstrates difference in noise, as well as DOF and detail trade-offs among the two sensors, but the crop-sensor image is of sufficient quality for reproductive purposes, meaning this camera is a strong contender for anyone needing a combination of good IQ and high-grade responsiveness and auto-focus at a reasonable price-point.  Click the image for a larger version if it’s not displayed large enough on your monitor.

Therefore, if you can’t afford a full-frame sensor or pro-grade flagship model with more expensive telephoto lenses, but still require a performance rig for action or wildlife photography, the new 7D Mark II crop-sensor dSLR is a big contender (in my opinion by far the best on the market in its class as of November 2014), especially considering it’s high responsiveness, excellent AF system, good ISO performance at a moderately high ISO, and at a price-point of only $1,799.

I hope you found this information useful.  If you have any other questions on my experience thus far with the new 7D Mark II, please leave a comment or contact me at ImagesByBeaulin@charter.net.

Cheers, and happy shooting…..Beau

© Beau Liddell, ImagesByBeaulin.com, All rights reserved.