In-Camera Noise Options

Facebooktwittergoogle_pluslinkedin

Today’s Question: As a follow-up to the May 12 question about in-camera versus post processing long exposure noise reduction, do you have the same opinion on high ISO noise reduction?

Tim’s Quick Answer: No, there is an important distinction between long exposure noise reduction and high ISO noise reduction in the context of noise reduction applied in the camera. The high ISO noise reduction in the camera does not directly affect RAW capture data. Results will vary, but in general I find that I prefer to apply the high ISO noise reduction in post-processing rather than in the camera.

More Detail: The long exposure noise reduction I referred to in the May 12th edition of the Ask Tim Grey eNewsletter does apply to RAW captures. The second “dark frame” exposure is used to subtract noise from the image in the camera, changing the information in the RAW capture.

High ISO noise reduction operates differently. For JPEG captures, this high ISO noise reduction will of course affect the original JPEG capture. However, for RAW captures the high ISO noise reduction will not affect the actual data recorded by the image sensor. Instead, this information will be added a special metadata within the image. However, in general the only way to make use of that special metadata regarding noise reduction is to use the software provided by your camera’s manufacturer to process your RAW captures.

In other words, if you are using software such as Lightroom or Adobe Camera Raw to process your RAW captures, the high ISO noise reduction settings in your camera will not apply to those RAW captures at all.

As noted above, for the limited number of cameras I have tested for in-camera high ISO noise reduction, I have found that I am happier with the results I can achieve with post-processing noise reduction. For example, Lightroom and Adobe Camera Raw now provide excellent noise reduction, which I have generally found to be better than the in-camera noise reduction that is available.

So, while I do recommend making use of long exposure noise reduction in the camera, I generally don’t recommend using in-camera high ISO noise reduction. Note that even when you have applied long exposure noise reduction in the camera you will likely want to apply additional noise reduction when processing your photos after the capture.

Double Exposure Time

Facebooktwittergoogle_pluslinkedin

Today’s Question: In your answer about in-camera noise reduction you explained that a second “black frame” is captured. Doesn’t that mean you are essentially doubling the time it takes to capture a photo? Wouldn’t a 30-second exposure then take a full minute to actually capture? And isn’t that a good reason not to use in-camera noise reduction?

Tim’s Quick Answer: Yes, in-camera noise reduction does double the total time required to capture a photograph, but in my mind it is well worth the extra time in terms of the improvement in quality (relative to noise levels) in the final capture.

More Detail: As noted in the previous edition of the Ask Tim Grey eNewsletter, in-camera noise reduction generally operates by capturing a “black frame”, which is essentially a duplication of the “real” capture with the shutter closed to prevent light from reaching the image sensor.

This black frame provides information about how the image sensor is performing in terms of noise in the same overall conditions as the photo you are capturing. That information is then used by the camera to subtract the noise from the capture.

To be sure, there are situations where it is more important to be able to work as quickly as possible, or to be able to capture more photos within a given timeframe. But when it comes to overall image quality I consider in-camera noise reduction to be tremendously helpful for long exposures.

You will generally find that in-camera noise reduction is not employed until you reach exposure times of around thirty seconds or more. As a result, the impact of the second black frame capture is somewhat significant in terms of total time. It is also significant, however, in terms of the reduction of noise in the capture.

So, recognizing that using in-camera noise reduction essentially doubles the amount of time required for each long exposure capture, I do recommend making use of this feature whenever time allows.

Noise Reduction Options

Facebooktwittergoogle_pluslinkedin

Today’s Question: I would like to know the difference between the long exposure noise reduction option in camera and the noise reduction with Lightroom or Photoshop. What is the best option for long exposure with a 10-stop ND filter?

Tim’s Quick Answer: These two types of noise reduction are actually fundamentally different, with the in-camera option relating to how the camera actually behaves while the post-processing option can only analyze pixel values in the photo. For long exposures I recommend employing both in-camera and post-processing noise reduction.

More Detail: The in-camera noise reduction option provides a unique advantage, in that it is able to compensate for the actual behavior of the image sensor at that specific time under the current conditions.

As you may be aware, there are a variety of factors that can impact overall noise, including the design of the image sensor, the duration of the exposure, and the amount of heat buildup on the sensor. In-camera noise reduction provides the best opportunity to compensate for these various factors. In most cases this function operates by essentially capturing two exposures. First, your actual exposure is captured. Then a “black frame” is captured, where a photo is captured with the same exposure duration, but with the shutter closed to prevent light from reaching the sensor.

This black frame exposure can then be used to determine the noise behavior of the image sensor under the current conditions, so that the camera can then process the actual capture to subtract out the noise. This is, of course, a rather sophisticated operation, and it can be very effective at reducing the noise in the initial capture.

When you are applying noise reduction in post-processing, you only really have the pixel information to work with. Thus, noise reduction software uses a variety of techniques to evaluate and reduce the appearance of noise in the photo.

Since both in-camera and post-processing noise reduction employ a different approach to reducing noise, and since they compensate for different limitations, I recommend using both of them. I generally consider in-camera noise reduction to be the more important of the two, but there will likely be some degree of problematic noise remaining in the image even after in-camera noise reduction has been applied. The careful application of additional noise reduction in post-processing can help ensure the best image possible from the perspective of noise.

Maximum Frame Rate

Facebooktwittergoogle_pluslinkedin

Today’s Question: Can you explain why a digital SLR that is limited to somewhere around ten frames per second for photos is able to shoot sixty frames per second for video?

Tim’s Quick Answer: There are two key factors here, which I’ll over-simplify in an effort to keep things as clear as possible. First, video is typically shot at a lower resolution than still photos, so there is less data to process per frame. Second, for still photos there is generally a lot more work being done by the camera, such as moving the mirror and potentially establishing autofocus for each frame.

More Detail: Let’s assume a “typical” digital SLR that has a 20-megapixel image sensor. That means that each photo you capture will contain 20 million pixels. If we assume a frame rate of ten frames per second, that translates into 200 million pixels recorded per second.

If we then assume full high definition (HD) video at 1080p, we would be recording frames that are 1920 by 1080 pixels in overall size. That translates into just over 2 million pixels per frame. Even if we assume a higher-than-typical frame rate of 60 frames per second, the total number of pixels being recorded per second with video is “only” about 124 million pixels. At 30 frames per second that value goes down to about 62 million pixels per second. So, fewer pixels being processed per second with video.

Of course, things do get a bit more complicated. For example, the new Canon EOS-1D X Mark II (http://timgrey.me/atg1dx2) supports 4K video up to 60 frames per second, while capturing still images with a 20.2 megapixel image sensor. The resolution of the 4K video in this case is 4096 by 2160 pixels. That is “only” about 8.8 megapixels. However, when shooting that video at 60 frames per second you are processing more pixels per second (about 530 million) that you would for still photos captured at the maximum frame rate of 16 frames per second (about 319,000 pixels per second).

However, there are additional complexities here that make this possible. As noted above, there is generally more work being done by the camera when capturing individual still images in sequence. In addition, the “more” data being recorded for video doesn’t always translate into more data being stored on a memory card.

In the case of the Canon EOS-1D X Mark II, 4K video is captured in the Motion JPEG format, which employs compression to reduce overall file size as well as the overall amount of data being handled. With HD (1080p) the result is an MPEG-4 video.

So again, in general the answer here is relatively simple, because in most cases video represents less information being processed compared to still photos with a typical digital SLR. But, of course, there are more complicated issues involved in some cases, especially when it relates to how the video frames are processed and what specific tasks the camera must perform for still photos compared to video.

Folder Mismatches

Facebooktwittergoogle_pluslinkedin

Today’s Question: About a week ago I happened to notice that the folders and subfolders I see in the Lightroom Library view are not matched by what I see in the Finder view of my folders and subfolders on the drive where they are stored. That is, folders and subfolders (as well as some photos) that I have moved and deleted within the Lightroom Library module are still appearing in their previous locations in the Finder view of my external drive. In addition, in the Import dialog in Lightroom, the folder hierarchy matches the one I see in the Finder, not the one in the Lightroom Library module. Am I seeing normal Library behavior?

Tim’s Quick Answer: This does sound like perfectly normal behavior in Lightroom. The key is to understand how Lightroom handles the removal of photos and folders within your Lightroom catalog, and to recognize that Lightroom’s behavior varies based on the circumstances.

More Detail: In this case, it sounds like the primary source of confusion relates to removing versus deleting folder and photos.

When you choose to remove one or more photos from your Lightroom catalog, you have the option of removing or deleting the photos. In other words, you can remove the photos from the Lightroom catalog but keep them on your hard drive, or you can actually delete the photos from your hard drive (removing them from your catalog at the same time). This option is available in the confirmation dialog that is presented when you choose the Remove Photos command.

My personal view is that if you are going to remove photos from your Lightroom catalog, you should probably delete them from the hard drive at the same time. After all, if you are using Lightroom to manage your photos, the only way you’re likely going to discover those photos is within Lightroom.

Note that if right-click on a folder that contains photos and choose the Remove command, the folder and photos within that folder will be removed from your Lightroom catalog, but neither the folder nor the photos will actually be deleted from your hard drive.

If, on the other hand, you delete all of the photos within a folder and then choose the Remove command for that folder, the folder will be removed from your Lightroom catalog and deleted from your hard drive.

In other words, there is a bit of variability in terms of the behavior of the options for removing photos and folders within Lightroom. It is possible to delete folder and photos from within Lightroom, but it is also possible to remove photos and folders from your catalog while leaving the actual files and folders in place on your hard drive.

If the folders on your hard drive that are not reflected in Lightroom are indeed empty or contain photos you don’t want to keep, you can delete those folders and photos. If, on the other hand, you want those images included in your Lightroom catalog, you can use the Import feature with the Add option to add those existing folders and photos back to your catalog.

As for the Import feature, that will reflect the actual contents of whatever storage device (such as a hard drive) you browse within Import. The Library module in Lightroom only presents folders and photos that are actually being managed by Lightroom. Any folders or photos that were “left behind” based on the use of the Remove command will still be visible in the Import dialog just as they are visible within your operating system.

Lightroom Storage Upgrade

Facebooktwittergoogle_pluslinkedin

Today’s Question: My Lightroom Catalog and Previews are on my main internal hard drive. I wish to keep them there. Therefore, I only want to move my photo files to a larger external hard drive. Is it correct to assume that I close Lightroom; then after copying all of my photo files to the new external hard drive, I give it exactly the same name as the old external hard drive? Having done that, when I next open Lightroom, will it automatically find my photo files?

Tim’s Quick Answer: Yes, it is possible to “upgrade” to a larger hard drive for your photos, while keeping your Lightroom catalog (and therefore previews) on the internal hard drive on your computer. The process you described in today’s question is one way of accomplishing this task.

More Detail: The key to this approach is that Lightroom is looking for your photos based on the path to the photos referenced within the Lightroom catalog. If you “upgrade” to a larger hard drive to store your photos, you can simply make that hard drive appear with the same path as the “old” hard drive, and Lightroom essentially won’t know the difference.

The first step is to copy all of your photos from the existing external hard drive to the new drive. You can quite literally copy all of the contents of one drive to the other, or use synchronization software such as GoodSync (http://timgrey.me/greybackup) to perform this task.

Once you have copied all of the folders and photos from the “old” drive to the “new” drive, you’ll need to update that drive. First quit Lightroom, and then disconnect the “old” hard drive from the computer. Then update the “new” hard drive so that it matches the path of the old drive.

For Windows users, this means changing the drive letter (which you can do in the Computer Management application) to the same drive letter that had been used for the old hard drive. For Macintosh users you simply need to update the volume label for the new drive to match that of the old drive, which you can accomplish by right-clicking on the new drive and choosing Rename, then applying the same name as the old drive.

At this point the path to all of your folders and photos will appear exactly as those paths had been before the storage upgrade. You can then launch Lightroom and you will find all of your photos available just as if nothing had happened at all. Except now you’ll have more space available on the new larger hard drive.

This process was also outlined in the August 2014 issue of Pixology magazine, in the article “Step by Step: Storage Upgrade”.

Painting a Color Fix

Facebooktwittergoogle_pluslinkedin

Today’s Question: I’m looking for a tutorial that you did on removing what I’ll call a color cast from a portion of a picture. What I remember was a horse in a field with a partial greenish color cast caused by vegetation that was in the way when the picture was taken. Do you happen to remember where I may have seen this?

Tim’s Quick Answer: I don’t recall where that lesson may have been published, but it is a technique I have taught at a variety of workshops. The basic process involves painting with the desired color on a new layer in Photoshop, employing the Color blend mode.

More Detail: The first step is to add a new empty image layer directly above the Background image layer. To do so, first click on the thumbnail for the Background image layer. Then click the “Create a New Layer” button (the blank sheet of paper icon) at the bottom of the Layers panel.

To help stay organized I recommend renaming the new layer to something like “Color Fix”. To rename, simply double-click on the name of the layer, type a new name, and press Enter/Return on the keyboard to apply the change.

Next, change the blend mode for this new layer to Color. The default blend mode is Normal, which you’ll see reflected on the popup near the top-left of the Layers panel. Click that popup and choose Color.

You can now choose the Brush tool, and configure the tool to use a soft-edged brush (a low Hardness setting). I generally use a simple round brush, but you can use different shapes if it is helpful for your painting.

To select the color you want to paint with, hold the Alt key on Windows or the Option key on Macintosh and click within the image to select the desired color. For example, one of the images I have used to demonstrate this technique featured a horse with some green foliage in the foreground causing a color cast. In this situation I would hold the Alt/Option key while clicking on an appropriate color from the horse.

With the desired color selected, you can now paint in the area of the image that exhibits the problematic color cast. Note that to produce the best result you’ll generally need to sample a variety of different colors from the image while painting.

For example, if the horse in my sample image was brown, there would actually be many different shades of brown found within the horse. Therefore, I couldn’t simply choose a single shade of brown and paint over the entire area requiring color correction. Instead, I would sample various different shades of brown as I was working to resolve the color cast.

This technique is quite simple to implement, but can help you produce excellent results even with complicated color problems in the image. In effect, you are retaining all of the underlying texture while only altering the actual color appearance of the image, thanks to the Color blend mode.

Selection Merge Confusion

Facebooktwittergoogle_pluslinkedin

Today’s Question: I created a layer mask [in Photoshop] to separate my subject from the background so that I could apply more image noise reduction to the background than what the subject will get. I am also using the same mask to blur the background but not the subject. If I combine the two layer masks into a single selection, the entire image is not selected. How can that be?

Tim’s Quick Answer: The short answer here is that adding two selections together doesn’t produce a result where the luminance values in the shades of gray of a selection are exactly added together. In effect, adding luminance values of 50% plus 50% does not yield 100%, but rather 75%.

More Detail: In most cases this behavior is not a real problem, because the specific impact of a particular shade of gray on a layer mask isn’t critical. The result achieved with white on a layer mask (the full effect is applied) is important. And the result achieved with black on a layer mask (none of the effect is applied) is important. Shades of gray in between tend to be less important.

Still, this behavior can be problematic when combining different layer masks or selections together. For example, let’s assume a simple selection that is 50% selected. If you invert the selection that area would still be 50% selected. But if you merge the two selections together, you don’t end up with 100%.

In other words, merging two selections together, where the second selection represents the inverse (the exact opposite) of the first selection will not produce a result where all pixels in the image are selected.

I’m sure there is some complicated mathematical explanation for why this situation makes sense, rather than what we might consider a more “logical” approach to combining selections and layer masks. But the bottom line is that if you want to select the entire image, you should use the “Select All” command found on the Edit menu, rather than combining existing selections or layer masks.

GPS Accuracy

Facebooktwittergoogle_pluslinkedin

Today’s Question: In a recent episode of Tim Grey TV you made reference to the GPS receiver in your camera. What happens in a situation where you don’t have a reliable GPS signal? Do you get inaccurate location information?

Tim’s Quick Answer: When there is not a reliable GPS signal, the GPS coordinates embedded in metadata may be inaccurate, or they may be missing altogether. Therefore, it is important to keep in mind that even when location information is embedded in your photos automatically by your camera, the information may not be especially accurate.

More Detail: The implantation of the GPS feature varies among the different cameras that include a GPS receiver. My primary camera (in part because it includes a built-in GPS receiver) is the Canon EOS 7D Mk II (http://timgrey.me/tgtv7d2). This camera (similar to other cameras) will present an indication that the GPS feature is enabled, and that a signal has been acquired.

A variety of factors can impact the accuracy of a GPS signal, so even when you have a “good” signal the resulting coordinates may not be completely accurate. The relative accuracy of a GPS signal is often expressed as a distance. For example, a good signal might represent accuracy within 3 meters, while a not-so-good signal might represent accuracy within dozens of meters or more.

During my current travels I actually experienced several situations where a reliable GPS signal could not be achieved. This happens indoors, of course, but while photographing along the inside passage in Alaska there were a couple of situations where no GPS coordinates were recorded by my camera. For example, when in the Russell Fjord photographing Hubbard Glacier, the surrounding landscape was high enough to interfere with acquiring an adequate GPS signal, and so no coordinates were embedded in the metadata for my photos.

In other situations I have observed GPS coordinates that were not very accurate. In some cases the coordinates may have shown that the photo was captured down the street from the actual location, and in other situations I’ve seen coordinates that were off by more distance, perhaps around half a mile or so. Various factors can impact the relative accuracy of the GPS signal, and different cameras will handle the situation differently. But again, there are circumstances where no GPS coordinates will be embedded in your photos even when that feature is enabled, as well as circumstances where the coordinates that are embedded are less accurate than you would normally expect.

In situations where it is important that you know the location where photographs were captured with some precision, I do recommend reviewing the location information for your photos. You can do this, for example, in the Map module in Lightroom, even moving photos on the map to update GPS coordinates as needed based on the map.

Extreme Shutter Speed

Facebooktwittergoogle_pluslinkedin

Today’s Question: I appreciate that you include the camera settings with your photos when you share them, as I find that information informative. However, I’m confused about the settings for one of the images you shared on 500px:

https://500px.com/photo/151520179/mountains-at-hubbard-glacier-by-tim-grey

Your metadata shows that the shutter speed was 1/8,000th of a second. Why would you use such a fast shutter speed for a subject that wasn’t moving?

Tim’s Quick Answer: This was essentially a case of “insurance speed”. I was on a moving ship (though moving slowly), was using a long telephoto lens (600mm focal length for the photo in question), and I was shooting hand-held. Thus, I wanted the fastest shutter speed I could reasonably achieve to help ensure a sharp photo under the circumstances.

More Detail: The photo shown on my 500px page at the link above was captured near Hubbard Glacier in Alaska. We had started to depart the Russell Fjord when I spotted the snow-covered peaks. The ship had already been moving, and I was using a Tamron 150-600mm lens (http://timgrey.me/1WyqJ4n), frequently zooming all the way to 600mm.

I was not using a tripod at this point, in large part because the ship was already moving. The vibration of the engines can be transmitted through the ship itself, through the tripod, to the camera, causing blurring. In addition, maneuvering on the deck of the ship can be a challenge with a tripod. Although the sky was overcast, the layer of clouds was relatively thin and so I was able to achieve fast shutter speeds. So I opted to work without a tripod.

When shooting hand-held, the general rule of thumb is that you want the reciprocal of your shutter speed to match (or exceed) the focal length of the lens. So at 100mm I would want about a 1/100th of a second shutter speed, and thus at 600mm I would want about a 1/600th of a second shutter speed. However, with super telephoto lenses (even a lens that has image stabilization technology built-in, such as the lens I was using) I prefer to be far more conservative.

Because there was relatively good lighting even with the overcast sky, getting a fast shutter speed didn’t require much compromise. I was using an aperture of f/8 in an effort to ensure the sharpest images possible. I wanted the fastest shutter speed I could get, and by boosting the ISO setting up to only 400 I could achieve the maximum shutter speed (1/8000th of a second) for my camera.

I am perfectly comfortable with the degree of noise produced by the camera I was using (a Canon EOS 7D Mk II, http://timgrey.me/tgtv7d2) at an ISO setting of 400. So in this case the decision was easy. I could shoot at f/8 with an ISO setting of 400 and achieve a fast 1/8000th of a second shutter speed. That worked out well, and more importantly the scene was very nice, with snow-covered peaks seeming to blend in to the overcast sky above.