When you think of a 35mm sensor camera, what characteristics come to mind? Probably the first is shallow depth of field. It is part of what many consider the “look” of that imager size and a principle defining difference between it and other formats. But while some consider this the chief attribute of the format, others may find it to be a detriment in certain situations. For those shooting documentary, run-and-gun productions, a deeper depth of field may be preferable to maintain focus. Fortunately, there is another major facet to the latest crop of 35mm sensor cameras – a very low noise floor. This means that in a camera such as the Canon C300, a wide range of ISO settings can be applied with little objectionable artifacting to the image. Using the concept of the “variable depth of field camera,” you can adjust the available parameters of the camera to yield the depth of field desired for any given image. Watch the video to see this concept demonstrated.

24p is ‘jumpy’ by its nature, but the look of 24p video in a digital video camera can still come across a little more ‘jumpy’ then you’d expect. There are several reasons for this, which I explain in article for HDVideoPro magazine called “Did I Judder.” As part of the article, I put together these two videos to show the difference between native 24p video and 24p converted to 60i. I did this because much of the 24p content we see has been converted to 60i already, which has a smoothing effect that is very noticeable on television.

Watch the above video to see true 24p and then the video below in 60i. The difference is subtle on a computer monitor, so it’s difficult to show online. However, the difference helps explain why so many of us are surprised by 24p judder. To learn the other reasons for 24p judder, check out the full article over at HDVideoPro.com and on magazine shelves now.

In my last entry, I discussed the 3-way trade off between quality, size and complexity in codecs. Panasonic’s move from DVCPRO HD to AVC-Intra in their latest generation of P2 cameras provides an excellent example of a trade off between quality and complexity, as both codecs record at the same bit rate of 100mb/sec (at 30fps). How much better is AVC-Intra? To start off with, it preserves a lot more image information than DVCPRO HD. It records full raster (1920×1080, when recording 1080p) 4:2:2 at 10 bits per channel compared to 1280×1080 4:2:2 at 8 bits per channel, which means it’s starting out with almost double the data. Here’s a comparison of a frame of each (the same difference frames we looked at last time).

This is DVCPRO HD. Notice the lines around any fine detail. That’s the loss of sharp edges created by the reduction of the frame width from 1920 to 1280 pixels.

And AVC-Intra:

Clearly there’s a dramatic improvement in quality. So what’s the trade off? Well, I can comfortably play back DVCPRO HD footage on my old 1GHz G4 Powerbook. To play back AVC-Intra requires a nearly top-of-the line Mac Pro. It’s certainly a worthy trade off for an acquisition format, but due to its performance advantage, DVCPRO HD still makes a superior offline editing codec.

Note: While Panasonic is moving from DVCPRO to AVC-Intra, Sony is also migrating to a new generation of codecs, XDCAM and HDCAM SR. Sony’s made a different (and more complicated) set of compromises with their codecs. With XDCAM they’ve made a much more modest improvement in quality over HDCAM, but with a much higher compression ratio. With HDCAM SR, they’ve made quality the first priority (and it shows: see the difference frame from my last entry).

For more information, visit ibe-optics.com.

Image compression (really any kind of compression) involves a tradeoff between space or bandwidth (how much data you need to store or move around), complexity (how much computing power is needed to encode or decode the material) and quality. The very idea of compression implies a compromise of some sort, or we’d only be dealing with uncompressed files in the first place.

Every codec balances these qualities differently and, as computing resources, bandwidth and storage become more abundant, the decision on which codec to use for which purpose will need to be reevaluated periodically. (For an excellent example of this, consider Panasonic’s transition from DVCPROHD to AVCI, which provides improved image quality at the same bandwidth, at the cost of increased complexity. More on this in a future installment.)

So, how do we go about evaluating a codec? Well, evaluating the space and complexity (or at least their effects in your environment) are relatively straightforward – just compress some footage and monitor the size and resources used. But how do we judge quality?

This is what I do: I start with some uncompressed footage. I have a few clips set aside that I use to compare codecs. If you’re planning on doing codec testing for a specific project, make sure you pick some footage with attributes that are reflective of what you will encounter, as many codecs work better with certain types of footage than others (the greatest factor will be the amount of detail in the original footage, but long-GOP codecs will also be affected by the amount of difference between frames. Also, different codecs will down-sample the luma and/or chroma components of the image which will affect sharp color transitions). I then compress the footage in the codecs to be tested so I have several versions of the same clip – an uncompressed version, as well as various compressed versions. After looking at the different clips to see if anything jumps out at me, I then look at only the difference between the original and compressed versions of the clip, magnified to enhance any difference. (I use Shake for this. Here’s what my Shake node tree looks like)

I can now step through the clips and see how each codec handles different challenges, such as fine detail and fast movement. Here are some examples:

Here’s a crop from the center of our original image (Note that for the sake of bandwidth all images in this post have been cropped and compressed as high quality JPEGs. I think the quality is enough to make my points, but if you want to see the original images contact me and I’ll arrange it.):

Here’s the amplified difference image for Avid DNX 115:

Apple ProRes:

Avid DNX 175:

Apple ProRes HQ:

Sony HDCAM SR:

Notice that the more lossy a codec is, the more detail shows up in the difference image. Any sort of scaling in a codec (due to sub-sampling in the luma channel) creates very obvious loss around any fine detail. Other artifacts such as macro blocks and ringing also become clear.

Note 1: As an example of why it’s important to look at the original images in addition to the difference image, check these images out. I’ve deliberately created these images for illustrative purposes, but I’ve come across exactly this situation with real codecs. First the differences:

Example 1 Amplified Difference

Example 2 Amplified Difference

Notice that the second image is somewhat lossier than the first. But take a look at the actual images:

Example 1 Original

Example 2 Original

Look at the maroon carpet and the rear surface of the cart in both images. In fact, despite the fact that the second image is lossier (I’ve scaled it and added a fair bit of noise), it is still a much better looking image than the first.

Note 2: it is possible to condense the entire difference image down to a single number called the Peak Signal-to-Noise Ratio (or PSNR for short). This can be useful for certain purposes, but I find that looking at the actual image is much more informative. Firstly, the PSNR number itself is meaningless. It depends on the signal being compressed, and there are a few slightly different ways of calculating it resulting in different numbers, so it is only useful as a comparator for other compressed signals originating from the same source material. Secondly, it doesn’t tell you exactly what is being lost. Often what is being lost is more important than the overall level of lossiness.

Note 3: This is for the geeks out there. I noticed something interesting when preparing the images for this post. I compressed all of the examples as fixed quality JPEGs, and it turns out that if you sort the difference images by size (in decreasing order), they will sort in order of PSNR or the original image (lowest to highest). This makes sense because the compressibility of a signal is inversely proportional to the amount of information in contains, so the more detail in an image the less compressible (without signal loss) it is.

The Big Picture

If you make your living in the film or television production business, you probably have a pretty good intuitive sense of what kinds of pictures you like. You probably have also developed a vocabulary to describe images and the qualities that you’re looking for. They may be warm or cool, punchy or flat. You may describe pictures that have something wrong with them as being thin, noisy or plasticky. Sometimes you may describe certain pictures as being filmic or video-like.

Generally, there are a number of factors that create a particular look — the way colors are reproduced, the sharpness of edges, noise or grain, etc. One of the main goals of the testing that I do is to identify those specific qualities in an image that evoke a certain feeling.

There are many tests that produce easily quantifiable results: more resolution is generally better than less, a faster download is always better than a slower one, etc. However, many of these factors tend to counteract each other. For instance, all things being equal, higher resolution tends to imply lower sensitivity. When evaluating a product, it’s important to take into consideration the specific applications for it and find the best compromise for your use.

The Fine Details

Each further entry in this series will concentrate on an individual realm of testing. Some topics will include resolution and sharpness, compression, dynamic range and noise. With all of these subjects we strive to follow certain principles:

Specificity: Because there are so many factors that go into making an image, it can be difficult to isolate which individual factors are affecting the image in which ways. Therefore, we design tests that isolate these factors as much as possible.

Reproducibility: In order for tests to be meaningful, it is necessary for them to be reproducible. We need to make sure that the same tests are performed exactly on different equipment. If we perform the same test on the equipment, then we expect to get similar results or we know something is wrong with the test.

That Certain Something

With all of the concentration on objective tests, it’s easy to lose sight of the real purpose of the equipment we use – to present our artistic vision and provoke an emotional response. It’s important when we’re evaluating a piece of gear, that we keep our eyes open for the intangible qualities, whether it’s in the images created by a camera, or the way a piece of gear feels to use.

As an example, consider these two photographs of the same subject, taken at about the same time.

The first picture was taken with a digital camera, and by almost any measurement it is technically superior to the second. The second picture was taken with a cheap plastic toy camera, yet it is much more successful as a photograph. After all the testing is done and the numbers are in, we still need to look at the gear we review with an artistic eye. There are some qualities that just can’t be measured by any scientific method. So our final step in any evaluation is to determine if the piece of gear has a certain quality that appeals to our aesthetic values.