8mm, Super 8mm, and 16mm film to digital video tutorial.

Concepts and terminology of film to video and/or DVD

Resolution

By: Bruce Mayfield

A fundamental concept in all imagery science, is "resolution". In effect, resolution is visual information or visual data that makes up an image. The more information you have, in an image, the more detail you can actually see. The more data you can see, the more realistic -- true to life -- the image appears. When both 1) measuring and 2) talking about resolution, "more" (i.e., "higher") resolution is better. The "higher" the resolution, better the image looks.

In the film world, a larger piece of film (i.e., 35mm slide film) has more "resolution" than a smaller piece of film (i.e., 8mm movie film).  When comparing the film world to the TV world, film -- no matter what size -- is better than TV.  This is one reason the 'big screen" of the theater still rivals that of the "big screen TV".

Your 8mm film has "higher resolution" than your TV set. Your objective in a purchasing a film to video transfer, is to capture ALL the resolution of your original film.

In the analog world (your TV set), resolution is measured in "lines of resolution".  There are "vertical lines of resolution" and "horizontal lines of resolution".  In the digital world, resolution is measured in "dots or pixels per inch".  Dots or pixels per inch are expressed in for the vertical and horizontal axis of the display device, too.

To illustrate a layman's discussion of "resolution"  -- so even Grandma & Grandpa can understand it:

1. We will assume that "lines of pixels" and "lines of resolution" are the same. (They are not*)
    (for more information, see tech notes below)
2. We will assume that "dots" and "pixels" are the same. (They are not.**)
3. We will assume that "dots" and the "grains" in your film are the same. (They are not.***)

To get a basic feel for resolution as it relates to video: If you're in North America, your television set's picture is NTSC and has an effective resolution of 525 pixels by 480 pixels (Picture Element, or one dot of the image). If you're in Europe, PAL does better at 720 by 576.

Digital "resolution" is determined by the "source data" (Mama's picture) arranged as a "logical matrix" (i.e., a two dimensional data array). The true resolution of the image is determined by the rows and columns of information. You could think of a digital image as a spreadsheet -- with each cell of the spreadsheet containing a chunk of visual information to be displayed -- not the physical matrix of pixels in screen. To put it simply, a matrix of visual data  (one frame of Mama's image reduced to ones and zeros) is re-computed, mathematically, to fill a matrix of pixels. If there are more pixels in the display device, than there is digital data, then the display device averages or even doubles pixel activity to blend the actual visual data over the given physical space that must be filled.  The fancy word for all this visual "slight of hand" is called "decompression".

However, most people think of a "line of pixels" and a "line of resolution" as the same. 

Film, even regular 8mm film, is fairly high resolution. It is difficult to get a good comparison between film and digital image resolution since the silver halide crystal clusters (grain) not only vary in position but also vary in size -- depending on the "speed" of the original film. For the sake of simplicity we'll make some assumptions for our hypothetical piece of film. It will be a slow outdoor film with 90 lp/mm (line-pairs per millimeter) and it will be Super 8mm film with a frame size of 5.6mm by 4.1mm - yielding 756 by 553 "resolution" after we average the line pairs for variance in the grain.

So regular television signals are lower resolution than Super 8mm film - this is important! Keep this in mind when considering your film transfer options.

Compression

Digital video -- of all flavors -- uses compression of some sort. It's necessary to get the information to fit on available media, as a truly raw high-resolution digital stream can take as much as 94 gigabytes (one billion bytes) of space for every hour of video - and that's without sound.

Two basic classes of compression are available: Lossless and Lossy.

Lossless makes for magnificent video - but to use it you need 1) vast amounts of very fast storage and 2) you can't compress it to much more than 85% of the raw stream size. Lossless formats are used by professional editors and broadcast studios. Equipment that uses this type of format is very expense and is generally not economical for private use.

Lossy compression is much more common and easier to make use of. Some methods can manage as small as 15% of original stream size - but you wouldn't want to watch it. That level of compression is typically reserved for squeezing a little video over old analog modems. For most applications compression is kept in the vicinity of 40% to 60% of original size. The good methods let you use that level of space-saving while leaving you with a visually clear picture - though often the tricks used to achieve this come with costs down the line.

MPEG2 is a VERY lossy form of compression. MPEG2 compresses the images in a video stream two ways. First, it takes a key-frame and compresses it in a way similar to JPEG - averaging out areas of very similar colors and throwing away a certain amount of contrast information. Second, it does the same to the frames following the key frame, but then throws away any part of this new "delta frame" that isn't significantly different from the preceding image.

Why does this frame "look funny"? A JPEG image of a frame of highly compressed MPEG2 video. We over compressed this frame to show you what compression looks like -- if you zoom into a picture. This compression is the reason you can not blow-up MPEG2 frame, like you can a MiniDV frame. Once data is lost due to compression, it is lost forever.  Still want to use a DVD to archive your home movies?

This allows for a very high rate of compression while still keeping the image quite clear. it has its drawbacks, like any other form of compression. The most visible being the occasional visible line around the delta frame's "patch" of the preceding frame. Someone skilled in using this compression technique knows how to balance between visual quality and space saving, avoiding the risk of these ugly and distracting digital artifacts. The great filmmakers of the world employ skilled people for that very reason, which is why you almost never hear about or see those artifacts. Almost never.

DV Compression uses only a "lite" version of key-frame compression. DV does not use delta frames. DV (Same as MiniDV) has a bit stream of 25 Mbps (Megabits per second) or about 26 gigabytes per 2 hours of video with sound.

In comparison, MPEG2 on DVD averages about 5 Mbps or about  4.7 gigabytes for two hours of video -- when you're talking about DVD-R/+R.

The result, with DV or MiniDV, is a stream of greater clarity with no visible artifacts (assuming nothing interferes with the encoding).

The DV compression method is the preferred method for use in editing digital video for experienced amateurs, "prosumers" and professionals working with smaller formats because the full-frame technique means you can cut, splice and re-arrange frames without risking re-building a key-frame from a series of delta frames. Every time this is done the artifacts from the delta frame patch are preserved and amplified, degrading the image quality.

There are a broad range of other lossy digital video compression methods out there - DivX, XviD, MPG4, Sorenson3, and WMV9, are some of the more popular among many others - but can be difficult to find or use if you're not an enthusiast or professional. They all have their drawbacks and strengths as well, and the variation of quality between these techniques is sometimes astonishing. This is important to keep in mind when choosing a transfer technique for your film.