Why do images captured on celluloid have a sort of "glossy look" when well-mastered dv-footage doesn't? - Page 3

5) DV tends to make faces look a bit "plastic". There are probably several reasons for this:

5.1) DV's poor handling of highlights means that any highlights on the face look like homogeneous blobs of colour without any detail, a bit like a plastic doll. (again, make sure faces don't clip unless you really want them to)

5.2) DV has an almost linear gamma curve where as film has a more S-shaped gamma curve. (Shout if you don't know what a gamma curve is... Or check wikipedia).

Linear gamma curve

Simple S-shaped gamma curve

This difference in gamma curve means that film tends to render objects exposed at the middle of the curve (i.e. the midtones) with quite a high contrast (i.e. the middle of the curve has the steepest gradient) whilst still holding detail in the shadows and highlights. Faces in "glossy" shows are often exposed such that the majority of the face lies somewhere in the middle of the curve (although it's also quite common to use a "kicker" to put a bright highlight on the edge of the face). Film will render these faces with a high contrast, making them look visually interesting and accentuating facial features like cheek bones.

5.3) DV's colourimetry is inferior compared with film. In other words, DV gets colours a little wrong. As viewers, we're very sensitive to "slightly off" skin tones and we'll always read "slightly off" skin tones as unattractive.

6) Film tends to allow a shallower depth of field. There are several reasons for this:

6.1) Film lenses are often faster than video lenses (i.e. film lenses often have a wider aperture)

6.2) The size of the imaging target is much smaller on a DV camera compared with a 35mm frame. This means that, for a given focal length and T-stop, you'll get a shallower depth of field on a 35mm camera.

You can still achieve a shallow DOF on a DV camera by shooting wide open on a long lens so it would be inaccurate to say "DV has a shallow depth of field".

7) DV cameras usually shoot interlaced by default; film is progressive. A static shot looks identical no matter if it's shot progressive or interlaced. But when things start to move, interlaced imaging does two ugly things: 1) it produces "combing" at the edges of the moving object and 2) moving objects get rendered at 1/2 the resolution that a still image would be rendered at (kind of). And, of course, audiences associate interlaced video images with video and not with film.

8) Cultural aspects. Even today, many "glossy" fashion and film shoots are shot on film and so audiences tend to associate that "filmy" look with "glossy", even if they can't articulate why.

OK, I've waffled for long enough. Some closing remarks:

Don't bother trying to make DV "look like film". Instead, try to make DV look like "glossy DV". Light very carefully because you'll never get those burnt highlights back. Turn off everything "clever" in the camera's imaging processing pathway (sharpening etc). Try using a mild softening filter in front of the lens (experiment and be careful... It's quite easy to get a cheesy 90's porn look with softening filters). Shoot on a long lens, wide open. In colour correction, use your vectorscope, a good monitor and lots of patience to get the skin tones as natural as possible. When shooting outside, use a polariser filter too make the skies look deep (but don't do huge pans with a polariser or you'll find the sky goes weird as you pan). In colour correction, try adding an S-shaped gamma curve. Use your waveform monitor to make sure you're only crushing and clipping what you want to crush and clip. If you're not sure how best to use your colour correction tools and you haven't done so already then buy yourself a copy of "Color Correction for Digital Video" by Hullfish & Fowler.

Good luck.

Discuss this article in the forum

(delicious | digg | reddit | technorati)