DISCUSSION TOPIC
"CINEMETRICS AND CINEMASCOPE RESEARCH"

BACK TO THE DISCUSSION BOARD
Posted by: Sam Roggen Date: 2015-07-03

My name is Sam Roggen and I am a PhD student at the University of Antwerp, Belgium. Since I am, as you might have noticed, frequently using Cinemetrics as part of my empirical research, I thought I'd inform you on my research project (and thank you for this extremely useful study tool). 

 In my research, I look at the style of the earliest CinemaScope films from the 1953 until the end of the decade. As I learned from David Bordwell, these early Scope years offer a prime case for studying how a phase pf technological change can encourage classical studio directors to move between formal continuity and adaptation. My project takes the shape of four chapters, each of them devoted to a particular stylistic problem, and the various solutions Scope filmmakers came up with. I am, for instance, currently developing the chapter on stylistic strategies filmmakers used in order to highlight important visual information in the wide Scope frame. In this chapter, I use Charles Barr’s idea of gradation of emphasis, in order to examine subtle strategies used for emphasizing crucial details or objects.

In the case study that accompanies this chapter, I look at how Anthony Mann applied gradation of emphasis in his films. I am examining how elements as genre, aspect ratio (and the systems that support it) and creative collaborations influenced not only this particular strategy, but also, related to that, ASL and shot scale in his films from the 1950s.

 The measurements I perform with Cinemetrics are a first empirical step I take, as I aim to combine the data with formal analyses and historical (archival) research, in order to answer two questions: (1) how do these films look, and (2) why do they look the way they do?

 I’ve used Cinemetrics regularly over the last few months, and I am planning to keep doing this. Not in the least because, from October 2015 until May 2016, I will coordinate a students’ research project, in which fifteen of my students will examine film style in American widescreen cinema, thereby frequently using Cinemetrics in order to measure ASL, shot scale (and perhaps camera movement). 

Replied by:Barry Salt Date:2015-07-19

Your project is very promising, but also very ambitious. My database on the Cinemetrics website contains a large amount of data on ASLs for 'Scope and non-'Scope films. My results, extracted from that database, for 222 US 'Scope films and 382 non-'Scope films made from 1953 to 1959 (inclusive) is included below. It expands and validates the comments I made on this point in my "Film Style and Technology.." on page 274.

Replied by:Barry Salt Date:2015-07-22

Having looked at the 'fifties, it is obviously worth looking at the whole matter of comparative cutting rates in American 'Scope films and American non-'Scope films all the way up to 2009, decade by decade. By 'Scope I mean all films with a screen aspect ratio greater than 1:2.2, whether they are called CinemaScope, Panavision, 70 mm., or whatever. These two categories cover the 1065 'Scope films and 3781 non-'Scope films from 1953 up to 2009 (inclusive) in my database. The results are graphed below:

As you can see, a surprising feature of this graph is that although the mean ASL for all the 'Scope films in the decade starts out as greather than that for the non-'Scope films in the 'fifties, sometime in the 'seventies the position is reversed, and from that point onwards, the mean ASL for non'Scope filmsis greater than that for 'Scope films. This is because long-take films almost entirely stopped being made in 'Scope at that point. Of course, long-take films (films with ASLs bigger than 9 seconds) are also almost entirely art films in recent decades. As you can see, a surprising feature of this graph is that although the mean ASL for all the 'Scope films in the decade starts out as greather than that for the non-'Scope films in the 'fifties, sometime in the 'seventies the position is reversed, and from that point onwards, the mean ASL for non'Scope filmsis greater than that for 'Scope films. This is because long-take films almost entirely stopped being made in 'Scope at that point. Of course, long-take films (films with ASLs bigger than 9 seconds) are also almost entirely art films in recent decades.

If we use median values rather than mean values for all the films in the various decades, the difference between the two classes almost vanishes after the 'seventies. This is a message about basic statistics, and a warning of what you are likely to miss if you are an obsessive believer in non-parametric statistics, and refuce to believe that mean values exist.