What's the X Curve?

Professor Rumsey looks into the role the X Curve played in improving cinema sound
Author:
Publish date:
31634.jpg

Back in the early days of talking pictures, cinema sound systems were frankly pretty dire and no one could be sure what a movie would sound like in different places. One of the first standardised attempts to deal with this was the Academy Curve, designed as a replay chain equalisation for cinema sound mixing and reproduction environments, which to some extent reflected the frequency response of the worst systems. According to Dolby’s Ioan Allen it was developed by a panel of listeners judging the optimum high frequency attenuation for a wide range of material on typical theatre loudspeakers. If better theatres and mixing rooms emulated this reproduction characteristic (measured at the output of the power amplifier using a standard frequency response sweep reproduced from optical film), some attempt could be made at consistent sound quality. There were other reasons, too, for the Academy Curve, not least that early optical sound tracks without noise reduction were very hissy and needed some de-emphasis to minimise the effect. Perforated screens and large rooms didn’t help either. The Academy Curve was only flat from 100 Hz to 1.6 kHz. Fast-forward to the 1970s when Dolby measured numerous cinema “B chains” (the part of the reproduction signal chain from the main volume control through power amps to loudspeakers and into the room). They found a typical response that rolled off above 2 kHz and below 100 Hz. They also tried to find out what EQ had to be applied to typical cinema loudspeakers in the far field for the best timbre match with unequalised loudspeakers close to the listening position. (Typically, cinema loudspeakers are quite far from listeners, resulting in the high probability that listeners will be in the reverberant field.) Most interestingly, and possibly counter-intuitively, the distant loudspeakers needed a high frequency (HF) roll-off to make them sound the same. These two things strongly influenced the development of the “X Curve”, a target response for the B-chain that largely replaced the Academy Curve and was ratified in 1977. It is an ISO standard (ISO 2969) as well as being used in SMPTE 202M. It rolls off above 2 kHz, normally at 3 dB per octave, and also slightly below 100 Hz. It would normally be measured with a one-third octave analyser and a pink noise signal. You equalise the monitor system so that the measured curve at one or more listening positions (typically two-thirds of the way back in a cinema) matches the target curve. Although “X” might initially have stood for “eXperimental”, latterly it seems to have become “eXtended range”. Dig a little deeper into the reasons why we might need the X Curve and you begin to find a whole pile of different possibilities, some more plausible than others. For my money, the following are the most likely candidates. Firstly: sound reproduced in large rooms tends to sound brighter than the same sound reproduced in small rooms. Henryk Staffeldt suggested that this is because the ear is more sensitive to high frequencies in the diffuse field. Others proposed a psychoacoustic expectation that distant sounds or images will sound duller because that is what happens in natural acoustics (partly owing to air attenuation). If large room systems are not rolled off at HF then they will appear unduly bright. Secondly: Ioan Allen and others have emphasised the important contribution of reverberant sound. Allen suggested that the increased low frequency (LF) compared to HF reverb typically encountered in large rooms leads to a flat test noise having a raised LF and rolled-off HF response when integrated over a longish time period. However shorter sounds will not be so affected by that reverb, and short sounds constitute perhaps 80% of movie sound material, such as dialogue. So equalising the large room system to have a rolled off HF will probably sound reasonably flat for shorter sounds. This still doesn’t seem to account for the amount of HF EQ inherent in the curve, however, which is probably more a feature of the monitor systems in place at the time of the original project than the foregoing factors. The natural corollary of all this is that small rooms need less equalisation, which is indeed what was found. Later versions of the X Curve had a dependency on room volume to account for this. There is a small room version that only rolls off at 1.5 dB per octave after 2 kHz, or a modified version that rolls off 3 dB per octave after 4 kHz. However none of these “small room” curves are really small enough to account for the typical home theatre, which is a much more recent phenomenon. Indeed some have suggested that home theatres really don’t need any “house curve” monitor equalisation, because it is a myth that the X Curve leads to excessive HF in movie mixes. Given the arguments about the effect of reverberation on typical large room measurements, this is persuasive. However Tom Holman found that home theatres still sounded too bright when compared to cinema reproduction, leading him to develop another form of equalisation for the THX Home Cinema known as Re-EQ. This appears to have unity gain at 5 kHz, rolling off above it at about 3 dB per octave, with a small shelf boost below 5 kHz. The reason for Re-EQ appears to be little to do with an attempt to emulate something like the X Curve and more to do with the fact that home cinemas have more HF reverberation than modern movie theatres or dubbing stages, leading to a hard, bright sound, as well as the loudspeakers in home theatres having different directivity characteristics to the typical constant directivity horns found in modern movie theatres. (Some DVDs have been released for the home market with something like this equalisation already applied, so it shouldn’t be applied again in the home theatre system.) Holman says that the X Curve works in cinemas and dubbing theatres because they use loudspeakers with similar directivity characteristics, which therefore excite the room in a similar way. However it doesn’t necessarily adapt well to other contexts where acoustical conditions vary more widely. It has also been suggested that the curve is outdated, because we now have access to modern measuring techniques that can distinguish between early-arriving sound and late reverb. The better acoustics and vastly improved quality of cinema reproduction systems today have to some extent negated the need for such a strong equalisation as the X Curve, but with careful interpretation and intelligent application this somewhat dated standard can still form the basis of an approach to monitor system EQ in movie sound systems.

Related