View Single Post
Old 18th February 2010 | Show parent
Registered User
Ben B's Avatar
I think the problem here is one of semantics. Dialnorm is not a measurement. It refers to the metadata flag in an AC3 file, which tells the playback system how much to attenuate the mix. I think it adds to the problem of level standardization when terminology is misused. I contend that the station should not specify a specific dialnorm value, because it adds to the confusion, and could very well be totally wrong for the program in question.

For example, suppose I mixed a soundtrack, and my dialogue was already at an ideal level. In this case, I would want to encode my mix with a dialnorm of -31 (no attenuation). Choosing any other value would result in my mix playing back too low. The station should specify target levels, and a standardized method for assessing them (using actual measurement techniques, not by substituting the word "dialnorm"). For example, they could specify a value for "peaks not to exceed" on any channel (using dBfs), and the desired LEQ (or some other measurement standard) for dialogue.

From my observation, level variation from one station to the next is worse than it's ever been (on cable TV anyway). The different broadcast companies have different delivery specs, which isn't helping. I suspect that too many audio people are not using the dialnorm feature of their encoders in the right way. I also suspect that a lot of folks have no idea how to accurately measure average dialogue levels.

Requesting a dialnorm value of -24 is akin to asking someone to turn their mix down by 7dB regardless of its original level. If they were to replace the word "dialnorm" here with their preferred method for level assessment (such as LEQ), it would make a hell of a lot more sense, and probably solve a lot of problems.

-Ben B