Meeting Topic: Compression: Make It Musical
Moderator Name: Peter Doell
Speaker Name: EveAnna Manley, Manley Labs, Paul Wolff, Fix Audio Designs, Ross Hogarth, Hoax Productions
Meeting Location: SAE Institute, Hollywood, California
The Audio Engineering Society Los Angeles Section held its February 27, 2018 meeting at the SAE (School for Audio Engineering) Institute in Hollywood, California. The location is on the old Kodak campus, and the meeting room is behind SIM Digital, within the same campus. About seventy-five people were in attendance to hear from Paul Wolff, EveAnna Manley, and Ross Hogarth, on the topic of Audio Compression as a Musical and Production Tool. EveAnna Manley joined us from her company Manley Labs, Paul Wolff from Fix Audio Designs, and Ross Hogarth from Hoax Productions, where he had just come off a 2018 Grammy win for blues recording TajMo. Pete Doell, AES-LA Executive Board member and an audio mastering engineer who works at Aftermasters, introduced the panelists and moderated the discussion.
Both Paul Wolff and EveAnna Manley are audio hardware designers, and Russ is an audio tracking engineer and mixer, noted for his work with Roger Waters, the Black Crowes, John Mellencamp, and Jewel. The panel began with a discussion of audio compression, that is, the task of taking a signal with widely varying amplitude, and constraining it to a more even range of amplitude, and the types of hardware audio compressors that have been developed over the years: Electronic tube-based compressors, light-coupled compressors, FET (field effect transistor) compressors, and VCA (voltage controlled amplifier) compressors. Each handles the task of audio compression in a different way, depending on the technology. During the presentation, Paul Wolff sketched several diagrams on the whiteboard to illustrate how audio compressors react to signals, and how they use various types of feedback and feed forward to control their outputs.
The sonic effects of compression are largely determined by the following key factors: response to attack of a transient signal, time to release the effects of that attack, and the ratio of compression desired and achieved, based on a set threshold. Tube-based compressors tend to react more slowly to a transient signal, as well as release their limiting effect more slowly. Examples of these are the Manley Labs Variable Mu and the Gates Sta-Level. Light and sensor based compressors can vary on this score, as a compressor with an incandescent bulb will take time to react as the bulb brightens and dims, a process that takes time, but if instead an LED is used as the light source, that latency is much reduced, and thus the response time depends much more on the response properties of the light sensor. The Teletronix LA2A and the Manley L-Op are examples this type. FET-based compressors such as the Universal Audio 1176, on the other hand, can react almost instantaneously, and VCA compressors such as the DBX 160 can be quite responsive as well.
All the panelists were very focused on the use of compression as a musical tool, as a technique to change the quality and tone of a sound. While compression for, say, a television broadcast is intended constrain the levels audios within strict bounds while in a sense, staying invisible and inaudible, compression in a studio, in a recording session or a mixing session, can be used to modify the sounds in significant, and hopefully pleasing, ways. As Ross Hogarth said, "Most of the time I'm using compression or limiting as a musical tool, as something for tone and something to create a feel, and help the groove." He continued, "Ultimately mix is emotion, and I'm trying to get emotion to come out of two speakers so that when you close your eyes you feel whatever is the intent to feel." They discussed the use of compression on a recording track to make a singer's performance more pleasing, to punch up drums to emphasize more punch, and certain kinds of compression over an entire mix to "glue" the track together.
Paul Wolff described how important the rates of attack and decay are: "If you have a snare peak that goes through and you have the attacks set real low, it may allow that peak to come through, and then it compresses the signal, so the pop comes through, but then the body goes down. If you set the attack real fast you can remove that peak and you can just make it flat." Thus different compressors can create radically different tonal effects. He also noted how it sometimes makes sense to use high- and low-pass filters to process and compress certain signals independently, pointing out that the amount of energy in the bass is generally much higher than in the treble section, and while our ears react as if these high and low frequency signals are roughly the same in energy, the electronics react to whatever contains the most ab-solute energy, namely the bass, and thus can effectively ignore the treble signal, resulting in a distorted signal. Sending the high and low signals through separates paths and compressors and recombining them later will often give far better, more natural sounding result than a single path.
EveAnna described a common use for a FET compressor: "A FET limiter can have under a millisecond attack response time, and so a FET limiter's first job might be to prevent overs for an A to D. For digital you know you just can't exceed zero or you've got a big horrible sound. But there's an inherent attack and release to each kind of compression device. Tubes are probably the slowest; they're never going to be as fast responding to an attack as a FET limiter."
A lively Q&A session followed the panelists' initial discussion, and the audience asked many questions regarding specific applications of compression to devices such as the classic Roland TR-808 drum synthesizer.
The Audio Engineering Society wishes to thank the SAE Institute for the use of their hall, and panelists EveAnna Manley, Paul Wolff, and Ross Hogarth for their time and deep insights.
Written By: John Svetlik