RAP: What codec are you using with ISDN?
Jay: I'm using the Zephyr now with Layer III, single ISDN pair, basic two conductor copper wire. It's actually a drop that was used for an ordinary telephone, and the phone company switched it over and said they were going to use this pair for the ISDN. Dirt cheap. Costs me in the vicinity of fourteen bucks an hour. I get 20K bandwidth stereo, bidirectional. I did a session Friday where I directed an announcer in Miami and recorded him in Boston for a promo for a radio station in Salt Lake City. He was right there on my monitor speakers. It sounds a tiny little bit fuzzy, but it sounds better than FM radio. It just doesn't sound quite as crisp as when I'm recording the guy here.
RAP: How much compression is being used with the Layer III?
Jay: That's twelve to one compression with the Zephyr MPEG Layer III.
RAP: Has this compression presented itself in any nasty ways down the line?
Jay: No. I haven't got any complaints. I do the one pass and I mix it with other elements, and I do a little bit of equalization and some other tricks to it to make it as clean as I can. I have a generalized voice processor that I wrote for the Eventide to fix up signals as they come in off the Zephyr, but even the raw signal coming out of the back of the Zephyr is music quality. A little more distortion than we're used to these days and a lot better than the best MTR10 or MTR12 ten years ago.
RAP: Is this bit of distortion a result of the compression?
Jay: Yeah, it's from the masking. It's not that they compress the signal down twelve to one like you would with PKZIP. It's that it goes through and decides what it is that most listeners won't hear. For example, if you've got a real loud sound and then another softer sound at almost the same frequency modulating it, very few people will hear that softer sound. So this strips them away. Then it says, okay, I've got much simpler data, and sends it down the line. At the other end, it gets instructions to spread out that simpler data as if it were the entire sound. Similarly, you take a very loud sound and follow it in time by a very soft sound, even if they're not related at all in frequency. Most people won't hear the soft sound. It's also the secret of Stan Kenton's band arrangements. He would modulate by doing a big, loud drum riff so that when he came back, everybody would have forgotten what key it was in.
A couple of companies--if I could put in a plug here, not only for Telos but also for Comrex--are working real hard at educating us poor schnooks in the studios, telling us what's what with these algorithms, making it possible for one box to talk to another.
RAP: The Zephyr seems to be the codec that keeps coming up in conversation. Is this the best one out there?
Jay: If you can afford it, it's going to be the most versatile and will pay off the best for you. If you can't, if you don't have five thousand bucks to sink into it, Comrex has some lower cost boxes that do a very, very good job and that cross connect with the Zephyr, maybe not 20K bandwidth stereo, bidirectional, but 15K over a single ISDN line--just the thing for a voice.
RAP: And when you say bidirectional, you mean you can be talking to the person on one side of the line while they're sending the audio down the other side?
Jay: Yeah. And there's a slight delay in all these systems except for the real expensive 3D2. They store up half a second of sound and looks at that half second and say, "okay, here's what I'm going to do to save some data." So Nick Michaels talks to me from Miami, and I hear him a half a second later. I give him an instruction over the talkback, and he hears it a half second after that. I've got a couple of relays in my talkback system so he doesn't hear himself a whole second later, which would throw anybody off. There are work-arounds.
I do a lot of stuff where I rent studio space at Westwood One in New York. I get a New York announcer in there at Westwood and I'm directing him on one standard leaving the Zephyr, and they're sending back to me on a different standard coming back to my Zephyr. So I'm getting 20K of announcer, but he's only hearing 7.5K of me, which is still fine for directing, but he's hearing me instantly.
RAP: What are your thoughts on digital delivery systems like DGS or DCI? A lot of stations are on these networks.
Jay: Those are both very good systems. I looked into them and decided they wouldn't do for what I'm doing because they charge by the minute transmitted. If you have one spot to get to twenty stations, and you want it to be there in four hours, it's a great economical way to do it. It does use data compression, but if you don't need it instantaneously, it works. The problem that I had was that if I did a twenty-minute session, they charged it like twenty spots, and it isn't real time, even under their best service. So I'd have to direct the guy on a phone, and if I'm going to do that, he might as well FedEx me the tape.
RAP: You have a directory of ISDN equipped studios on your Home Page. How did that come about?
Jay: When I first got the Zephyr a little more than a year ago and looked around, there was one company, DigiFone, that had a list. You pay thirty bucks and you get four or five Xeroxed pages. Most of the sites on it were commercial recording studios using the three hundred dollar an hour system. I said, "I don't want this." So, I posted a note in the Broadcast Forum on CompuServe and in the rec.radio.broadcasting news group and told some friends. I called Telos and said, "I want to gather this information and make it publicly available to anybody who wants it totally free with no liability on my part. If you call somebody on the list, you work out your own deal with them. I want to do this just because nobody else is doing this, and the data should be out there." So Telos cooperated very nicely and said, "Well, we're not going to let you put out names and addresses from our guarantee cards, but we will give you the fax numbers so you can send a solicitation. And if they want to be on your list, they'll give you the rest of the information." Comrex has done the same thing quite recently, and by the time this interview comes out, I'll have a bunch of Comrex sites on there, too, since they're all inter-operable. One of the other manufacturers made nasty legal noises and said I can't even mention their box on my list, and I made nasty legal noises then and said, "Fair use--I'm a journalist and if somebody says they have your box, I'm going to report that fact." So they went off in a huff.
RAP: You are Vice-Chair for the Boston Section AES. How did that come about?
Jay: They needed somebody who knew enough about advertising to be able to put together the newsletter. The local section had also been very technical and very "technoid." There are a lot of very good people--incredibly gifted circuit designers, speaker designers--the people who actually understand what's going on at the electron level with the things we deal with, but the meetings were so full of math that only those guys could understand, and there was very little participation by the local advertising or production community. So they asked me to join the Executive Board then a year later be Vice Chair and see what I could do to goose it a little bit for all those other audio engineers, the guys who show up at the New York AES but never attend the Boston section meetings.
A lot of this, the ISDN list I do, the AES, and even to an extent the writing I do in Digital Video magazine and a couple of other places, is trying to bring back something to the community, because when I was twenty-five, there were a lot of people who were very good to me. I've had my gurus.
