Making it sound old

If you’ve followed my blog before I began the YLTLSBC post-a-day thing, you may have noticed I haven’t been sharing as many Onion videos lately. Rest assured I’m still designing sound for ONN, but for the time being I’m only posting the videos that feature interesting sound design or prompt me to say something interesting about the sound design process (or anything else for that matter).

O-SPAN Classic: CIA Accidentally Overthrows Costa Rica

Congress, 1924: Rep. Demands Horses Wear Dresses To Hide Foul Penises

I’m surprised at how inauthentic the “vintage audio” simulations I hear on television or the radio often seem. Sometimes it sounds like the producer simply applied some EQ to cut the bass/treble frequencies, boosted the mids, and called it a day. While the result sounds more boxy and nasal than a clean recording, it usually lacks the character that gives old recordings their unique sound.

For these Onion videos, I was initially asked to transfer the digital recordings to an analog tape and try degrading the sound by making a copy of a copy of copy… But I didn’t do that. First, I listened to a bunch of examples of historically similar recordings to use as a references. For “CIA Apology,” I primarily used a recording of this famous JFK speech.

And I used William Jennings Bryan’s 1923 “Cross of Gold” speech as the model for “Horse Dresses.” (“Horse Dresses” is set in 1924, so I should note here that film with synchronized sound is more or less an anachronism.)

Then, I thought about how I could digitally model each step of the analog recording process to in order to mimic the various changes and distortions that might have occurred in an “actual” recording from the 1960s:

The press secretary would be speaking into a dynamic mic, so I applied an EQ plugin set to roll off the bass and treble, with some very subtle boosts to the upper midrange, mimicking the mic’s frequency response. Then that mic’d signal would go to a P.A. system, so I used a guitar amp simulation plugin to distort and further modify the EQ. The P.A. would project into an auditorium, so I added some room reverb. Then I figured the sound would be recorded by a reporter with another dynamic mic (so I repeated the first step). They would record to tape, which can change the sound in all sorts of subtle ways; tape hisses, varies in pitch/speed, compresses and distorts. (I used a few different plugins to model these, but here’s a good free plugin that gets you part of the way there.) Then the tape would then be broadcast, so I added some more heavy compression and EQ to simulate an AM radio.

At this point, my simulation sounded pretty good, but as a final step I used a special plug-in in Logic called “Match EQ” to help me check my work. Match EQ analyzes the frequency characteristics of two audio files and and creates an EQ profile designed to make one file sound more like the other. Again, I used my references as models.

In “Horse Dresses,” I used the same kinds of effects but to a much greater extreme along with a healthy dose of record crackle.