The Sound of Her Voice

Welcome new subscribers! It’s such a thrill to see your interest in this project, which asks: what if the history of audio could teach us something about how we perceive the world?

Take for instance, how you perceive your own voice when you hear a recording of it. Do you think, “That’s not me! I don’t sound like that…do I?” 🙉

Yes, you do sound like that — but also not as bad as you think. There are anatomical and neurological reasons why our recorded voices may feel weird to us. Those reasons have to do with how sound waves are funneled into our inner ears and turned into electrical impulses that are then interpreted by the brain. The perception of our voice involves the auditory mind picking up not only the sound waves we produce by vocalizing, but also the vibrations of our larynx through the bones in our heads. Ask any neurologist — our auditory systems are hella complicated when it comes to speech. 

Anyway, when we hear a recording of our voice, those internal bone-vibrations are absent, and so a crucial input we’re accustomed to is missing. Thus it sounds strange, and — because it’s supposed to be our voice — just wrong.

This is all related, in a roundabout way, to this piece I just published on the public media site Transom. It’s about radio star Mary Margaret McBride, who hosted interview and chat shows on various networks and New York-area stations from the 1930s to 50s. (The header photo for this post shows McBride on the right, interviewing Missouri historian Goldena Howard).

McBride was one of the first creators of the “confessional interview” format on US air — a sort of Oprah or Terry Gross of her time, but even more epic if you consider that radio in McBride’s day was live and did not air repeats. She and her team worked hard for two decades to research and book thousands of interviews with authors, actors, public figures and ordinary folk. McBride’s audience was huge. She is largely forgotten today, though scholars are rediscovering her.

I went on a few side quests while researching the fascinating story of McBride’s success. So I will feature some of those here, because what else is a newsletter for, if not side quests? 

Today’s quest originated with a scene in a memoir McBride wrote. She describes how how, soon after starting in radio in the mid-1930s, she heard an air-check disk of one of her programs.

McBride was relating something more than just standard-issue vocal self-consciousness. She goes on: “I sounded like all those other women, and from that day to this, hearing my own voice has been torture.” By “other women,” she meant women who spoke on the radio. McBride did not want to think her voice sounded like those other women on air. She did not have to ask herself why.

By the 1930s, a popular perception had taken hold that women sounded really annoying on the radio. Despite the fact that (white) women were the norm as voice operators at the telephone company, radio had quickly become an exclusive man-on-mic club. The few women such as McBride who hosted their own shows were relegated to mid-day “zero hour” time slots, when — the thinking went — only housewives were listening and might put up with the torture of hearing their own sex speak words on the radio.

“From the beginning of radio women had vied with men before the microphone both as artists and as announcers and commentators. The feeling was now quite general that however acceptable the female voice might be in song or dramatic parts, yet in ordinary talk the microphone-voice of women was much inferior to that of men,” the business historian Gleason Archer wrote in 1938. 

Archer then raises and, to his mind, resolves a vocal mystery in one fell swoop:

This voiceband pinches

Apparently both lower-pitched and higher-pitched voices suffered when heard on radio receivers of the time. But the end result was a sonic sexism that only got more and more entrenched.

Here I have to turn this into a Tina Tallon stan post. Tallon is a composer and Ohio State associate professor who wrote a piece for The New Yorker called A Century of “Shrill”: How Bias in Technology Has Hurt Women’s Voices. It points to early regulations in the US that limited the bandwidth of broadcast radio signals in an effort to reduce interference. The so-called “voiceband” got set in the 1920s to between 300 and 3400 hertz. 

“Unfortunately, the researchers and regulators who were deciding on this range primarily took lower voices into account when doing so,” Tallon writes. 

Why did this bandwidth range kneecap higher-pitched voices? It’s science! Tallon explains how, when we speak, the sound of consonants “result from short, high-frequency noise bursts that punctuate the more continuous, lower-frequency pitched components that we perceive as vowels.” With the majority women’s voices, consonants were vocalized above the voiceband’s range, so very hard to perceive. 

I don’t know about you, but I feel that consonants are important if you enjoy comprehending human speech. Tallon agrees:

This stuff may not make immediate sense if you’re not a sound engineer, and I certainly don’t trust myself to explain all the nuances. But I grasped the issue right away as I watched Tallon’s sonic demonstration in a presentation for the Radcliffe Institute. 

What is going on in the screenshot above: Tallon processed two identical voice samples — a higher-pitched “female” voice and a lower-pitched “male” one — according to the specs of 1920s AM radio, as well as the typical microphone equipment of that time. When you hear the two audio samples side by side, it’s clear how hard it would have been to comprehend higher-pitched voices on the air.

(Seriously check it out! I’ve cued it up to the correct time-code, 31:40 here, and it’s quick. But Tallon’s whole presentation is really worth watching if you have the time).

No Girl Stuff

OK, so women’s voices sounded bad on the radio 100 years ago, so what? First of all, this established a listening perception that remains with us today. Just click on “reviews” for any female-hosted podcast and you will likely find a hearty sample of one-star reviews saying “annoying voice,” “high-pitched, shrill,” and blah blah blah. Male hosts and reporters get comments on their voices too, but not to the same extent.

Better audio technologies such as FM broadcasting did not seem to fix the issue. When I interviewed NPR’s first program director Bill Siemering for this 99% Invisible feature, he recalled how station managers would complain about all the female reporters he’d hired: “That on FM you’ve got all these high frequencies — and women’s voices, you should see the [VU] meter, Bill, when women’s voices come on.’ That kind of thing.”

In the Radcliffe presentation, Tina Tallon demonstrates how modern sound-compression algorithms such as the .mp3 may also act to reduce our comprehension of higher-pitched speaking voices. Just as biased medical research resulted a diagnostic and treatment system based on male bodies, the almost exclusively male sound- engineering world seems to have established a vocal baseline that still haunts us today. 

Tallon makes the important point that radio’s early engineers knew exactly why women’s voices sounded weird on air. But did anyone think to ask, is it bad we might be creating a huge problem for half the human race? Of course not, because the research labs that created microphones and broadcasting technology had almost no women present.

Instead, the broadcasting world basically decided the ladies would just have to avoid the mic as much as possible, which conveniently locked one gender out of a major source of cultural and industry power. Many of the “how to break into radio” manuals I’ve read from the 1930s and 40s were especially fond of expanding on this original sin.

For women — or anyone with a higher-pitched voice in radio and podcasting — the persistence of sonic discrimination is incredibly frustrating because it clothes itself in technological self-absolution: “I’m not a bigot — I’m talking science here. And science says no girl stuff!” 

We are familiar with this kind of move from such popular stupidities as eugenics, phrenology, and The Bell Curve. But with speech, people often don’t even reach for science, because everyone thinks they’re an expert in how voices should work. If someone sounds “shrill” or “annoying,” it must be the speaker’s fault. They’re annoying you on purpose just by talking. The delivery method for that speech is not considered because it’s invisible. You can hear how some of this voice shaming plays out in this excellent NPR animated feature Speaking While Female.

Revenge of Goo

Mary Margaret McBride and her production team had one answer for all this nonsense: making bank. Despite being called flustery, a babbler — and in one TIME magazine review, simply Goo — McBride developed a huge audience who adored her show. Her haters truly became her waiters when McBride fans would pack Madison Square Garden and even the Yankee Stadium if given the chance.

At McBride’s broadcasting peak, sponsors waited in line to get a coveted spot in her advertorial rotation. When she decided she liked a product enough to endorse it, she went all in. Yes, critics were baffled and annoyed by her gooey testimonials to cake mixes and canned soup, but listeners eagerly wrote in with recipes and testimonials of their own. And so McBride had the financial independence to do her own thing, to get longer programming slots, and to use her journalistic training and instincts to interview whom she wanted. If she hadn’t died in 1976, she would definitely be trying to book Taylor today. (And Tina Tallon — though Brooke Gladstone got there first.)

Mary Margaret was a large woman who was not afraid to dig into her sponsors’ cakes and cookies on air with her guests in pursuit of the day’s endorsement. One must presume the stingy broadcast voiceband worked in her favor then, as it left the higher frequencies of their lip-smacking feasts unheard.