Similarly, from TPM Reader RS:
There isn't one answer to your question. I listened carefully to the audio and they are doing some very sophisticated things, changing the audio processing and imaging with each change of scene. They combine his voice with the background audio of each place, and then process his voice to make it sound different in each place, using a combination of reverberation, high-pass filtering (taking out the low part of the sound), and delay. They also change the stereo position of his voice and these effects to change our perspective on where his voice is coming from. I think the particular sound you were asking about is a combination of high-pass filtering (take out the lows to make it sound tinny) and reverberation (to make it sound more distant). But what is most interesting to me is the way they give each change of scene more power to grab our attention by changing the quality of the sound. Of note also is they didn't alter his voice - change the pitch or timing - which is also possible these days. For instance, you could use auto-tuning software to make it sound like he is singing in tune : )
Another from TPM Reader JH:
I'm a sound designer and I work at a studio in Los Angeles, I have to do this particular effect about twice a day.
If you listen carefully, you notice that it changes from environment to environment, inside the closed factory it sounds like an ambient public-address system, like "Daisy," but on the cut to the factory exterior it sounds like an exterior reverberation, as if he were singing it in the very-empty parking lot. The interior shot of the conference room sounds like it's done over a speakerphone/conference call, etc...
The "Daisy" ad only had echo over the countdown, which was probably not added by Johnsons's ad men, but was already a part of the stock countdown sound effect they used. I have similar recordings in my library and they all have the same echo, because they were recorded outside on a real launch pad, with the countdown being read through a tannoy.
(Nerd digression: Nowadays this effect done with what's called a "convolution," which is a gizmo that allows the linear-time invariant characteristics -- basically the spectrum and reverberation -- of a space or a piece of equipment to be recorded, and then, applied later to any voice or sound effect you want. Within limits, you can take any voice or sound and make it sound like it came through not just "any" radio speaker, but your 1959 Phillips B3x91A, or not just any stadium, but Wembley Stadium from 10 rows back. ...
Can I gush for a moment? Our readers are so awesome. This is just a sampling of at least a dozen highly informed emails from readers that came in within an hour or so of my question, on a Saturday no less. You guys rock.