it depends when you mean. Early recordings could barely pick up low frequencies, if you listen to old jazz they don’t have drums and a banjo player generally keeps time because the drums couldn’t be recorded properly, even earlier recordings can be hand cranked and have terrible trouble with consistent speed which messes with the frequency.
It’s actually a combination of the two.
Recording fidelity was much lower 100 years ago so voices do tend to sound different, but at the same time people did talk differently. It happens to language over time as speech, the vocal pitch and enunciation, evolve over time.
You can hear that today as younger women these days tend to employ much more vocal fry in their speech than women did 20 years ago.
It is also worth noting that due to the lower “capture” quality of microphones and speakers, there was the potential for a lot of muffling and loss or mishmash of sounds that could render a voice unintelligible.
Broadcasters and people doing recording would often use accents that adjusted sounds and timing to accommodate those shortcomings of the equipment. This was not entirely universal, but it was common.
And, like film, many older recordings were recorded on analog devices that could sometimes vary a bit in speed even at the time, and when we “lift” the information into modern systems we have to make a bit of adjustment if the modern equipment moves at a slightly different speed; or if older equipment is available but your copy of a device has some slight gearing difference from the one that made the recording due to wear and tear, repairs, etc. in the intervening years.
Latest Answers