For a project I’m working on, I’d like to make some of the voice audio I’ve recorded (.wav) to sound like it’s coming through a telelphone to simulate a radio call-in show. I know it has to do with some kind of filter and perhaps down sampling (8khz?) but can someone explain exactly how to do it spefically – which program, which functions?
IANAn audio guy, but I took a spoken word wav file and subjected it to a high pass filter found in Audacity and it kinda sounds a like it’s coming over the phone.
Telephones have a very narrow bandwidth. Find some sort of graphic equalizer (Windows Media Player has one) and do your best to kill off everything below 300 Hz and above 4 KHz (4,000 Hz). This should give the voice that distinctive “tinny” telephone sound.
Using a very low sampling rate is probably the easiest (if not the best) way. You can do this with the tiny “Sound Recorder” program that comes with Windows. Just open the file, then Save As… and somewhere in there is a button/tab that says “quality” or “rate” or something. Bring it way down and see how it sounds. Be sure to keep the original saved so you can undo it.
Silly idea, but why not reciord through a phone? You can get a device at radio shack that plugs into the handset jack of a standard phone that provides a microphone output.
No phones use carbon microphones anymore. I haven’t even seen one in well over 15 years. They almost all use either voice coil or piezoelectric mics these days.
If the Adobe Audition “Telephone handset” FFT preset is any indication, you want it to cut off everything below 300Hz, then exponentially ramp up so that at 1600Hz, it’s full strength until you get to 3300Hz, where it drops off linearly to 0% strength at 4000Hz. Here’s a kind of ugly ASCII art rendition of the filter:
100-| BxxxxxxxC
| x x
| x x
| x x
| x x
| x x
| xx x
| xxx x
0---| xxxA Dxxxxxxx
---------------------------------
| | | |
300Hz 1600Hz 3300 4000Hz
I bet engineer_comp_geek’s suggestion of just chopping it off at 300 and 4000 would be a pretty reasonable approximation for most purposes.
They may not be making new telephones with carbon microphones, but there are many old telephones and handsets that are still in service. Telephones used to be designed for a service life of decades.
I did something like this using Audacity; I actually had to blend some modern recordings in with samples taken from a 1950s cereal commercial (in order to make a radio jingle); I can’t remember the precise order of what I did, but it involved high and low pass filters, downsampling and compressing the dynamic range - all of which have been separately suggested above, but I actually needed all of them in concert.
The end result was very convincing - if I still have a copy, I’ll post a link.
Another chiming in for Audacity - I used it to simulate both bullhorns and telephone conversation for a sound design project I was working on.
I used the ‘lo-fidelity’ effect, then hi and low pass filters and the result was pretty darn close for a telephone effect. The bullhorn required a bit more finagling, but also worked out in the end.