I’m getting old. My hearing vanishes above 5.5kHz. (I was the young idiot hippy who could always be found sitting in front of the speaker stacks at the concert…)
I’m forever accumulatng WAV files with very noisy speech on them. Trying to figure out what’s being said is getting to be frustrating. And no, no hearing aids yet.
It occured to me that Audacity might be the tool I need to try to reproduce the tricks used by hearing aids. Now, bumping frequencies’ power levels to help intelligibility is simple,
but there’s another trick that I’m told hearing aids can do to help us deaf ol’ phartes - that’d be to grab a swathe of frequencies in the problem ranges and slide them
down into the hearable range on top of the audio that’s already there. Ex: Shift 6kHz-8kHz down to 3-5kHz.
Can anyone add any observations on this? I kind of like the idea of a ‘Deaf ol’ Bugger’ preset in Audacity.
Cutting very low frequencies (below about 150 Hz) can often help improve the intelligibility of speech.
Try Equalization settings similar to this, then use the Normalize or Amplify effect to bring the peak level to -1 dB.
For best results, try experimenting with the Equalization settings to see what works best for you.
I don’t know that frequency shifting is going to help a lot. Human voices generally fall into the telephone effect – 300Hz to 3000Hz. Your hearing is still good to AM radio – 20Hz to 5000Hz.
You may run into the same problems that people trying to get the “Announcing Effect.” You can’t just shift stuff around. Crisp “SS” sounds that for you are history are converted into muffled rumble when you push them down.
You might try the telephone filter.
Effect > Equalization > Select Curve > Telephone. You might very well find that stripping off lows and highs works better for you. You can also tailor the curve after you load it by pushing all the little points around so the points on the right are further to the right (higher pitch) and try that again.
The telephone “effect” was burned into telephones for a reason. It gave you very good intelligibility over noisy lines.
Let us know how it comes out or you need help with the tool(s).
There’s another idea. I have lumps and bumps in my hearing I didn’t used to have. Get them to give you your hearing test frequency diagram. Select CostCo stores will do that free. Try to transfer those readings to Audacity. Where you have peaks, dip the sound a little, etc. etc. I have to wear hearing protection now when I use the high pressure air hose to clean a computer. The hissing sounds like ice picks in my head.
There is a Change Pitch effect. Of course, that will shift-down the low frequencies too.
You could try splitting the audio into two bands with high-pass & low-pass filters. Then pitch-sift down the high frequencies and re-mix the hign & low band files.
Trying to figure out what’s being said is getting to be frustrating. And no, no hearing aids yet.
If you are getting frustrated… it’s probably about time… Or if you haven’t done so already, at least see an audiologist. He/she can tell you if you are “borderline” of that you really-really need hearing aids.
You could try splitting the audio into two bands with high-pass & low-pass filters. Then pitch-sift down the high frequencies and re-mix the hign & low band files.
Yes, except I still don’t think that’s going to work. Duplicate the voice track so you get two of them, one above the other. Edit > Duplicate. Then you can turn each copy on and off with the SOLO and MUTE buttons on the left. Effect > High Pass Filter to the top one and Effect > Low Pass Filter to the bottom one. Split them around 5000Hz.
You can then apply the pitch shifter to the top one to bring all the sounds down, and, you can hear what each one sounds like by itself and in combination, again with the MUTE and SOLO button of each track. With no buttons applied, you will hear a simple mix of both tracks.
That approach does work, but it’s a lot more complicated than using the Equalizer, which is why I suggested trying the Equalization effect first.
Regardless of which approach is taken, reducing the level of frequencies below “voice range” can help intelligibility as is removes “non-voice” sounds.
For frequency band splitting and pitch shifting it would be better to use the “Sliding Time Scale / Pitch Shift” effect as it is higher quality than the simple Pitch Shift effect. Also it would probably be a good idea to use a high frequency low pass filter to remove noise above about 7 kHz so that noise is not being shifted into the “audible” range.
Pitch shifting above 5 kHz down by about 10 semitones (about -40 %) may help with distinguishing “S’s” and “F’s”.
