"Hollow" voice over

Hello, I am new here and have a question,

I use for my voice over in my video montages a podcast from Rode, in shockmound and ofcourse audacity for register the voice. When I play that again the sound is “hollow” like I speek in a metal pipe. The room where i work is small.
Is there a filter or other system in audacity to make my voice over bether ?

There are very few tools for repairing a damaged voice. You might try turning off Windows Enhanced Services.



If the hollow effect is due to the reverberation in the room then IMO there’s no filter which will fix it. If effective de-reverb filters do actually exist then they are horrifically expensive.

A possible solution for future recordings : shield the microphone from sound reflected from the walls of the room with acoustic foam …

Have done that. I try again.

Thanks I have buy two sheets of accoustic foam past week, I gonna try again after this installation.

If you conform to the modern idea that all your surfaces need to be polished oak and you subscribe to the clean, spare look, then recording voice is going to be an adventure. We’re currently struggling with a friend on the US east coast who volunteered to do promotional announcements in one of those rooms. It was pretty, but it sounded awful. We’re fitting her with a proper microphone, but it’s still only going to help a bit. The problem is not going to vanish.

This is a different shoot with the same problem.


This clip will always sound echoey. There is no filter for room echoes.

The left-brain analysis tells us that room echoes are your own voice arriving at the microphone later and later than the original speech by bouncing from the walls. So basically, you’re asking the software to filter you out of yourself. Rough to do.


Hello, very interessting what you have to say but wat do you think of my voice over (attachement) how can I make it bether?
catmen like I say its like I speek in a pipe.


Try getting closer to the microphone. That will also help with the very low recording level.
Use a “pop shield” so that you don’t blow directly on the microphone


I send you a foto of my installation, I use a røde podcaster with plopshield and shockmound, i speek only 15 cm from the micro, so what can go wrong?

Step #1. Normalise to 0dB …
normalise to 0dB.gif
(“Normalise” is in the “Effect” menu)

Step #2. apply this (notch-type) equalization …
the equalization applied to ''aquadukt,mp3''.gif
(“equalization” is also in the “Effect” menu)

the result …

OK, this is realy bether!
thanks you verry much I go further this way,
sorry for my horibble englisch.

With the røde podcaster the sound should be better than you are getting … http://www.youtube.com/watch?v=cE-O7mY7nSE#t=2m40s

Possibly you could accidentally be recording the computer’s built-in microphone along with (or instead of) the røde microphone ?.

The computer’s built-in microphone is not high-quality (not broadcast-quality) like the røde podcaster.

Only enable the USB recording device (the røde), disable the other recording devices, (e.g. disable the built-in microphone) …


“Something” is going wrong because a røde podcaster should sound much better than that. It “sounds” like you are about 50 cm from the mic or more.

What settings do you have in the Device Toolbar?

Can you take the equipment outside? Just for a test if you have a quiet neighborhood, test the system exactly as you have it, but without the room. Most of the presentation sounds OK, but you have that constricting echo sound and I think that’s what’s killing you.

I’ve been known to build a studio out of furniture moving blankets to get away from this effect.


You can build a tiny cave with acoustic foam, but the microphone is still going to record the wall behind you.

I know people who record in a clothes closet lined with quilts to get a “studio.” Echo is a pretty serious problem and as above, there’s no software solution.


There is a tool in audacity that allows one to see the reverberation. In the Analyze menu pick plot spectrum. Use the autocorrelation function. The peaks occur at the reverberation frequencies of which there are several sets probably corresponding to the walls and ceiling in the room as reflectors. Trebor’s procedure can be seen to greatly reduce the reverb as the ear also agrees with.

The second recording does not show as great a reduction in reverb as the original poster’s recording does with Trebor’s settings. The room for this recording may not have parallel walls or for some other reason show such well defined peaks as the original poster’s room. It may also be that much of the spectral content of the woman speaker’s voice is so much higher than the reverberation frequency that the reverberation is less pronounced in the autocorrelation.