How to record good quality narration .WAV files?

How can I record good quality sound files of my voice to embed in PowerPoint presentations? I know how to embed the file, and can follow simple instructions to use various recording functions on a PC. But, how do I make them sound good and clear and not irritating or difficult to listen to?

I don’t need to use music or anything else. I just want to narrate slides, like I do when presenting live. And I’m going to do each slide separately to get things right. For good or ill, I already have lots of information and experience on making good and enjoyable presentations - people tell me they like them.

What sort of a microphone should I buy? How far away do I hold it? What sort of room should I be in? Can I just use the sound recording function in Windows? What else should I know or try or avoid?

For whatever it matters, I tend to make my “S” sounds whistle a bit, which I can’t hear myself but is very obvious to me in recordings. Anything I can do to make that less annoying rather than more would be a plus.


Audio Engineer/Producer here.

For what you need, the simplest solution comes in a few forms:

  1. a USB microphone.
  2. free recording software like audacity and your laptop mic.
  3. a handheld recorder like the Tascam DR03…$90, small handheld unit, built in stereo mics, can plug in via USB to your computer to xfer files or you can remove the microSD card and use it directly.

I recommend #3 for simplicty and the fact that it has several other potential uses, recording meetings, concerts, what have you. Also, no real learning curve or variables to mess with.

oh, to remove or lessen sibilance there are two common techniques:

you can reduce the EQ at approximately 5-7kHz (sweep around till you find the best range) by 3-6 dB

or use a tool called a de-esser, which is a frequency specific dynamics processor that achieves the same goal, usually a little more elegantly.

to do these, you’d need some audio editing software. You can use Garageband for this if you’re a Mac fellow, if you’re PC I’d recommend Audacity, it’s free and would work just fine.

Thanks, picker! I already own an Olympus Digital Voice Recorder (I think it’s aVN 6200PC, but whichever model it is, it cost around $50 and can put .wav files on a flash card). I think of it as being for leaving myself voice memos while driving. Would it be good for this? I don’t know if it is smart or dumb to try to get appropriate recording quality out of it.

I’ll probably get Audacity.

Any advice on how to hold it, what kind of room to choose, other tips like that?

I have a similar voice recorder, and also a more high-end sound recorder (Zoom H2, about $150). There is a huge difference in sound quality.

use a mounted or a boom microphone to easily keep mouth to microphone distance and orientation the same.

I’m pretty sure PowerPoint supports MP3s. I’d use that instead of waves. WAVs will be quite large. I’d use Audacity to convert it to an MP3. To do so, you’ll need the LAME MP3 encoder*.

You can get a pretty low bitrate if you’re just dealing with speech, but you’ll have to experiment. I’d try the default 128kbps and see if it’s good enough, and then move up and down from there.

*LAME actually stands for “LAME Ain’t an MP3 Encoder,” which was true when it was first created, but is no longer. The silly naming convention is something that happened early on in the open source community, and the acronyms are so silly that we tend to ignore that they ever were acronyms in speech. Hence calling it that LAME MP3 encoder, which is similar to saying ATM machine or PIN number.

Meh, is there really going to be that big a difference in microphones for just spoken word recordings? I’d just use the Olympus OP already has and convert with care to mp3 – BigT already has that covered, and I agree with the rec for Audacity and LAME. Audacity is far better than most low-end pay-for-pray “recording” software engines, and I use it all the time, even though I have some other expensive software – Audacity just works and it doesn’t bug up your system by hogging all the resources or demanding certain drivers for a soundcard you don’t even have plugged in at the moment. You could mix down a multi track album with Audacity, I’m sure, with great results, with a little elbow grease, and get just the stuff some n00b with pirated FruityLoops or something could get.

ETA the few times I’ve recorded spoken voice, I’ve almost always used the worst mic possible, a Shur SM57 – with good treatment in postproduction, it’s fine. It’s not meant to be hifi or poesie sonore, I take it, so, who cares? Get rid of the hiss and sibilance in post, I say!

I’ve gotten really good results just recording using a $40 USB headset directly into Audacity. Audacity has plenty of filters that you can use to tone down the “s” sounds. One thing that you do need to be aware of is that Powerpoint only will allow an audio file of up to a certain size to be imbedded into the presentation; with your plan of doing a seperate file for each slide you shouldn’t run into any issues though.

More good stuff, thanks.

I’m using PowerPoint 2003, which I read does not allow embedding any file formats other than WAV. Since I intend to distribute the file and don’t want headaches with links being messed up (and yes I know about the “Pack and Go” and other bundling abilities), I wish I could stay with WAV.

But a web calculator for uncompressed audio file sizes says 3 hours worth at 44 kHz, 2 channels, 16 bit, will be almost 2 GB. This is certainly unappetizing. Can I record useful WAV files having 1 channel and lower rate and bit settings? How low can I go?

I wish I had better data, but of course you don’t need two channels, especially for a mono microphone source. You can tweak all the other parameters as needed.

If that fails – and I hope someone corrects me, because I’m interested in the correct answer as well – you might try “transcoding” to another lossless format like FLAC and re-“transcoding” back into WAV using the specs you want. It would likely be easier to just insist on the lower fidelity from the git-go, though.

If you really don’t care about the quality, you can transcode to a compressed, lossy format like mp3 and back into WAV and let the encoder do most of the work for you. Something like this would almost certainly be good enough for a PP presentation. You are most certainly not going to have hi-fi or mid-fi monitors providing FOH sound to your audience – I wouldn’t worry about it, and just compress it as much as you can to make it fit.

OK, I got Audacity and think it’s pretty nifty, and I got a cheap mono mic with a 1/8" plug, and I experimented.

When I record using the mic in the laptop’s mic jack or the docking station mic jack, the sound is very faint, even with record and playback turned all the way up. Audacity VU meters and plots of volume are barely registering unless I actually get the mic in contact with my lips. There is a constant whirring and whining sound, too. Sometimes it picks up sound around the laptop better than it does around the mic, as if it is recording using a mic in the laptop. Sometimes it goes dead completely. Sometimes the sound is faint but I can hear my voice, and it doesn’t matter whether the mic switch is on or off. I have spent about an hour hunting around the almost nonexistant docs for the laptop for clues, as well as reading about bugs and other weirdness on the internet, but so far have not even resolved whether the mic jack is mono or stereo (there are clues in each direction and the docs are amazingly vague). Sometimes it does not even matter whether I select Mic Jack, Dock Jack, or Line In, in Audacity. I suspect some weird automatic thing is automatically confusing me.

I have also used this mic in my Olympus VN-5200PC (not 6200) recorder, and the volume is now great, but 30 seconds of recording takes 40 seconds to play back, even in carefully timed readings like reciting the progress of the seconds indicator on a clock.

Both of these are pretty big issues - any ideas?

The last time we did this, I rented a soundbooth. :slight_smile: I found that voice recording is a great way to pick up drones and hums from appliances or ventilation fans, things that the human listener has tuned out as background noise. If the recorder is sitting on a hard surface, that may transmit the sound of its motors (if it’s tape-based). Speakers may vibrate the microphone’s case and affect the recording. And of course there are external noises like passing traffic. You have to be careful.

Does your recorder have variable-rate playback? Mine does; there are separate dials for pitch and speed. It does some kind of digital ‘stretching’ to lengthen the sound, for example, without lowering the pitch.

issue 1 is probably an input gain issue, or mic type. Your mic could be omnidirectional (unlikely) or you might need a preamplifier. First thing I would do is try to adjust INPUT signal coming into the machine. With Audacity, I believe that input gain is controlled first by your computer o/s, and after that by the software.

You want to see the metered level of your input signal (either through record-enabling or input-monitoring the signal on the channel strip) When the mic is in position, speak…meters should spike around -3 to -6 dB. Adding a noise gate will help a lot, adjust threshold to ignore background noises but open upon your speaking. Attack should be quick…say 1-15 ms, and release set to taste…start around 30-40 ms and adjust from there.

you can also try creating an EQ filter to minimize extraneous sounds. use an EQ program to put low- and high- pass filters on the signal. starting points ~ 180 Hz on the low end and 8-10kHz on the high end. Adjust to taste.

  1. is most likely a sample rate mismatch. If you record at 48k and Audacity is set to 44.1k, you’ll see your audio slow down proportionally (and vice versa) So double check your rates and adjust to be the same. For speech, recording at 16 bit/44.1 is more than adequate for non-entertainment (music, audiobook, etc) tasks.