Audio guidelines

(If you’ve reached this page via a web search for help with voice recording, we hope you find something useful here. While you’re here, why not take a look at our journal as well? Our home page is at http://www.14by14.com.)

What are the general requirements for voice recording for 14 by 14?

I’m new to computer voice recording. What do I need to get started?

I have a separate digital recorder. Can I use that?

Why would I need a USB microphone?

What software should I use for recording?

How can I convert existing recordings to mp3 format?

What technical settings should I use for recording?

Are there techniques for good voice recording?

Why is my recording so quiet?

Why does my recording sound distorted?

Why is my voice almost swamped by hiss or hum on the recording?

How can I get rid of popping noises on P and B sounds?

What editing might I need to do?

 

What are the general requirements for voice recording for 14 by 14?

When recording your own reading of one or more sonnets for 14 by 14, please follow these general guidelines:

  1. Make a separate file for each poem.
  2. Record in mono: there’s really no point in stereo recording from a single microphone.   
  3. Use an appropriate recording level and distance from the microphone to ensure reasonable quality.   
  4. Say the title of the poem at the beginning (not your name) and then read the poem.
  5. Preferably, save and send files in MP3 format, with a sampling rate of 44kHz and a bit rate between 56 and 80Kbps. 

See the answers below for more details on some of these points.

I’m new to computer voice recording. What do I need to get started?

You’ll need a PC or Mac with a microphone and speakers, and suitable software. If your computer has a built-in microphone, that may do to get started. Otherwise try plugging a basic microphone into the appropriate socket, which on PCs is commonly colour-coded pink. Free or trial software for audio recording and editing is available online for download (some links are included under the Software question below).

It’s possible you may have trouble with background (system) noise and/or with the maximum recording level being too low. Usually a USB microphone will solve such problems — see below under Why would I need a USB microphone?

I have a separate digital recorder. Can I use that?

If you have a digital recording device (or an mp3 player or phone with recording capability) you may be able to use it to create a voice recording initially. The quality from a purpose digital recorder should certainly be good enough, but the voice recording feature on some mp3 players and most phones will not produce adequate quality. Listen to the recording through headphones to check the quality.

All such devices include means of transferring or copying the recorded file to a computer, via cable or wirelessly. Even if the recording is of adequate quality, it’s likely to need some editing — perhaps to cut out errors or to increase volume. So it is still highly recommended that you install and use suitable software on your computer. (See the later questions and answers on this page.)

Why would I need a USB microphone?

The input signal from a standard plug-in or built-in mike goes through the sound card, which typically adds unwanted hum and other background noise to the recording. It might also be hard to get the recording level high enough, forcing you to speak really close to the mike. If you do that, the result will probably be unwanted pops on plosives, and/or breath sounds on sibilants. These issues may or may not arise on your system, depending on the sound card, microphone, and recording technique.

A USB microphone will usually give better results and help to overcome the problems mentioned. It costs a little more than a standard microphone but is often worth it for the reduced frustration. A USB mike doesn’t go through the sound card; it plugs into any USB port and, on Windows XP and later, is automatically recognised as a sound input device when plugged in. You just need to select it as your recording source in the software you’re using.

What software should I use for recording?

The very basic Windows Recorder software is not recommended; it offers very few control and editing facilities. Here are some programs that can be downloaded in free or trial versions. Trial versions typically expire after 14 days or restrict use in some way.

WavePad has built-in support for saving in mp3 format. With Audacity or GoldWave you will need to install an add-on, the LAME MP3 Encoder, which can be downloaded for free (details at those program sites).

Note that the software products listed above all offer some editing features as well as recording. It’s highly desirable to use a recorder/editor and not just a bare-bones recorder. You might want to edit out mistakes, for instance, or change the level after recording. Be aware also that some recording and editing software might not support a USB microphone — the above three do.

How can I convert existing recordings to mp3 format?

If your existing recordings in WAV or some other format are of suitable quality and the level is good, you can convert them to mp3 with the free version of  Switch: http://www.nch.com.au/switch/index.html. Or you can use any of the recording/editing programs listed above to open an audio file and save it in mp3 format.

What technical settings should I use for recording?

As mentioned, choose mono rather than stereo recording. And if your software gives you a choice, save the file at a sample rate of 44100Hz (sometimes shown as 44kHz) and a bit rate of 80kbps or less. 56 kbps is usually fine.

The 44100Hz sampling rate is for compatibility with the Flash player system we use, and the modest bit rate is to keep the file size down. These are not essential but strongly preferred — as is sending the file in mp3 format. If push comes to shove you can send an original WAV file or whatever, but that will likely be a much bigger file attachment and will make extra work for us, something you surely would rather not do. 

Are there techniques for good voice recording?

These are some guidelines that may help.

  • Speak over or past the mike rather than directly into it, especially if you’re close to it.
  • Don’t over-enunciate consonantal sounds (especially P, B, T, S and SH), otherwise your recording may be marred by unpleasant pops, clicks and storm effects.
  • Try to use a recording level that won’t make it necessary to get too close to the mike. (More on this below.)
  • If you make a mistake, don’t stop the recording or you might lose your flow. Just pause for a second or two, then start reading again from the beginning of the phrase or line. That way it will be easy to edit the mistake out afterwards.

Testing and setting the recording level first is important. Good software lets you test and see the level before you start recording. Experiment to find the best combination of recording level and distance from the mike. With a standard mike going through the sound card it may be necessary to set the recording level near or at maximum to get a strong enough input signal. Then you still might need to increase the volume after recording — also, unfortunately, amplifying any hum or hiss from the sound card. A USB microphone will make things easier; you will probably be able to set the recording level somewhere in the middle, speak at a comfortable distance from the mike, and get an acceptable result.

After making the recording, listen to it and look at the waveform in your editing software. Here’s how a voice file looks in a recording/editing program when recorded at a good level.


Recording level OK.Waveforms should reach the 0.5 and -0.5 levels or beyond but should not touch the top and bottom edges of the window.


Why is my recording so quiet?

Here’s how a file looks when recorded at too low a level.

If you get this sort of result even with the recording level set as high as it will go, and you’re speaking close to the mike, then the mike is too “low gain” — not sensitive enough.

In this case the recording might be rescuable by editing to increase the level, but any hum or other underlying noise will be increased as well. The better answer is to re-record with a USB mike.


Why does my recording sound distorted?

Probably the recording level was too high. You will hear a blaring kind of distortion. The following is an example of how a file looks when recorded at too high a level.

This recording will not be usable. Most of the waveforms are clipped at the top and bottom — the sound is distorted. Reducing the volume by editing can’t fix this.


Why is my voice almost swamped by hiss or hum on the recording?

This problem arises with a low-gain microphone whose input goes through the sound card — typically the situation with a standard mike and a basic, rather than a better, sound card. You can’t get the recording volume up above a low level and you can’t get the background noise down.

Here’s an example of such a waveform as it appears in a sound editor. This was recorded through a standard PC mike plugged into the microphone socket (which connects the mike through the sound card).

   

Speech volume is rather too low here, but increasing it after recording will make the background noise louder too. The background noise comes mostly from the sound card, and may be be more or less noticeable according to the quality of the card.

You could spend a lot of time editing such a recording in hope of getting a reasonable result. Ideally you would remove the background noise cleanly, or greatly reduce it, using a “noise reduction” filter, and then increase the volume of the remaining signal (the voice sounds). But the sound frequencies contained in the hiss/hum overlap with those in the voice, so in practice reducing the noise is likely to make the voice sound a little odd as well. It may need experiment with different settings to find a good compromise where most of the noise is removed and the voice isn’t too obviously changed. Often it will be faster to re-record than to spend time trying to fix such a recording. Set the recording level higher and get as close to the mike as you can without producing exaggerated popping noises on P and B sounds.

Changing the recording level won’t alter the volume of the background noise, only of the input signal recorded from the mike. Increasing the latter will make it stand out more from the background noise.

If you can’t overcome the problem with adjustments of level and microphone distance, switch to a USB microphone. This bypasses the sound card and typically provides a better input level with close to zero noise.

How can I get rid of popping noises on P and B sounds?

This is a common problem with a cheaper microphone, especially where you have to speak right into it at close range in order to record at sufficient volume. It can happen even with better mikes, including USB mikes. If you can’t get around it by speaking further from the mike and/or speaking over or past it, the problem can be reduced with a pop shield. Some mikes come with one of these — usually a black foam sleeve. It’s possible to improvise a pop shield from a thin piece of coarse plastic or rubber foam, and this can make quite a big difference.

If you have pops on plosives and don’t want to redo the recording, you can try reducing them in your sound editor, which is likely to have a pop/click reduction filter. Use this with care and be prepared to undo the action and experiment with settings, so that wanted sounds aren’t removed along with the unwanted ones. 

What editing might I need to do?

Mistakes are easily edited out if managed using a pause-and-repeat approach during recording, as suggested above.  In the editing software select the section you want to cut out, then apply a Delete or Cut command. 

You might need to adjust the volume of the entire recording or of individual sections. Select the required section of sound and apply a volume change by percentage or by setting a different maximum volume.

If your recording contains pops or clicks, you might need to try applying a removal filter as noted in the answer to the previous question.

Some editing software offers an option to slow down or speed up the recording without changing the pitch of the voice. This can occasionally be useful if you are generally satisfied with the recording but think your reading was somewhat too fast or slow.

Thse notes are provided for authors who have had a sonnet accepted for 14 by 14. Please do not send sound files with submissions. If your work is accepted, you will be reminded that a recorded reading is an option.