Wikinews:Audio Wikinews/News Briefs/Show/Guide
Please read the Style Guide before doing any recording. This will help you sound great and make the show better overall. The tips included on the style guide page are part of the minimum quality guidelines needed for recording the AW (just as articles published on wikinews have minimum standards they too must meet). If you cannot meet these requirements, your segment may have to be edited out of the show.
To record, you do not need a fancy microphone, but I do use a USB microphone with an XLA to USB input and get great results (My microphone is a Behringer ULTRAVOICE XM8500 and it cost only $30.00US), however, many headphone style (gaming) microphones do not produce a high enough quality of sound. Basically, if your recording sounds like you are speaking from a telephone, then you will need to invest in something with a little higher quality.
PROTIP : I cover my microphone with a small cloth bag. I do this because it eliminates the popping sound some consonants produce and eliminates a bit of the background hum of my computer's power supply. Also, after you record, please check your waveform. If your waveform is overflowing the top and bottom of the clip, your mic is to loud (hot), so please make adjustments to your levels so that things sound "normal". See the screen captures below for a good example of how the waveforms should look.
Please use Audacity. It's open source and easy to use / learn. Also, all files you record need to be saved in the .ogg format. We do this for simplicity sake and because .ogg is also open source (unlike an .MP3, .MP4 or .WMA).
For our example, let's use an actual story from Wikinews. The story is Nelson Mandela's great-grand daughter dies in car crash and is fairly short but will highlight basic techniques of recording with Audacity.
- The great-granddaughter of former South Africa President Nelson Mandela has died in a car crash following a concert to open the World Cup.
- The Nelson Mandela Foundation released a statement which said that thirteen-year-old Zenani Mandela, who celebrated her birthday on June 9, died in a single vehicle accident and that no one else was injured. The statement continued: "The family has asked for privacy as they mourn this tragedy."
- South Africa has a poor road safety record and ranks ninth in the world for traffic fatalities. Traffic safety is feared to be a threat of injury to supporters to the World Cup.
First, record each paragraph separately. Doing one long take will only make it harder to go back and edit and since the show is pre-recorded you must edit all your mistakes (stumbles, pronunciations, background noise, etc ...), so break the reading up into small pieces.
I use Notepad++ as where I paste text into for reading because I can make the text really big on my screen (CTRL & mouse scroll wheel) for easy reading. I then keep the Audacity window open as a small, second window where I can use the mouse to hit the record and stop buttons easily. Once I hit record, I pause for 1-2 seconds, then start speaking then pause for another 1-2 seconds when I'm done with the section. This is important because not only does it slow you down and keep you from rushing, but it gives you some extra room to work with in Audacity that you will really need.
When I make a mistake (and I make a ton), I just stop recording, delete the waveform and try again. It's a slow process, especially when you do 10-15 takes of just one sentence, but that's what it takes to do it right as even professional announcers and broadcasters do this. If you are doing everything in 1 take every time, then not only are you the first person in the history of recording to do so, but you are also doing it wrong. In other words, expect to do multiple takes of each paragraph. You are not just reading the news, you are a presenter and entertainer as well - so make it sound good.
A very common recording (and speaking) mistake is to not follow punctuation correctly. In the following sentence :
- "The Nelson Mandela Foundation released a statement which said that thirteen-year-old Zenani Mandela, who celebrated her birthday on June 9, died in a single vehicle accident and that no one else was injured."
there should be a pause (and a tonal shift) between "thirteen-year-old Zenani Mandela", "who celebrated her birthday on June 9" and "died in a single vehicle accident and that no one else was injured." because there are commas there. Reading it straight through does not sound correct, so please follow punctuation - remember, you want to make the story interesting for the listener because flat, dry readings are not interesting.
See the style guide for further tips on how to improve your reading and speaking.
Raw Audacity Output
Once you have your recording done, Audacity should look like this :
Notice how each paragraph has its own waveform file : there are three paragraphs and thus there are three waveform files.
Removing Background Noise
Now, the first step is to clean up your background noise. Unless you have a million dollar recording studio, you will have background noise, so you need to get rid of it.
First, find a spot in any one of the waveforms where there is empty space. Usually, I use the few moments pause I add at the end of a segment, and highlight just that portion.
Next, you want to get a noise profile. Getting a noise profile will store the sound of the background noise you just highlighted and when we are done, it will remove that sound/noise from the entire recording.
The noise profile tool will bring up this box:
Click the Get Noise Profile button.
The box will close, and you will be back to the main screen.
You can now highlight everything (CRTL+A) ...
... and all the waveforms will be highlighted.
Now, go back into the Noise Removal tool ...
and the Noise Profile box will pop up again. Slide the arrow in Step 2 on the box all the way to the left and hit Remove Noise. Mileage will vary with this feature so play around with it as best you can. Some people may have more noise and thus not want to slide the arrow all the way to the left - experiment with the feature until you get it sounding clear.
I use a super cheap pair of old ear-bud headphones to test how much background noise I have. Good speakers and headphones will actually hide noise, but cheap ($5.00) headphones don't and thus are perfect for analyzing your recording. Since you never know what your audience is using to listen to the show, assume they are using the cheapest and worst equipment possible and make it sound as good as possible for them.
Normalize The Track
You have one more step to go before editing the whole story together.
You want to get the audio all one volume so that parts aren't louder than others. To do this, use the amplify tool:
This tool will balance out the volume and set the maximum volume of the recording. Now, this is a bit tricky, but what you need to do is get the volume of the recording not too loud or too quiet.
What I do is test the volume of my speakers to find a good, comfortable volume (you can play an mp3 or video) and then match that volume level in Audacity by using the slide bar in the Amplify tool. I set mine to Amplification (dB) 15.0 (your mileage will vary, however).
Hit OK and it will amplify and normalize the file and your waveforms will now look like this:
Trimming The Tracks
Go ahead and trim the beginning and ends of each waveform to get rid of excess empty space (don't trim too much, you want to have room to work with).
Editing Tracks Together
Once they are all trimmed up, they're now ready to be edited together to sound like one long, mistake-free take. To do this, use the Time Shift Tool in the upper left of the screen (it looks like a horizontal line with an arrow on each end).
Move your waveforms into place (you can use CTRL-F to fit all the waveforms into the window if you are having trouble seeing them).
Final Quality Control
Finally, hit the play button at the top of the screen (big green triangle) and listen to the recording all the way through. Listen for any errors and make sure it flows nicely. If you put the beginning of one waveform too close to the end of another, it will sound like you are talking too fast so make it sound like your normal conversational cadence. Remember, it is very important to listen to the entire recording before finishing, otherwise mistakes will get past you.
Once your segment sounds great, save the project ...
... and, finally, export as Ogg Vorbis (this creates a .ogg file of your recording) using the appropriate file naming convention for the project you are recording for.
The Finished Segment
Congratulations, you are done!
This section will deal with adding audio from other sources (importing) as well as stripping audio from a video file in VLC.