Audacity is installed and used by millions of users for its excellent audio recording and editing capacity. To enjoy the great audio recording service of Audacity, you need to cope with the complex interface and preset. You must set up the audio source or audio tracks before starting an audio capture.
- Before we start we need to find a small section of the voice track that has no vocals recorded on it. We will use this section of the track to give Audacity a sample of the noise we want to remove. In the example track I have used the mouse to select a portion of the file but very often there will be a silent 'lead in' at the front of the.
- There are several Windows VST plug-ins that can be used in Audacity for vocal removal and isolation. For installation of VST plug-ins, see below. Voicetrap is a commercial plug-in featuring center channel removal by frequency-based and cepstral methods (that is, not a simple subtraction of one channel from the other).
- Audacity is shipped with many effects that you can use to modify your audio. Real-time preview of LADSPA, LV2, VST and Audio Unit (macOS) effects. Plug-in Manager handles plug-in installation and addition/removal of effects and generators from the menus. Generate tones, silence, noise, instru.
- Audacity is audio editing software that enables you to modify your audio files with advanced controls. One of Audacity's included features is the ability to listen to while you record or monitor a new track. This tutorial will walk you through the steps. Enable Software Play Through in Audacity.
The Isolate options all return (dual) mono output. The narrowness of the center slice can be adjusted with the 'Strength' slider. Center isolation is not possible using the classic subtraction method.
The Analyze option displays the amount of correlation between the stereo channels and the degree of likelihood that center removal or isolation will work effectively.
- See also the tutorial on Vocal Removal and Isolation
- Accessed by:Effect > Vocal Reduction and Isolation..
Action
Bandwidth Limited Actions
The Low Cut and High Cut sliders affect these actions. |
- Remove Vocals: to mono: (default). If the audio is center-panned, removes the vocal range defined by the Low Cut and High Cut sliders, and returns it as a dual-channel mono track. Audio is said to be 'center-panned' if it is common to both left and right channels. Cancellation of center panned audio is achieved by the well known 'invert and mix' method.
- This setting is identical to the 'Vocal Remover > Remove frequency band' effect in previous versions of Audacity.
- The Strength slider is not used when 'Remove Vocals: to mono:' is selected.
- Remove Vocals: If the audio is center-panned, removes the vocal range defined by the Low Cut and High Cut sliders, and returns it as a stereo track.
- Isolate Vocals: If the audio is center-panned, extracts the slider-defined vocal range and returns it as a (dual) mono track.
- Isolate Vocals and Invert: If the audio is center-panned, extracts the slider-defined vocal range and returns it as an inverted (dual) mono track.
- This option is equivalent to applying the Isolate Vocals option and then inverting.
- The 'and Invert' options may be useful when processing using duplicated tracks. See the examples below.
Full Spectrum Actions
The Low Cut and High Cut sliders have no effect with these actions. |
- Remove Center: to mono: Removes all audio (the whole frequency spectrum) that is common to both left and right channels and returns a (dual) mono track. Cancellation of center panned audio is achieved by the well known 'invert and mix' method.
- This setting is identical to the 'Vocal Remover > entire spectrum' effect in previous versions of Audacity.
- None of the sliders are used when 'Remove Center: to mono' is selected.
- Remove Center: Removes all audio (the whole frequency spectrum) that is common to both left and right channels and returns a true stereo track.
- Isolate Center: If the audio is center-panned, extracts the whole frequency spectrum and returns it as a (dual) mono track.
- Isolate Center and Invert: If the audio is center-panned, extracts the whole frequency spectrum from and returns it as an inverted (dual) mono track.
Analysis Action
- Analyze: Tells you how great the chances are of a successful vocal reduction or isolation. It returns also the average Pan position for the selected audio. See below for an in-depth explanation.
Strength
The shape of the center is not a thin band but rather like a tent with a pole in the middle. This means that there will always audio from other pan positions included. The Strength slider modifies the shape of the center. Higher values increase the degree of reduction or isolation. Zero is off (no reduction or isolation). This slider has no effect with the 'Remove Center Classic: (Mono)' choice.
Remove
- Less than 1.0: Produces a notch with a V shape like a ship's keel, but also with a sharp point. Use this setting to preserve some amount of the center.
- 1.0 - the default: Produces a notch with a V shape. The power distribution is preserved. This is the ideal setting for most cases.
- Greater than 1.0: Produces a notch with a U shape. This removes some audio adjacent to the center, as well as the center. Note that this will not remove stereo reverberation since it is distributed over the entire pan width.
Isolate
- Less than 1.0: Produces a peak like a U shape upside down, similar to an arch. Only extreme left and right are eliminated.
- 1.0 - the default: Produces a peak like a V shape upside down. The power distribution is preserved. You might want a higher value when a lot of side audio leaks in. This is also the recommended setting when working with two tracks (options with ..and Invert).
- Greater than 1.0: Produces a peak with an A shape like the Eiffel Tower (with a sharp point). This will attenuate most of the side energy but might produce artifacts or musical noise. Normalize the audio after using Isolate with a high strength setting.
Low Cut for Vocals
All actions with Vocals in the name use this value. The frequencies below will either not be removed or not be included in the isolated audio. It is useful to exclude bass and kick drum. The default value of 120 Hz is good for most lead vocals, or even low male voices. You can enter a higher value around 170 Hz for female voices and about 230 Hz for those of children.
High Cut for Vocals
All actions with Vocals in the name use this value. The frequencies above will either not be removed or not be included in the isolated audio. It is useful to exclude high sounds like bells, cymbals or Hi-hat. Note that human sounds like S or Z can also be very high in frequency - 5000 to 8000 Hz, so listen carefully to the preview.
Buttons
Clicking on the command buttons give the following results:
- gives a dropdown menu enabling you to manage presets for the tool and to see some detail about the tool. For details see Manage presets
- plays a short preview of what the audio would sound like if the effect is applied with the current settings, without making actual changes to the audio. The length of preview is determined by your setting in Edit > Preferences > Playback, the default setting is 6 seconds.
- applies the effect to the selected audio with the current effect settings, but unlike the effect runs in debug mode. This is primarily of use when writing or editing Nyquist plug-ins.
- In addition to the normal plug-in behavior, a 'debug window' opens to display error messages, normally the debug window will be empty
- applies the effect to the selected audio with the current effect settings and closes the dialog
- aborts the effect and leaves the audio unchanged, closing the dialog
- brings you to the appropriate page in the Manual, this page
Interpreting the analysis results
It is recommended to analyze the audio prior to actual processing of the effect. Analysis is very fast compared to the time needed for processing.
The two important values are:
- Pan Position: This gives the average pan position for the whole selected audio. The value of the track pan slider is not included in this calculation.
- Correlation: This is a measure for the similarity of the two channels. Values of +/-100 mean that the channels are exactly the same, even if one is inverted. A value of 0 means no relationship. It is fruitless to attempt removal or isolation in these cases. The ideal value is normally around +50. Values below 0 are rare and indicate that the stereo width is more than 100%. Inverting one channel makes the correlation positive.
- The correlation should be fairly high (around 100) which would indicate that there is indeed only the voice.
- Copy the value for the Pan Position. Double-click in the Pan slider to open the adjustment dialog then paste in the value with reversed sign (that is, put a minus sign in front of a positive value and a positive sign in front of a negative value).
- Mix and render the track. The voice will now be exactly centered and you can remove or isolate it.
- Don't forget to readjust the selection.
Limitations
- The input must be a true stereo track and not mere (dual) mono.
- Stereo Reverberation will not be fully removed because it is independently distributed over the whole stereo image.
- Naturally the plug-in does not know what kind of audio is in the center. All is removed or isolated equally, whether vocals, bass or solo instruments.
- This effect is quite slow, except for the 'Remove Center Classic: (Mono)' and 'Analyze' actions.
The whole processing is done in memory. Long selections (over half an hour) might therefore cause Audacity to crash. |
Examples
This effect can be used in a variety of ways:- Creating a Karaoke version from an original song
- Choose Remove Vocals if you want to keep some bass and drum beat.
- Creating an acapella version from an original song
- Choose Isolate Vocals with a high Strength value. The result will probably still have some music in it. It is therefore better suited for a remix with a different accompaniment. Pure vocal versions have to be post-processed with other tools.
- Alternative to 'Auto Duck' with podcasts
- Duplicate the track and choose Isolate Center and Invert on the second track. Record your voice. Silence the second track where you don't speak and the original music should be heard. On the same track, make fades at the boundaries (where the speech starts or stops).
- Removing excessive stereo reverb from a vocal or instrument track.
- Use Isolate.. to make a single vocal or instrument track 'dry' again.
- Removing random system noise from a two channel recording of a mono source.
- use Isolate.. to remove random noise produced by the audio interface, the cables or the computer itself.
- Removing single words or phrases
- You can remove offending content from a song or movie dialog by selecting it and choosing Remove Vocals. Mask it with other sounds, if necessary.
- Amplifying dialogs in a movie.
- Duplicate the track, isolate the vocals and adjust the gain.
- Converting stereo files into 3.1, 5.1 and other multi-channel formats.
- Duplicate the track and choose 'Isolate Center and Invert' on the second track. Render all to a new track. Delete the first track, invert the second track and split the third track. Rearrange the tracks accordingly. Thus you end up with 'Front Left', 'Center' and 'Front Right' eventually.
- Extracting instrument solos
- Isolate the center if the solo is there, otherwise remove the center, split to mono and delete the superfluous track.
- Blending between original, karaoke version and vocals only
- Duplicate the track and choose Isolate Vocals and Invert on the second track. Choose the envelope tool or fade effects to gradually change from one mode to the other. Silencing the first track results in the vocals being isolated. Silencing the second track produces the original audio. Playing both tracks together removes the vocals.
- Correcting the pan position
- Analyze the whole track and copy the value for the pan position. Paste the value into the pan dialog and reverse the sign, that is, set a minus sign in front of positive values and vice versa.
- Measuring similarity between two mono tracks
- Make the two tracks stereo and choose Analyze. The correlation value is the measure for the similarity in percent.
- Applying a brick wall filter to mono tracks
- Duplicate a mono track, make it stereo and choose Isolate Vocals for band pass and Remove Vocals for band stop filtering. Adjust the frequencies accordingly.
Links
|< Index of Effects, Generators and Analyzers
You may not need to change any of these settings as Audacity's default settings are designed to work well for you right from the start, so much so that you can probably skip-read this page.
Your first recording with a microphone is probably best done without playthrough of additional sounds, and unless you know you have a stereo microphone, using mono.Nothing goes 'wrong' if you attempt to record in stereo with a single microphone, though you may find that only one of the two stereo tracks has sound.
To just use the sound from the microphone:
- Go to Edit > Preferences (or Audacity > Preferences on Mac) then select the Recording section, and make the following settings in the Playthrough section
To record in the most standard way:
- in the Sound Activated Recording section
- in the Options Section
- in the Sampling section
- Default Sample Rate:
- Default Sample format:
Project Sample Rate
The Project Sample Rate in Hz can be seen, and changed if required, in the Selection Toolbar at the bottom of the Audacity window:
The sample rate defines the number of digital audio samples per second taken from a continuous analog signal.
44100 Hz is standard for creating files that can be burned to Audio CD or for high-quality MP3 files.
This is the default setting in Audacity and is best left at that setting for your recordings.
Project Sample Format
Voice Recording In Audacity
The Project Sample Format (also known as Bit Depth or Word Size) is the number of computer bits present in each audio sample. It determines the dynamic range of the audio.
Audacity's default is set to 32-bit float. This is good for editing and processing audio. It is strongly recommended that you retain the default setting, however if you really need to change your sample format this can only be done in the Quality section of Preferences.
You can easily downsample to 16-bit (the standard for creating WAV files that can be burned to audio CD) when exporting later.
Temporary Files
Most users have only one hard disk. However, if you have multiple hard disks, you will want to make sure that Audacity uses your largest or fastest hard disk to store audio. Open the Preferences (in the Edit menu, or the Audacity menu on Mac) and click on the Directories tab. Make sure that the directory listed is on your preferred hard disk.
If your home directory is mounted from a remote server, you definitely do not want Audacity's temporary directory to be there! |
Software Playthrough
Now click on Transport > Transport Options and, if you are recording from a microphone, make sure that Software Playthrough is not checked - Software Playthrough will cause undesirable feedback from the computer speakers to the microphone.
If you are recording a guitar or keyboard and want to hear the instrument through the computer speakers, make sure that Software Playthroughis checked.
Hardware playthrough: This is a method available on some Windows and Linux machines that lets you unmute the input directly in the sound device. To use this, open the operating system sound mixer, then the Playback section, then unmute the input device (microphone or line-in) and turn the volume up. You can find the system mixer in the Control Panel at Sound on Windows. On Linux, it may be possible to unmute input devices in ALSAmixer then set playback and recording devices in Audacity to the (hardware) device.
If hardware playthrough is required it is best to use an external USB or FireWire audio device that has a headphones jack for no-latency monitoring (such as the Behringer UCA 202 or Zoom H2). |
Overdub off
Now click on Transport > Transport Options again and make sure that 'Overdub (on/off)' is not checked. When this option is enabled Audacity will play other existing tracks while recording a new one. You can decide which tracks will play according to their mute/solo buttons on the Track Control Panel.
It is unlikely you will need overdub for your first recording. If you do want to overdub by (for example) singing over a recorded instrumental track, you should be aware of the Latency correction preference.
Latency correction
Open it zip. On most computer systems there will be a delay known as latency between singing or hitting your note and it being laid down in the recording. When Transport > Transport Options > Overdub (on/off) is set to 'on' and you record with another track already present, Audacity will push the recorded track backwards by 130 milliseconds to compensate for the delay. If your latency is constant, you can adjust the Latency correction value so that your recorded tracks should end up properly synchronized with the other tracks after correction. To set a custom latency value for your system, see the Latency Test page.
Sound Activated Recording
Also at Transport > Transport Options make sure that 'Sound Activated Recording' is not checked. 'Off' is the default setting. When i movie.
When this is enabled, recording will automatically start or resume when the recording volume rises above the chosen threshold level, and pause when the level falls below that threshold. You cannot pause Sound Activated recordings manually using the Pause button or corresponding menu item or shortcut.
Audacity Preferences
If you prefer, you can make the above settings in your Audacity Preferences. To access this, use Edit > Preferences.. (or Audacity > Preferences.. on Mac).
The image below shows the Recording section of Preferences:
There are many other settings that can be made in Preferences. Do take some time to explore and understand these.
Audacity Music Recording Software
For example Sample Rate and Sample Format can be changed in the Quality section of Preferences.
Links
> Forward to: Tutorial - Selecting Your Recording Device
Audacity Recording App
< Back to: Tutorial - Connecting Up