Noise Reduction
- Peter 31Jan15: I archived the moving "spectral selection ability advice note" to the Talk page - removed the attendant P1.
- Peter 28Jan15: Updated image with defaults as set for 2.1.0 in r13893 28Jan15 nightly alpha
- Peter 28Jan15: @Paul: I removed your ednote on the functionality that is "hidden" for 2.1.0 - it broke the page formatting - and we have it already safely (a full copy of the previous page) on the Talk page. And anyway it is easily retrievable from the page history stored in the Manual wiki (try clicking on the blue "history" tab above.
- Peter 28Jan15: ToDo-2 Page almost ready for 2.1.0 release, subject to final editorial reviews.
- Bill 31Jan2015: Final edit? Expanded the Frequency Smoothing description, corrected error in the first bullet point under Tips, added 'musical noise' to artifacts description.
IMO this page is ready to go. - Gale 01Feb15: I'd like to review it within 24 hours. I had some specific things to add and want to check if they're already added.
To use Noise Reduction, you need a region in the waveform that contains only the noise you want to reduce.
Be aware that it may be impossible to get a satisfactory removal when the noise is very loud, when the noise is variable, when the music or speech is not much louder than the noise or when the noise frequencies are very similar to those of the music or speech.
If your problem is mains hum or a high-pitched whistle, the use of a Notch Filter may help, which should be carried out before applying Noise Reduction.
- Bill 13Dec2014: I disagree with James. For mains hum the frequency discrimination of the noise reduction effect is too coarse (about 22 Hz per band). Notch filters work much better on mains hum.
- Accessed by:
Step 1 - Get Noise Profile
This first step teaches Audacity about the noise you want to remove by identifying the noise floor of the different frequencies comprising the noise.
- Select a region of the waveform which contains only noise. A minimum of 2048 samples (0.05 sec. at 44100 Hz sample rate) is needed, below which an error will be shown. A longer profile is better. If there are very different types of noise in different places in the track, they are best dealt with by grabbing the profile for the first type, reducing the noise for it, then grabbing the profile of the next type of noise and reducing that.
- Click .
- Click Get Noise Profile.
Step 2 - Reduce the Noise
- "The Sensitivity slider biases the thresholds of all frequency bands. Each unit increase of Sensitivity increases the average time between artifacts approximately tenfold when the profile and the treated sound have a typical white or red noise spectrum. Higher settings will thus reduce the number of artifacts, but at the risk of introducing the opposite discrimination error, in which parts of the desired signal are misclassified as noise and so reduced.
When the Sensitivity is excessive, "tinkle-bells" will be heard in the Residue [...] and where the original had louder sounds, rather than in the pauses between sounds."
- Select the entire region of waveform from which you want to reduce the noise, then set the Noise Reduction parameters. This is often best done by trial and error, adjusting the sliders and using the Preview button to listen to a few seconds of audio after noise reduction. Listening to the Residue (the sound that will be filtered out when you apply "Reduce") can also be useful in determining how much damage is being done to the desired (non-noise) sound.
- Click .
- Noise Reduction (dB): Controls the amount of volume reduction to be applied to the identified noise. Use the lowest value that reduces the noise to an acceptable level. Higher values than necessary may make the noise even quieter, but will result in damage to the audio that remains.
- Sensitivity: Controls how much of the audio will be considered as noise, on a scale of 0 (minimum) to 24 (maximum). Greater sensitivity means that more noise will be removed, possibly at the expense of some damage to the desired signal. Set this to the lowest value that achieves effective noise removal without the introduction of artifacts.
- Frequency Smoothing (bands): This control should be left at 0 except in extraordinary circumstances. If you are still hearing artifacts in the noise-reduced audio, and have set the Sensitivity as high as possible without damaging the desired audio, try setting this control to a value between 1 and 6.
The Frequency Smoothing control can reduce the perception of artifacts by spreading them out over one or more frequency bands. The width (in hertz) of a frequency band is dependent on the sampling frequency of the track. Width = sampling frequency / 2048 = 21.53 Hz at 44100 Hz sampling rate.
Because frequency smoothing is applied to all bands, not just those where artifacts are present, it will often damage the desired audio.
- Noise:
- Reduce: Select this option to filter out the noise from the selection.
- Residue: Select this option to hear the sound that would be filtered out if you chose "Reduce". This is useful for finding the optimum settings that do not damage the audio. If you can hear recognizable bits of the desired sound in the residue, it is likely that you have set Noise Reduction too high or Sensitivity too high.
Artifacts can happen if the noise floor is higher than was estimated, either because Sensitivity was set too low, or because the noise profile was not representative of the noise throughout the track. Although most of the noise will be suppressed most of the time, random artifacts may occur, which can be even more objectionable than the original noise. For a more technical explanation of artifacts, see How Noise Reduction Worksin the Wiki.
Tips
After creating a noise profile, CTRL + R or will apply Noise Reduction at its current settings.
Reducing noise usually results in some distortion. This is normal and there is nothing you can do about it. When there's only a little bit of noise, and the signal (that is, the voice, music or other desired sound) is much louder than the noise, this effect works well and there's very little audible distortion. Unfortunately, when the noise is very variable or very loud (the signal is not much louder than the noise) the result may be too distorted.
I think "in the result" is ambiguous. Are we listening for tinkly bells and recognisable bits of the signal in the residue, or are tinkly bells and recognisable bits the same (recognisable bits that sound tinkly)?
If you are still having problems the following tips may help:
- Select
Residue from the Noise: options, click , listening for traces of the desired signal. You can usually reduce the amount of desired sound in the residue by decreasing Sensitivity or decreasing Noise Reduction.
- Decreasing Sensitivity increases the chance of there being artifacts sounding like "tinkly bells" in the result.
- Decreasing Noise Reduction will increase the amount of noise in the result, but may result is less distortion of the desired audio.
- When you have found settings of Sensitivity and Noise Reduction that appear to do the least damage to the desired sound (based on there being little of that sound in the residue), click the Reduce radio button then click to apply the effect.
- Do any Notch Filtering or Click Removal before doing noise reduction
- Do any amplification, compression or other effects after doing noise reduction, not before. It works best as close to the source of the noise as possible.
|
When using Noise Reduction in Chains the profile is captured as below.
|
Advanced Tips
- For a more technical explanation of the algorithm used in Noise Reduction, see Noise Reduction algorithm in the Wiki.
- For a more detailed explanation of artifacts and how to deal with them see Noise Reduction artifacts on the same page in the Wiki.
- Also see Sample workflow for LP digitization which lists some alternative noise reduction software that you can use as part of your audio restoration routine.
