Spectral Editing Software

[Athenasby Regalado]

Amongother purposes, spectral selection and editing can be used for cleaning up unwanted sound, enhancing certain resonances, changing the quality of a voice or removing mouth sounds from voice work.

To define a frequency and time range change the track to Spectrogram view, click and hold the mouse button at a vertical position that you want to be the upper or lower frequency boundary to be.

Drag vertically (with or without continuing to drag horizontally) to define the bandwidth (range of frequencies) to be acted on. A "box" containing a combined frequency and time range is now drawn in a colored tint as shown below (the exact color of the tint will depend on the version of Audacity and the settings of your monitor):

If either the low or high frequency limits are not defined, no center frequency will be shown. In the illustration below, the selection has been dragged from about 5,000 Hz to the bottom of the track. The high frequency selection is about 5,000 Hz and the low frequency selection is not defined, so the selection includes everything below 5,000 Hz:

When you hover the mouse pointer over the center frequency line it changes to a double triangle to show that the center frequency can be dragged. When you click and drag to move the center frequency and the current frequency range to a new position, the center frequency will snap to frequency peaks. To move the center frequency smoothly (without snapping to frequency peaks) hold down the Shift key then left click and drag.

To adjust the bandwidth of the selection hover the mouse over the upper or lower boundary until the pointer changes to a double triangle then left click and drag. This moves the upper and lower frequency boundaries so they remain equally spaced around the center frequency (thus not changing the center frequency).

To adjust the upper or lower boundary while holding the opposite boundary constant hold down the Shift key then hover the mouse near the upper or lower boundary until the pointer changes to a single triangle then click and drag (this will change the center frequency).

If you add a label at a time region, that region's Low Frequency and High Frequency are stored in the label and are recalled when you select the label. Similarly if you add a label at the cursor position, the frequency at which you clicked to set the cursor is stored in the label.

If you subsequently want to change the Low Frequency and High Frequency stored by the label, select and right-click the label (or use the Menu key) then choose "Edit" to open the label in Label Editor.

Press the shortcut Q key to change from the last selected spectral selection (one that has at least the lower or upper boundary defined) to one where the lower boundary is 0 Hz and the upper boundary is the Nyquist frequency of the track. In this case there is no longer a spectral selection (all frequencies are selected which is the same as making a time selection in waveform view). Press the Q key again to restore the previous upper and lower boundaries. This command works whether in a Spectrogram view or not.

The permitted upper frequency of the spectral selection is determined by the Project Sample Rate and not by the track sample rate. Therefore if a project includes a track that has a lower sample rate than the Project Sample Rate, Spectral Selection Toolbar could display an "illegal" frequency selection above the Nyquist frequency for the track at the lower Project Sample Rate. Audacity will disregard this "illegal" selection when running Spectral edit effects.

Use this to greatly attenuate or remove part of the frequency range in the spectral selection, for example to repair damaged audio. Depending on how the track is displayed and what spectral selection has been made, this effect will do one of the following:

Use this like an equalizer to modify the tone quality, boosting or reducing the amplitude of the selected frequency band using the Gain control. As with an equalizer, frequencies outside the selected band may be boosted or reduced, but less so than those within the frequency band.

A shelf filter curve is flat at both ends rather than flat only at the frequencies being passed. Use these filters when you do not want to reduce the lowest bass or highest treble frequencies in the spectral selection as much as you would with a standard high pass or low pass filter, or if you actually want to boost either the low or high frequencies respectively.

Snap the center frequency line to the whistle by hovering the mouse over the center line until the double triangle cursor appears, then click and drag to snap the center line to the whistle frequency:

@StudioES said:

Virtual ANS 3 has a sonogram editor - "you can draw the spectrum and play it at the same time". Also converts wav or jpeg to sonogram. AUv3. Not sure how deep the editing is though.

Don't be mislead by screen graphics - desktop vendors like Izotope, Celemony and Zynaptiq operate at a frequency resolution that can't even be displayed.

(iirc 1/100 cent, or 1/10000 of a semitone)

Such amount of detail requires dedicated software routines to keep up with (near)realtime demands.

I just use it for correction noise issues in Reaper. I guess I haven't thought about how I could creatively use it like I should. If I thought about that more, I'd use it on iOS. I find the accurate touchscreen is more enjoyable for fine sample editing usually.

I am sure there is no universality of terms but what Spectral editing is in Reaper is a more elaborate way to edit audio. Like zoom in a see the frequency and time domain and you can fix noises, little blips and bloops. I am guessing it can be used creatively but I just haven't gone there.

@Multicellular said:

I just use it for correction noise issues in Reaper. I guess I haven't thought about how I could creatively use it like I should. If I thought about that more, I'd use it on iOS. I find the accurate touchscreen is more enjoyable for fine sample editing usually.

The hint by @audiblevideo about Infinite fits this context. Such 'harmonics' may not be exactly where you'd expect them by math rules. This natural displacement is a main feature of Infinite, subtle, but essential.

But Infinite works reverse to spectral editing: it has methods to generate such content, while spectral editing kind of tries to discover rules from whatever output is presented on screen.

It can't do this fully on it's own, so usually one has to mark areas supposedly belonging together on the screen.

Machine learning might improve the handling when it gets more popular.

Imh understanding (at the moment) spectral editing is handled in a form of re-synthesis.

This doesn't come without artifacts or at least with a significant change of the original's sound character. It's a thin line between nice and exaggerated.

@Telefunky said:

But not a very convenient operation in this context: while you zoom into one frequency range, it's harmonics (on screen) get further and further away from the choosen point of interest

Hard-cutting spectral data and doing the inverse transform can cause weird artefacts, especially by treating phase and magnitude differently.

I hate them but indeed they can be used for creative sound design.

The site is secure.

The ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Water suppression is typically performed in vivo by exciting the longitudinal magnetization in combination with dephasing, or by using frequency-selective coherence generation. MEGA, a frequency-selective refocusing technique, can be placed into any pulse sequence element designed to generate a Hahn spin-echo or stimulated echo, to dephase transverse water coherences with minimal spectral distortions. Water suppression performance was verified in vivo using stimulated echo acquisition mode (STEAM) localization, which provided water suppression comparable with that achieved with four selective pulses in 3,1-DRYSTEAM. The advantage of the proposed method was exploited for editing J-coupled resonances. Using a double-banded pulse that selectively inverts a J-coupling partner and simultaneously suppresses water, efficient metabolite editing was achieved in the point resolved spectroscopy (PRESS) and STEAM sequences in which MEGA was incorporated. To illustrate the efficiency of the method, the detection of gamma-aminobutyric acid (GABA) was demonstrated, with minimal contributions from macromolecules and overlying singlet peaks at 4 T. The estimated occipital GABA concentration was consistent with previous reports, suggesting that editing for GABA is efficient when based on MEGA at high field strengths.

I should add that I was able to select the frequencies again (I had forgotten to re-enable spectral selection when I re-started), so right now I have just the frequency range I want highlighted, but how do I cut just these out, so I can then paste that into a new file, save that, and then save the remaining un-cut frequencies in another file?

3a8082e126