Selecting a part of the waveform usually results in a contiguous selection, starting where you clicked the mouse up to the point where you released it.

A non-contiguous selection can be thought of as being composed of several simple selections, several regions are selected simultaneously. Here is how a non-contiguous selection can look like:

Although TwistedWave has always been able to handle non-contiguous selections, it has recently gained more visibility with the silence detector, because it uses a non-contiguous selection to mark the silences in an audio file.

Making a non-contiguous selection

When you click the waveform, the current selection is discarded, and you start making another selection. In order to make a non-contiguous, you can option-click, and the existing selection will remain, and you start adding to it.

Additionally, you can option-shift-click and drag the mouse to remove an area from the selection.

Exactly as with a simple selection, you can adjust a non-contiguous selection by clicking on its edges, or shift-clicking if you want to avoid accidentally deselecting everything by clicking just a bit too far.

What for?

What can this be used for?

The interesting point about non-contiguous selections is that all the effects that can be applied on a selection also work on non-contiguous selections, and if you want to apply an effect with the same settings on different parts of an audio file, instead of selecting a piece of audio, apply an effect, selecting another part, and applying again, you can select all the parts you want to process simultaneously, and apply the effect.

The nice part is that TwistedWave remembers the selection you applied the effect on. If you want to apply your effect with a different preset, hit undo, and the audio you were working on is automatically reselected.

Also, as indicated above, a non-contiguous selection is a very helpful tool to mark the silences automatically detected in the sound file.

TwistedWave 1.8 was just released. The detailed list of new features, fixes and improvements can be found in the release notes. It can be downloaded from the home page. The main new features being the availability of a 64-bit version of TwistedWave, automatically splitting a sound file by detecting silences, and the ability to export iPhone ringtone files.

TwistedWave 64 bit

With Mac OS 10.6, Apple has completed the transition to 64 bit computing, and it is time for TwistedWave to jump on the bandwagon.

By going 64 bit, TwistedWave benefits from the ability to use a lot more memory, and deliver faster performance.

Running TwistedWave 64 with 32-bit Audio Units

In theory, a 32 bit application can load only 32 bit plugins, and a 64 bit application can load only 64 bit plugins. This would mean that TwistedWave 64 will not be able to load 32 bit Audio Units.

The problem is that today, the vast majority of Audio Units available is 32 bit only. In fact, from the hundred of plugins currently installed on my mac, all of them are 32 bit, except from the built-in Apple Audio Units, since Mac OS 10.6.

Using 32-bit Audio Units is possible

Without being able to run 32-bit Audio Units, a 64 bit TwistedWave would be useless to many.

For that reason, TwistedWave 64 incorporates a new technology that makes it able to load 32 and 64 bit Audio Units transparently. You will not even have to know you are running a 32 bit Audio Unit with TwistedWave 64.

Silence detector

If manually cutting a long file into many different parts, and saving them as individual files looks like a big waste of your time, you will be happy to use the silence detector built in TwistedWave that allows you to do exactly that.

In just a few steps, you can:

• Detect the silences
• Name the different parts
• Export as individual files

You can read more on how to automatically split an audio file.

Manually cutting a long file into many different parts, and saving them as individual files can be very tedious and quickly take a lot of time. In this article, I will demonstrate how easily this can be done with TwistedWave, thanks to the built-in silence detector.

Detecting the silences

When you want to split an audio file, the first task consists in detecting the silences, parts of the audio where there is no or little sound. Each time a silence appears in the file, this is a place where it should be cut. The silence detector is perfectly suited to this job. From the main menu, click Select / Detect Silences…, and here is what you get:

The simplicity of the interface reflects the fact that the task is fully automatic, and TwistedWave is able to detect the silences without any extra configuration. If you are working on a more complex sound file where the difference between the silences and the useful sounds is not obvious, you can see below how the expert mode of the silence detector can be used to fine-tune the way TwistedWave tells the silences from the rest of the file.

When the silence detector is opened, it first analyzes the audio, and previews the silences by selecting them, like this:

Seeing the image below may surprise you if you thought the selection always had to be contiguous. Well, it does not have to be, and this can be particularly useful in cases like this one. You can read more on non-contiguous selections.

When previewing the silences, you can start playing the audio, and TwistedWave will play all the selected silences. That way, you can make sure a useful sound was not selected by mistake.

After the audio was analyzed, from the pop-up button, you can select what to do with the silences in the audio file. When pressing the Apply button, one of several actions will be performed:

• Select Silences. This option will select the parts of the audio containing silence.
• Select Sounds. This is the opposite of the previous option, and only the areas not containing silence will be selected.
• Place Markers. Markers will be automatically placed at the beginning and end of each detected silence.
• Delete Silences. All the silences will be deleted.
• Delete Silences and Mark. All the silences will be deleted, and some markers will be placed where the cuts were made.

The option we are going to use is Delete Silences and Mark, which will give this when applied:

Naming the different parts

TwistedWave has automatically placed some markers at the places where the sound file should be cut. Now, it is convenient to give more explicit names to the markers, because they are going to be used to name the files.

Instead of double-clicking the markers one by one to rename them, the markers window provides a convenient way to rename all of them in one go. Double click the first one, and when you press return, the second is automatically selected to be renamed. You can also check the “Auto play” checkbox, and TwistedWave will automatically preview the audio starting at the marker you are renaming.

Splitting in individual files

Once you have placed the markers, and renamed them, the work is almost done. Select “Split by Markers” form the Markers menu, select the destination folder and file format, and TwistedWave does the rest:

The expert silence detector

On some occasions, there is no clear distinction between the silences and the rest of the file, and the simple detector will not correctly locate the silences in the sound file. In that case, you should switch to the expert mode, and you will have access to a number of settings that allow you to fine-tune the detection.

The expert silences detector looks like this:

A number of new parameters are available to control the silence detection. They are:

• Threshold. This parameter specifies the sound level in dB below which a sound is considered as silence. Note that when you open the silence detector, the sound file is analyzed, and this parameter is automatically set to a value that should be fine. You should only have to change this parameter in some cases where the distinction between sound and silence is less obvious.
• Minimum silence duration. This parameter specifies the minimum duration below which a silence is not considered. This can be useful when detecting silences from a file containing spoken text, and a small pause between two words should not be considered as silence.
• Minimum sound duration. This parameter specifies the minimum duration below which a sound is not considered as such, and will be treated as a silence. When a short noise is present during a silence, it can be desirable to still consider the silence as a whole, and ignore the sounds up to a given length.
• Left and right padding. When you want to cut the silences from a sound file, it may be necessary to keep around a fraction of a second of audio before and after each sound. If the sound fades out at the end, for instance, it may be necessary to keep it a bit longer even after it drops below the silence threshold.

For best results, here is how you should set the different parameters in the expert detector:

• First, set all the parameters to 0, except the threshold.
• If necessary, fine-tune the threshold so that all the silences are detected, even if some silence is found in the useful sounds.
• Increase the minimum silence duration just enough for the small silences in the middle of a sound to be ignored.
• If necessary, increase the minimum sound duration to ignore a short noise inside a silence.
• Increase the left and right padding values to keep a short silence before and after each useful sound.

Manually tweaking the silence detection

If the difference between the sounds and the silences is so small that the expert silence detector can’t make a correct detection, you still have the option to manually adjust the detection.

First, adjust the detection parameters as well as possible, and instead of cutting and marking the silences right now, select the option to “Select Silences”. Apply this, and close the silence detector.

Exactly as above, you should end up with a non-contiguous selection in the audio file:

You can adjust this selection by:

• Shift clicking its edges to adjust them.
• Option-click and drag the mouse to add a region to the selection in order to mark a silence that would not have been detected.
• Shift-Option-click and drag the mouse to remove a region from the selection to remove a silence that should be ignored, or to mark a sound that was ignored within a silence.

When this is done, make sure the option to Auto Mark Cuts is enabled, from the Markers menu, and press backspace, or Edit / Delete from the main menu. Then, export the files with “Split by Markers” and you are done.