User Manual

Widener

 

---

Intro
On stereo widening...
User Interface
Algorithms
Goniometer
Correlation Meter
Output Section
Credits

 

Intro

Softube Widener is designed to create a stereo signal from a mono track or widen an existing stereo track through the use of different widening algorithms. Each algorithm has its own unique sound, benefits and quirks, and since stereo widening is as much a matter of taste as it is mathematics, we have chosen to include a wide variety of types.

 

On stereo widening...

Stereo widening is the art of taking a mono track, or something almost mono, and make it stereo. (To simply increase the gain of the side in an M/S matrix won't make a mono track stereo, it will simply increase the amount of stereo already in there, so we're not talking about that now.)

Traditionally stereo widening has been achieved by using a lot of different techniques, for example recording a source with stereo mics, running mono audio through a stereo chorus, running it through a reverb, or using delays to create the stereo image. Each of these methods have drawbacks: recording something will also add the sound of the room, a delay will add audible echoes, and a chorus will sound, well … chorus-y. But they are all valid techniques, and if you want to get a wide stereo sound as well as a delay effect, using a stereo delay will kill two birds with one stone.

But what if we don't want a delay or a chorus effect? What happens if we take such techniques and try to minimize the side effects?

First we need to get comfortable with the fact that the is no way to add stereo information without changing the sound. It sounds simple when we say it like that, but just the fact that we're adding something (stereo stuff, remember?) means that we are also changing something. Let's call the unwanted side effects artefacts. The trick in stereo widening is to place these artefacts where they make as little damage as possible.

To illustrate this we can use a simple delay ("Haas delay"). The idea is to take a mono track and send it to the left channel and send a 5-10 ms delayed version of it to the right channel. This will create something that sounds like stereo. It will also sound tilted to the left, since the left channel arrives earlier (artefact #1). It will probably sound quite good on headphones. The problem arrives when we listen back in mono, because summing left and right will sum our original audio with the delayed copy. And these two signals will sound "phasey" or "tinny" together (artefact #2), and we have created what is called a "comb filter". That's the sound you get when you accidentally are listening both through the DAW as well as the direct monitoring on your audio interface and you get a slight delay between the audio.

Can we avoid this? Yes! We can delay to the mono signal but apply it to the left and right completely out of phase, so that the delayed part will cancel out in mono. But that only means that we have moved the comb filter artefact from the center out to the sides, in the left and right channels. (This is the basics of the Classic algorithm.)

So instead of hearing the artefacts when we mono check our mix, we now hear the artefacts in the left and right channels. And that is probably not a problem, as long as you are aware of it and can tune the delay times to minimize the bad effects.

 

Deciding on where to put the artefacts, for example in the mono part, in the left/right channels or in the high or low frequencies, is one thing that differentiate widening algorithms. What types of artefacts we are expecting (comb filter, echoes, chorus, etc) is another differentiating factor.

In Widener, we are using three different techniques in our five algorithms. The first (Classic) is delayed based, as described above. The next two (Frequency Spread and Rotation) are shifting the phase of the audio instead of delaying it. That leads to less comb filtering and no audible echoes but can sound less wide and sometimes sound like the audio has been panned slightly to the left or right. The last two (Ambience and Reflections) are using reverb-like techniques, where we basically add so many echoes so that the comb filtering artefact is masked. On Reflections we are also shuffling the echoes around to further mask the comb filtering, and we are doing that in a way to minimize the audible chorusing effect that can sound “pitchy” or out-of-tune.

The side effects of these different algorithms are also placed differently, the first two doesn’t touch the mid/mono part at all and are completely “mono safe”, but they change the frequency response of the left and right channels. Rotation doesn't change the frequency response of left and right, but adds a slight dip in the mid/mono part. It also doesn't do much for a track that's already stereo. Ambience and Reflections will change both the mid/side and left/right, but tries to do it equally so that there shouldn't be any surprises when you listen back in another environment.

So each algorithm has its advantages and drawbacks, and as long as you know what to listen for you will be able to widen your tracks in a natural way.

Stereo widening is fun!

 

User Interface

On both sides of the main interface you find input and output metering, as well as input and output levels. These are a part of Extended Features, which you can find in all Softube plug-ins. For more information, please visit https://www.softube.com/user-manuals/extended-features

The left side of the main UI consists of the following meters: a goniometer measuring the stereo width of a signal and a correlation meter at the bottom.

To the right you have the five algorithms: Classic, Frequency Spread, Rotation, Ambience and Reflection, as well as their associated parameters. The Amount parameter is the same for all algorithms, and many of the algorithms also share a High and Low Frequency control.

Below the algorithms are the output controls with a Width parameter controlling the gain of the difference signal ("side"), a Mono Bass control and an Output Gain.

Click here to view the full-sized image in a new tab.

 

Algorithms

Each algorithm has its own sound, advantages, but also drawbacks. Understanding the nature of each widening type is important so that you know what to listen for in your context.

Classic

This is a very common Mid/Side technique where the mono signal is delayed and added to the sides. It is completely mono compatible, since it doesn't affect the mid at all, but will color the left and the right independently.

 

Amount: Sets the amount of the mid that gets added to the sides, similar to a Dry/Wet control.

Size: Sets the delay of the added mid. A longer delay will lead to less coloration, but can make it sound like a slap back echo.

 

Low and High Frequency: Sets the low and high cuts for the added mid signal. If you only want to widen the stereo image for high frequencies you can turn up the Low Frequency control.

Frequency Spread

This algorithm is similar to Classic, but instead of delaying the mids with a fixed amount of time, it alters the phase of mids and then adds it to the sides. It is completely mono compatible, since it doesn't affect the mid at all, but will color the left and the right independently.

The Frequency Spread algorithm is often used to enhance mono drum kits or other percussive sources.

 

Amount: Sets the amount of the effect, similar to a Dry/Wet control.

Flip Sides: Swaps left/right of the added effect.

Low and High Frequency: Sets the low and high cuts for the added mid signal as well as changing the lower and upper bounds for the phase shifts.

Rotation

Rotation uses phase rotation on both the left and right channels to create a wide stereo image. It will affect the transient response of the audio, and the mids will have a dip at around the emphasis frequency. The frequency response of the left and right channels is completely unaffected by Rotation.

Rotation will create a very smooth and natural stereo image, almost without side effects. However, since the mono part will get attenuated at the Emphasis frequency, it is important to not set Emphasis at a frequency where important mono elements are, for example the lead vocals of a mix. Be gentle with the amount, settings between 0 and 60% are safe, but above that you risk getting left and right elements out of phase at the Emphasis frequency.

 

Amount: Sets the amount of phase rotation.

Emphasis: Sets the center frequency of the phase rotation.

Flip Sides: Swaps left/right of the added effect.

Ambience

A very short reverb is effective at widening a mono track, and the Ambience algorithm is sort of a reverb. It uses a network of phase rotating filters that are fed into the other channel (from left to right and right to left) while feeding back in a big figure of eight. This algorithm affects both left and right as well as mid and sides.

 

Amount: Sets dry/wet.

Size: Sets the size of the algorithm, in practice it is possible to use it to fine tune the frequency response of the algorithm.

 

Low and High Frequency: Sets the low and high cuts for the added ambience. If you only want to widen the stereo image for high frequencies you can turn up the Low Frequency control.

Reflections

This is another take on "a very short reverb", but instead of focusing on many echoes as in Ambience, this one emulates a small space. In order to avoid "phasey" artifacts like comb filtering it uses modulation to move around the echoes. For some applications you will be able to hear that modulation as a "chorus-y" effect,but for other tracks it's not audible at all. We have specifically chosen a type of modulation that has as little chorusing/pitching effects as possible.

 

Amount: Sets dry/wet.

Size: Sets the size of the algorithm, which in this case corresponds to the size of the room.

Low and High Frequency: Sets the low and high cuts for the added reflections. If you want to soften the sound of the reflections, you can turn down High Frequency.

 

Goniometer

The goniometer displays the relationship between left and right channel audio and is useful to gauge how much stereo information the audio has.

Mono audio will look like a vertical line on the goniometer, while audio that is panned completely to the left will look like a diagonal line pointing to the upper left corner. Similarily, if left and right channels are completely out of phase, this can be seen as a horizontal line on the goniometer.

The goniometer shows the correlation between left and right channel audio. If left and right channels are identical, a vertical line will appear in the middle. If only left channel audio is present, a diagonal line will appear. A typical stereo signal that is more mono than stereo will look like a thick vertical blob, while audio that lacks mono will appear as a horizontal blob.

With a mono input the goniometer will show a vertical line. If we choose a widening algorithm and increase the Amount, we can see how that line expands into a vertical ellipse, because we have added information to the sides. If we want to increase the amount even more, we can turn up Width to 300%, and might end up with a horizontally aligned ellipse. That means that we now have more information in the sides than what we have in the mono channel, and unless you do it on purpose it's typically not a good thing, since left and right are more out of phase than in phase now.

If the blob in the middle is very round in its shape, that means that the left and right channels are completely "uncorrelated", meaning that they have no (mathematical) relationship with each other.

 

Correlation Meter

The correlation meter also measures how similar the left and right channels are, the "correlation" between left and right. This is very similar to the goniometer, but displayed in another way.

If the correlation meter is +1, it means that left and right channels are identical, and the audio is in mono. This corresponds to the goniometer showing a vertical line.

Similarly, if the correlation meter reads -1, it means that left and right channels are completely out of phase, and no mono exist, only sides. This corresponds to the goniometer showing a horizontal line.

If left and right channels are completely different (in a mathematical sense) the correlation meter will show 0.

 

Output Section

These parameters affect the output of the plug-in.

Width: Sets the amount of stereo width on the output by applying gain to the side signal.

 

Mono Bass: Reduces the stereo width for bass signals by applying a high pass filter to the side signal.

Output: Sets the output level.

 

Credits

Cameron Clark – signal processing
Johan Bremin – product design
Thomas Andersson – signal processing
Niklas Odelholm – signal processing, product design, user manual
Anna Pohl Lundgren – UI programming
Arvid Johnsson – UI programming
Arvid Rosén – UI programming
Björn Rödseth – UI programming, signal processing
Tord Bärnfors – UI programming