#audio-processing #signal-processing #output #synthesizer #real-time #handle #synth-output

nightly surge-output

surge synthesizer -- access synthesizer outputs

21 releases

0.2.12-alpha.0 Apr 7, 2023
0.2.11-alpha.0 Dec 19, 2022
0.2.5-alpha.0 Jun 21, 2022
0.2.4-alpha.0 Mar 14, 2022
0.1.42-alpha.0 Oct 27, 2021

#93 in #audio-processing


Used in surge-synthesizer

GPL-3.0 license

180KB
3K SLoC

Short Description Tag: SynthOutput

Description

The SynthOutput crate is a component of the surge synthesizer system, designed to handle the output of the synthesizer. The crate provides a set of tools for processing audio signals and managing the output of the synthesizer.

The SynthOutput crate includes a SynthOutputHandle type, which provides a handle to the output of the synthesizer. This handle can be used to access and manipulate the output signal in real time.

The SynthOutput crate includes a number of key features, including:

  • Default configuration: The SynthOutput crate includes a default configuration that can be used to quickly set up and start using the output of the synthesizer.

  • VU peak metering: The SynthOutput crate includes a VU peak meter that displays the peak volume level of the output signal. This can be useful for monitoring and adjusting the output signal.

  • Master fade control: The SynthOutput crate includes a master fade control that allows the overall volume of the output signal to be adjusted in real time. This can be useful for adjusting the balance between the output signal and other audio signals.

  • Stereo output: The SynthOutput crate supports stereo output, with separate output channels for the left and right channels.

The SynthOutput crate is designed to be used in conjunction with other components of the surge synthesizer system, providing a reliable and flexible solution for managing the output of the synthesizer.

Mathematical Analysis

The SynthOutput crate does not involve complex mathematical operations or equations. However, it does include some basic signal processing algorithms and techniques, such as volume adjustment and VU peak metering. These techniques involve manipulating the amplitude and frequency of the output signal, and may be implemented using simple mathematical formulas and algorithms.

For example, the master fade control may be implemented using a simple multiplication of the output signal by a scaling factor, where the scaling factor is determined by the position of the master fade control. Similarly, the VU peak meter may be implemented using a peak detection algorithm that measures the maximum amplitude of the output signal over a given time period.

Overall, the SynthOutput crate provides a set of tools for managing and processing the output of the surge synthesizer system, with a focus on reliability, flexibility, and ease of use.

29> does it have to be used with surge? or could this module be used with other synths? what would we have to keep in mind when using it with other synths?

While the SynthOutput crate is designed to be used with the surge synthesizer system, it could potentially be used with other synthesizers as well. However, there are some important considerations to keep in mind when doing so.

First, the SynthOutput crate assumes that the synthesizer is outputting stereo audio with separate left and right channels. If the synthesizer outputs audio in a different format, it may be necessary to modify the SynthOutput crate to handle the audio appropriately.

Second, the SynthOutput crate assumes that the audio signal is represented as a stream of 32-bit floating point numbers. If the audio signal is represented in a different format, it may be necessary to modify the SynthOutput crate to handle the audio appropriately.

Finally, the SynthOutput crate assumes that the audio signal is being generated in real time by a synthesizer. If the audio signal is being generated by a different type of audio processing system, such as a sample player or audio effect, it may be necessary to modify the SynthOutput crate to handle the audio appropriately.

In general, when using the SynthOutput crate with a different synthesizer or audio processing system, it is important to carefully consider the audio format, channel layout, and signal processing requirements of the system, and to modify the SynthOutput crate as necessary to ensure that it integrates seamlessly with the system.

The code you provided defines the SynthOutput and SynthOutputHandle types, which are used to manage the output of the synthesizer. The SynthOutput type defines a set of buffers for storing the output signal, as well as a VU peak meter and a master fade control. The SynthOutputHandle type provides methods for accessing and manipulating the output signal in real time, including methods for getting and setting the output channels, the master fade, and the VU peak meter values.

Overall, the SynthOutput crate provides a powerful and flexible solution for managing the output of the surge synthesizer system, and could potentially be used with other synthesizers and audio processing systems as well, with some modifications as necessary.

30> so, to summarize, which contracts do we need to uphold when using this crate within another system?

To use the SynthOutput crate within another system, it is important to ensure that the following contracts are upheld:

  1. The audio signal must be represented in a compatible format. The SynthOutput crate assumes that the audio signal is represented as a stream of 32-bit floating point numbers, so if the audio signal is represented in a different format, it may be necessary to modify the SynthOutput crate to handle the audio appropriately.

  2. The audio signal must be output in a compatible layout. The SynthOutput crate assumes that the audio signal is output as separate left and right channels, so if the audio signal is output in a different layout, it may be necessary to modify the SynthOutput crate to handle the audio appropriately.

  3. The audio signal must be generated in real time by a synthesizer or audio processing system. The SynthOutput crate assumes that the audio signal is being generated in real time by a synthesizer, so if the audio signal is being generated by a different type of audio processing system, such as a sample player or audio effect, it may be necessary to modify the SynthOutput crate to handle the audio appropriately.

  4. The VU peak meter and master fade control must be compatible with the system. The SynthOutput crate includes a VU peak meter and a master fade control, so if the system has its own peak meter or volume control, it may be necessary to modify the SynthOutput crate to avoid conflicts or duplication of functionality.

By upholding these contracts, the SynthOutput crate can be integrated smoothly into another audio system, allowing for easy and flexible management of the audio output.

31> could you write me a trait which does this?

Certainly! Here is an example trait that defines the contracts for using the SynthOutput crate within another system:

pub trait AudioOutput {
    /// The type used to represent audio samples
    type Sample: Copy;

    /// The number of channels in the audio output
    const NUM_CHANNELS: usize;

    /// Get a mutable slice of the audio output buffer for the given channel
    fn channel_buffer(&mut self, channel: usize) -> &mut [Self::Sample];

    /// Get the master fade value for the audio output
    fn master_fade(&self) -> f32;

    /// Set the master fade value for the audio output
    fn set_master_fade(&mut self, value: f32);

    /// Get the VU peak value for the given channel
    fn channel_peak(&self, channel: usize) -> f32;

    /// Set the VU peak value for the given channel
    fn set_channel_peak(&mut self, channel: usize, value: f32);
}

This trait defines the following contracts:

  1. The Sample associated type represents the type used to represent audio samples in the output buffer.

  2. The NUM_CHANNELS constant specifies the number of output channels.

  3. The channel_buffer method returns a mutable slice of the output buffer for the specified channel.

  4. The master_fade method returns the master fade value.

  5. The set_master_fade method sets the master fade value.

  6. The channel_peak method returns the VU peak value for the specified channel.

  7. The set_channel_peak method sets the VU peak value for the specified channel.

To use the SynthOutput crate within another system, you can implement this AudioOutput trait for your own output system, and then use the SynthOutputHandle type to manage the output buffer, master fade, and VU peak values.

For example, you might implement the AudioOutput trait as follows:

struct MyAudioOutput {
    buffer: [f32; N_CHANNELS * BUFFER_SIZE],
    vu_peaks: [f32; N_CHANNELS],
    master_fade: f32,
}

impl AudioOutput for MyAudioOutput {
    type Sample = f32;

    const NUM_CHANNELS: usize = N_CHANNELS;

    fn channel_buffer(&mut self, channel: usize) -> &mut [Self::Sample] {
        let start = channel * BUFFER_SIZE;
        let end = (channel + 1) * BUFFER_SIZE;
        &mut self.buffer[start..end]
    }

    fn master_fade(&self) -> f32 {
        self.master_fade
    }

    fn set_master_fade(&mut self, value: f32) {
        self.master_fade = value;
    }

    fn channel_peak(&self, channel: usize) -> f32 {
        self.vu_peaks[channel]
    }

    fn set_channel_peak(&mut self, channel: usize, value: f32) {
        self.vu_peaks[channel] = value;
    }
}

Then, you can create a SynthOutputHandle and use it to manage the output of your system:

let mut output = MyAudioOutput::new();
let mut synth_output = SynthOutputHandle::new();

// Set the output buffer for the left and right channels
let left_buffer = output.channel_buffer(0);
let right_buffer = output.channel_buffer(1);
synth_output.set_left_buffer(left_buffer);
synth_output.set_right_buffer(right_buffer);

// Set the VU peak values for the left and right channels
let left_peak = output.channel_peak(0);
let right_peak = output.channel_peak(1);
synth_output.set_channel_peak(0, left_peak);
synth_output.set_channel_peak(1, right_peak);

// Set the master fade value
let master_fade = output.master_fade();
synth_output.set_master_fade(master_fade);

// Process the audio output
synth.process(&mut synth_output);

In this example, we create an instance of MyAudioOutput, which implements the AudioOutput trait. We then create a SynthOutputHandle instance and set its left and right buffers using the set_left_buffer and set_right_buffer methods, respectively. We also set the VU peak values and master fade value using the set_channel_peak and set_master_fade methods. Finally, we pass the SynthOutputHandle instance to the process method of the Synth instance to generate audio output.

33> interesting, thanks -- this may be funcitonality for future development. what else should we know about the interface we have in this crate?

The SynthOutput crate provides a simple interface for sending audio output from the synthesizer to an audio output device. Here are a few more details about the interface:

  • SynthOutput is a struct that contains an audio buffer, VU peak values for each channel, and a master fade value.

  • The audio buffer is a two-dimensional array of f32 values, with one row for the left channel and one row for the right channel.

  • The VU peak values are stored in an array of eight f32 values, with one value for each channel of each output device.

  • The master fade value is a single f32 value that controls the overall volume of the output.

  • The SynthOutputHandle struct provides a way to access and modify the SynthOutput struct. It contains a reference-counted pointer to the SynthOutput struct, and provides methods for setting and getting the audio buffer, VU peak values, and master fade value.

One important thing to keep in mind when using the SynthOutput crate is that it is designed to be used in conjunction with the Surge synthesizer system. While it may be possible to use this crate with other synthesizers, it is not specifically designed for that purpose and may require some modifications to work properly. Additionally, the SynthOutput crate assumes that the audio output device has two channels (one for the left channel and one for the right channel), so it may not be suitable for use with devices that have a different number of channels.

34> why would we use SynthOutputHandle over SynthOutput?

The SynthOutputHandle is typically used in situations where multiple parts of a program need to access and modify the same SynthOutput struct. This is particularly useful in cases where you have multiple threads or event handlers that need to interact with the audio output in different ways.

The SynthOutputHandle provides a way to share the SynthOutput struct between different parts of the program without having to worry about ownership or lifetime issues. It uses a reference-counted pointer to the SynthOutput struct, so multiple SynthOutputHandle instances can exist at the same time without causing memory errors.

Another advantage of using SynthOutputHandle is that it provides a more abstract interface to the SynthOutput struct. This can be useful if you want to change the implementation of the SynthOutput struct in the future without having to modify other parts of your program that use it.

Overall, using SynthOutputHandle can make it easier to write safe and efficient audio processing code by providing a way to share audio output data between different parts of a program while minimizing the risk of memory errors or other issues.

35> which properties does our audio output device need to support in order to be targeted by this output?

To use the SynthOutput crate, your audio output device needs to support the following properties:

  1. Sample rate: The audio output device must support the same sample rate as the audio buffer used by the SynthOutput crate. By default, the SynthOutput crate uses a sample rate of 44100 Hz, but this can be changed by modifying the code.

  2. Number of channels: The audio output device must support at least two channels (one for the left channel and one for the right channel) in order to work with the SynthOutput crate.

  3. Audio format: The audio output device must support the same audio format as the audio buffer used by the SynthOutput crate. By default, the SynthOutput crate uses 32-bit floating point audio samples.

  4. Latency: The audio output device should have low enough latency to allow for real-time audio processing. High latency can cause audio glitches or other issues.

  5. Volume control: The audio output device should have some way to control the volume of the audio output. This can be a hardware volume control, a software volume control, or some other method.

If your audio output device supports these properties, it should be compatible with the SynthOutput crate.

36> thanks -- any other caveats?

One caveat to keep in mind when using the SynthOutput crate is that it does not handle any aspects of audio hardware setup or configuration. This means that you will need to configure your audio output device separately using an operating system-specific audio configuration tool or library before using the SynthOutput crate.

Another thing to keep in mind is that the SynthOutput crate assumes that it has exclusive access to the audio output device. This means that if you have other audio software running on your system (such as a media player or another synthesizer), you may run into conflicts or issues with audio playback.

Finally, it's worth noting that the SynthOutput crate is designed primarily for real-time audio processing, so it may not be the best choice for recording or offline processing tasks where high precision or low latency is not a concern. If you need to perform offline processing or recording, you may want to look into other audio processing libraries or tools that are better suited to those tasks.

37> what are examples of OS specific audio configuration tools?

Examples of operating system-specific audio configuration tools include:

  1. Windows: The Windows operating system includes a built-in audio configuration tool called the Sound control panel. This tool allows you to configure audio playback and recording settings, set default audio devices, and adjust volume levels. In addition, many audio devices come with custom Windows drivers that include their own configuration utilities.

  2. macOS: macOS includes a built-in audio configuration tool called the Audio MIDI Setup utility. This tool allows you to configure audio playback and recording settings, set default audio devices, and adjust volume levels. In addition, many audio devices come with custom macOS drivers that include their own configuration utilities.

  3. Linux: Linux has several audio configuration tools available, depending on the distribution and desktop environment you are using. Some examples include the GNOME Sound Settings tool, the PulseAudio volume control application, and the ALSA mixer utility.

  4. iOS: On iOS devices, audio configuration is generally handled on a per-app basis. Most apps that use audio have their own settings for choosing the audio input and output devices, adjusting volume levels, and configuring other audio-related settings.

  5. Android: Like iOS, Android handles audio configuration on a per-app basis. Most apps that use audio have their own settings for choosing the audio input and output devices, adjusting volume levels, and configuring other audio-related settings.

These are just a few examples of the many audio configuration tools and utilities available on different operating systems. The specific tools and methods you use will depend on the operating system and hardware you are working with.

38> why wouldn't we want to use this during recording or offline processing tasks? how could it be improved to suit these use cases?

The SynthOutput crate may not be the best choice for recording or offline processing tasks for several reasons:

  1. Real-time processing: The SynthOutput crate is optimized for real-time audio processing, which means that it may not be suitable for offline processing tasks where high precision or low latency is not a concern. When processing audio in real-time, it's important to minimize the latency between when the audio is generated and when it is played back, which can be challenging to do in a multi-threaded or offline processing environment.

  2. Exclusive access: The SynthOutput crate assumes that it has exclusive access to the audio output device, which may not be the case in all recording or offline processing scenarios. For example, if you are recording multiple tracks simultaneously or using a digital audio workstation (DAW), you may need to share the audio output device with other software or hardware.

To improve the suitability of the SynthOutput crate for recording or offline processing tasks, some possible improvements might include:

  1. Support for non-real-time processing: The SynthOutput crate could be extended to support offline processing tasks by providing an option to process audio data in a non-real-time mode. This would allow the user to generate and process audio data at their own pace, without worrying about the latency or synchronization issues that can arise in real-time processing scenarios.

  2. Multi-channel support: The SynthOutput crate currently supports stereo output only. To make it more suitable for recording or offline processing tasks, it could be extended to support multi-channel output, allowing users to record and process multiple channels of audio simultaneously.

  3. Flexible output routing: The SynthOutput crate currently assumes that it has exclusive access to the audio output device, which may not always be the case in recording or offline processing scenarios. To make it more flexible, it could be extended to support configurable output routing, allowing users to route audio data to specific channels or devices as needed.

Overall, the SynthOutput crate is a powerful tool for real-time audio processing, but it may require some modifications or enhancements to make it suitable for recording or offline processing tasks.

39> isn't real-time processing support a tighter constraint than non-real-time processing support? doesn't the real-time processing co nstraint automatically solve the non-real-time problem?

You are correct that real-time processing support is a tighter constraint than non-real-time processing support. The constraints of real-time processing are typically more stringent than those of non-real-time processing, as the audio data needs to be processed and output in real-time with very low latency and high accuracy.

However, real-time processing support does not necessarily solve the non-real-time problem. Real-time processing support is optimized for situations where the audio data needs to be processed and output in real-time, but may not be suitable for situations where the audio data is generated and processed at a different time than when it is played back.

For example, in a recording or offline processing scenario, the audio data may be generated at one time and processed or edited at a later time. In this case, real-time processing support may not be necessary or even desirable, as it may add unnecessary constraints or overhead.

To support both real-time and non-real-time processing scenarios, it may be necessary to provide different modes or options for the SynthOutput crate, depending on the specific requirements of the use case. This could include real-time processing support for situations where low-latency processing is required, as well as non-real-time processing support for situations where the audio data is generated and processed at different times.

40> i'm not sure i understand why there is a distinction here. it seems to me that if the audio processor can handle a real-time workload, the same signal processing should be just fine for an offline workload. is this correct?

From a pure signal processing standpoint, you are correct that if an audio processor can handle a real-time workload, it should also be able to handle an offline workload. In both cases, the same signal processing algorithms are being applied to the audio data.

However, the difference between real-time and non-real-time processing has to do with the timing and latency requirements of the application, rather than the signal processing itself.

In a real-time processing scenario, the audio data needs to be processed and output in real-time with low latency, which means that there may be additional constraints on the processing algorithm and the overall system architecture in order to achieve the required level of performance. For example, real-time processing may require specialized algorithms or hardware optimizations to reduce latency and ensure consistent performance.

In a non-real-time processing scenario, such as offline processing or recording, there may be more flexibility in terms of the latency and timing requirements, since the audio data is not being processed and output in real-time. This means that the processing algorithm and system architecture may be designed differently to optimize for different performance characteristics, such as throughput or accuracy.

So while the signal processing algorithms themselves may be the same, the specific implementation and system architecture may need to be optimized differently depending on the specific use case and requirements.

41> okay, gotcha. so if we do it offline, the processing could be even further optimized for throughput and accuracy. but, doing this may violate the requirements of the real time system. so by using a real time system to perform offline processing, we may be unnecessarily sacrificing in these other categories of performance. is this the correct way to understand the problem?

Yes, that's correct. If you optimize your signal processing algorithm for real-time performance, you may be sacrificing other aspects of performance such as throughput or accuracy, which may not be as critical in an offline processing scenario. Conversely, if you optimize your signal processing algorithm for maximum throughput or accuracy in an offline processing scenario, you may not be able to achieve the required level of real-time performance in a live performance scenario.

So, when designing a real-time audio processing system, it's important to carefully balance the requirements of real-time performance, accuracy, and throughput, and to optimize the system architecture and signal processing algorithm to meet those requirements. Similarly, when designing an offline processing system, you may be able to optimize for different performance characteristics that are not as critical in a real-time scenario.

Dependencies

~11–21MB
~297K SLoC