How to use IAudioEndpoint and connect it to Audio Engine?

mkckr0 0

I want to create a new audio endpoint in user mode. So the audio player can work normally without a physical sound card. I find the API https://learn.microsoft.com/en-us/windows/win32/termserv/terminal-services-audioendpoint-api-reference.

As the doc says:

An audio endpoint represents an audio device, audio API, or any other audio source or sink, and is used to send data to or consume data from the audio engine. An audio endpoint must be connected to the audio engine through a connection, and each connection can have only one endpoint connected to it. After an endpoint is registered, the audio engine attaches the endpoint to the connection.

I don't know how to use the IAudioEndpoint and connect it to audio engine.

Jeanine Zhang-MSFT 10,041 Reputation points Microsoft Vendor

2024-03-19T02:30:40.25+00:00

According to the Doc: User-Mode Audio Components

An audio endpoint device is a software abstraction that represents a physical device that the user manipulates directly.

IAudioEndpoint interface is used to provide information to the audio engine about an audio endpoint.

As far as I'm concerned, the connection here represents a physical connection.

mkckr0 0

I find the AudioEngineEndpoint.idl in Windows 10 SDK has comments about IAudioEndpoint

//-------------------------------------------------------------------------------------
//  An Audio Endpoint object abstracts the audio device, audio API and any other
//  data source/sink from the AudioProcessor. An Audio Endpoint must implement the
//  IAudioEndpoint, IAudioEndpointRT, and one or both of IAudioInputEndpointRT and
//  IAudioOutputEndpointRT interfaces. The clients of the processor must attach at least
//  two Audio Endpoint objects to the Audio Processor, one for input and one for output
//  for useful work to occur. An endpoint may only be connected to one connection, and
//  each connection may only have one endpoint connected to it.
//
//  The audio endpoint object is designed to be a client-extendable object that is used
//  to get data into or out of an audio engine instance. Data transfer is handled by
//  connecting an audio endpoint to an audio connection. See
//  IAudioProcessor::AttachInputEndpointToConnection
//  and IAudioProcessor::AttachOutputEndpointToConnection for more information on attaching
//  endpoints to connections, and the sections on IAudioEndpoint, IAudioEndpointRT,
//  IAudioInputEndpointRT, and IAudioOutputEndpointRT interfaces for info on the audio
//  engine side of the endpoint interfaces.
//
//  The client-facing side of the endpoint objects is unspecified and may be written in any
//  fashion convenient to the needs of the client of the endpoint. The specified interfaces
//  must be implemented as stated above. If the endpoint is not being used to drive a graph,
//  skip to the next paragraph. Otherwise, if the endpoint is being used to drive a processing
//  graph, the processor will call the GetCurrentPadding interface before every processing
//  pass. At that time, the endpoint should return the amount of data that is held in the
//  endpoint. For an output endpoint, that would be the amount of data queued on the device.
//  For an input endpoint, that would be the amount of data captured from the device but not
//  yet delivered to the engine.
//
//  The processor will call either IAudioInputEndpointRT::GetInputDataPointer or
//  IAudioOutputEndpointRT::GetOutputDataPointer (depending on
//  whether the endpoint is input or output to the engine), at which time the endpoint should
//  return a data pointer of the requested size. After processing occurs, the processor will
//  call IAudioInputEndpointRT::ReleaseInputDataPointer or
//  IAudioOutputEndpointRT::ReleaseOutputDataPointer, at which time the endpoint can
//  handle the data appropriately. See the various interface calls given above for more
//  detailed information.
//
//  On the client-facing side of the endpoint object, data just must be capable of being
//  delivered to or gotten from the endpoint as appropriate. If that means that the client
//  wants to use an interface called WriteData to write data on the endpoint, that is fine.
//  The client side could use a callback mechanism, such that there is a method called
//  RegisterCallback(myFuncPtr) that causes myFuncPtr to be called each time
//  GetInputDataPointer is called (however, the myFuncPtr would have to conform to the
//  realtime requirements if that were the case).
//
//  This is what is meant by a client-extendable object: the interface used by the client
//  can be anything, as long as the interfaces required by the engine are present and operate
//  to deliver or consume data to or from the engine as specified in the required interfaces.
//  The "interface" to the client side doesn't even need to be a COM interface: it could be a
//  simple set of C calls or C++ object methods.
//
//  The AudioProcessor uses this interface to get information about the endpoint(s)
//  attached to it.
[
    object,
    uuid(30A99515-1527-4451-AF9F-00C5F0234DAF),
    pointer_default(unique),
    local
]
interface IAudioEndpoint : IUnknown{

But I can't find anything about IAudioProcessor::AttachInputEndpointToConnection and IAudioProcessor::AttachOutputEndpointToConnection.

mkckr0 0 Reputation points

2024-03-19T03:21:54.9633333+00:00

~~Delete repeated comment.~~
mkckr0 0 Reputation points

2024-03-19T03:43:40.19+00:00
As the diagram shown in User-Mode Audio Components,

"Endpoint Buffer" can link the API and driver in exclusive mode, and can also link the API and Audio Engine in shared mode. I think app can create a new endpoint buffer connected with Audio Engine, even when there is no audio device.

I find Remote Desktop create a new audio endpoint in Device Manager, but it doesn't add any new audio adapter. I guess it's a virtual audio endpoint and it use IAudioEndpoint to create the new virtual audio endpoint, so the audio player can work normally.

I also find the sample code of ApplicationLoopback

RETURN_IF_FAILED(ActivateAudioInterfaceAsync(VIRTUAL_AUDIO_DEVICE_PROCESS_LOOPBACK, __uuidof(IAudioClient), &activateParams, this, &asyncOp));

The program can record the sound in PC without any physical speaker. I guess VIRTUAL_AUDIO_DEVICE_PROCESS_LOOPBACK is a predefined virtual audio endpoint and there is nothing about driver.
Jeanine Zhang-MSFT 10,041 Reputation points Microsoft Vendor

2024-03-19T07:39:02.03+00:00

If a client of WASAPI opens a stream in shared mode, then the client writes audio data to the endpoint buffer and the Windows audio engine reads the data from the buffer. In this mode, the client shares the audio hardware with other applications running in other processes.

Virtual audio devices represent the filter graphs that render and capture audio content. As far as I'm concerned, Virtual audio devices need driver.

Reffer to the Doc:Loopback Recording

The implementation of loopback by WASAPI depend on the capabilities of the hardware.

Could you please tell us, what you want to achieve?
mkckr0 0 Reputation points

2024-03-19T09:53:04.32+00:00

Yes, The Loopback Recording by using IMMDevice interface requires a physical audio device. But ActivateAudioInterfaceAsync doesn't need.

As for what I want to achieve, I have said at the beginning:

I want to create a new audio endpoint in user mode. So the audio player can work normally without a physical sound card.

To achieve a virtual audio device in Windows needs to program a driver and use EV code signing certificate to publish it. Sample Audio Drivers SysVAD Virtual Audio Device Driver Sample It's very complicated and have to pay for the EV code signing certificate. I want to find a easier way.

In Linux, developer can create a audio virtual endpoint by using audio server without driver and even root privilege. I think Windows may have simlilar things. The Audio Engine is like a audio server and it can take over all things in user space and provide interface to create new audio endpoint.
Jeanine Zhang-MSFT 10,041 Reputation points Microsoft Vendor

2024-03-20T05:45:02.09+00:00

What do you mean "audio server"? Whether the "audio server" you are talking about is a "audio service"?

According to the Doc: Windows Audio Architecture

The audio service: Is used to setup and control audio streams. Implements Windows policies for background audio playback, ducking, etc.

As far as I'm concerned, in windows, you couldn't create an audio virtual endpoint by using audio service. If you want to create an audio virtual endpoint in windows, you need to use WDK to create windows device drivers.
mkckr0 0 Reputation points

2024-03-20T06:39:08.9133333+00:00

You just tell me how to use IAudioEndpoint. If it can not be used, I'll consider to create a audio driver.
Jeanine Zhang-MSFT 10,041 Reputation points Microsoft Vendor

2024-03-20T07:39:30.1366667+00:00

Refer to the Doc:IAudioEndpoint interface

Provides information to the audio engine about an audio endpoint. This interface is implemented by an audio endpoint.

IAudioEndpoint interface is not used to create an audio endpoint.

1 answer

mkckr0 0 Reputation points

2024-05-07T13:39:57.14+00:00

A client of WASAPI passes data to an endpoint device through an endpoint buffer. System software and hardware components manage the movement of data from the endpoint buffer to the endpoint device in a manner that is largely transparent to the client. Furthermore, for an endpoint device that plugs into an audio adapter with jack-presence detection, the client can create an endpoint buffer only for an endpoint device that is physically present. For more information about jack-presence detection, see Audio Endpoint Devices.

The client can create an endpoint buffer only for an endpoint device that is physically present.
Please sign in to rate this answer.

0 comments No comments
Sign in to comment

Use comments to ask for clarification, additional information, or improvements to the question.

Share via

How to use IAudioEndpoint and connect it to Audio Engine?

1 answer

Your answer