An explanation of my comment to your question...
In computing, a sink, or data sink generally refers to the destination of data flow.
snip
In several computer programs employing streams, such as GStreamer, PulseAudio, or PipeWire, a source is the starting point of a pipeline which produces a stream but does not consume any, while a sink is the end point which accepts a stream without producing any
https://en.wikipedia.org/wiki/Sink_(computing)
Therefore an audio sink is the destination of audio processing, such as speakers.
Sound card on another hand...
A sound card (also known as an audio card) is an internal expansion card that provides input and output of audio signals to and from a computer under the control of computer programs.
Sound cards use a digital-to-analog converter (DAC), which converts recorded or generated digital signal data into an analog format.
https://en.wikipedia.org/wiki/Sound_card
Sound card therefore is neither audio source nor audio destination (sink)
What sound card does is converting the digital signal.
Audio source is also known as media source ex media file.
Sound processing pipeline consists of media source to media processing software (which uses sound card) toward media sink.
See link below for an overview of Microsoft's implementation of audio processing, it mentions sinks and sources and audio processing:
https://learn.microsoft.com/en-us/windows/win32/medfound/overview-of-the-media-foundation-architecture
edit:
Audio processing on a PC is done by audio processing software and consists of an audio processing pipeline (software thing) that is similar to rendering pipeline in video game development.
In similar fashion how GPU is used for rendering in rendering pipeline (but is not a destination), so is audio card used for audio conversion but is not itself a destination of audio signal.
Sound card accepts input (ex. digital) and provides output (ex. analog) where both input and output is managed by audio processing software.
User interaction with audio software decides what audio sources and sinks (destinations) are used.
For instance if one is to use microphone for audio input, then that's audio source, if a user uses an audio file it's also an audio source.
Audio software will ofc. with the help of drivers interact with devices that are sources or destinations (ex. microphone or speakers)
Audio pipeline (software concept) starts with the source, in the case of microphone audio input is analog signal which the audio software with the help of sound card converts to digital signal in order to be usable by audio software.
Sound card is therefore an intermediate that is used solely for digital-to-analog and vice-versa audio processing which is needed for audio software because audio software deals with digital audio signal but not with analog, to understand analog signal audio software uses sound card for conversion.
Once the signal is converted by sound card the audio software will continue streaming audio signal down the pipeline toward destination aka. audio sink
Audio sink can be a new audio file (ex. if the user is recording voice and saving to file) or it can be speakers (ex. if the user is playing audio)
Audio sink is the end of an audio pipeline where source is the start of the audio pipeline.
Audio pipeline itself is managed by audio software.
Between audio source and audio sink the software is responsible for audio processing which includes using sound card for signal conversion, therefore sound card is neither a source nor destination but an intermediate unit in the audio pipeline that is used by the audio software.
Just like in rendering pipeline a GPU is neither the source nor sink for video game software but is used by the pipeline coded into game, so is sound card used in the audio pipeline by audio software.
edit2:
good to know, but who then plays the audio, i.e. send it to the output connectors? are those connectors the audio devices?
Audio pipeline breakdown:
Example: how using microphone works to listen to ones voice?
- A user speaks to microphone, the input to microphone is analog audio signal
- Audio software (or library) receives analog input signal from driver managing the microphone but does not understand analog signal
- Audio software then coverts analog signal input with the help of sound card driver to digital signal to be able to handle it in audio pipeline (sound card takes input and gives output - conversion)
- Audio software then receives and handles the converted digital signal, it then for ex. applies codecs, does mixing etc. (all with digital signal which the software is able to understand)
- Audio software then again uses sound card to convert digital to analog signal to be able to send it to sink (ex. speakers)
- This analog signal is then sent to driver managing speakers, in order for speaker to play audio the signal must be analog.
Therefore audio software uses sound card twice, first time to convert analog to digital, and then again digital to analog because hardware (microphone and speakers) don't understand digital signal but on another side audio software doesn't analog digital signal so it needs sound card for conversion to be able to handle it and send it to driver responsible for output (or receive for input).
In case of playing an audio file, the signal is already digital therefore sound card would be used only once, that is for digital to analog to be sent to drivers managing speakers. (ex. there is no analog input)
