如何在Python环境中加载Hornet Songkey MK2 VST实现实时调式检测？

Hey Julian, great question! Since you're new to how VST plugins work, let me start with a quick simplified breakdown before diving into practical steps: VSTs are binary modules (like .dll files on Windows, .vst/.vst3 on Mac) that depend on a "host" program (like Reaper or Steinberg) to feed them audio data, manage settings, and retrieve results. To use one in Python, we need a library that acts as this host.

Here's a step-by-step approach to make this work:

1. Install Required Python Libraries #

You'll need two core components: a VST host library to load the plugin, and an audio library to capture real-time input.

• VST Host Library: pyvst is a reliable choice for VST2 plugins (most mainstream plugins use this format; double-check if your Hornet Songkey MK2 is VST2). Install it via pip:

  pip install pyvst
  

  If your plugin is VST3, use pyvst3 instead (note: it’s a bit more experimental):

  pip install pyvst3
  

• Real-Time Audio Capture: sounddevice lets you grab live audio from your microphone or input device. Install it with:

  pip install sounddevice
  

2. Locate Your VST Plugin File #

First, find the actual binary file for Hornet Songkey MK2:

• On Windows: Typically in C:\Program Files\VSTPlugins or the custom folder you chose during installation. Look for HornetSongkeyMK2.dll.
• On Mac: Usually in /Library/Audio/Plug-Ins/VST or /Library/Audio/Plug-Ins/VST3, with a .vst or .vst3 extension.

3. Load the Plugin and Initialize It #

Once you have the path, use the VST library to load and set up the plugin. Here’s an example with pyvst:

import pyvst
import sounddevice as sd

# Replace with your plugin's actual file path
PLUGIN_PATH = "C:/Program Files/VSTPlugins/HornetSongkeyMK2.dll"
SAMPLE_RATE = 44100  # Match your audio device's sample rate
BUFFER_SIZE = 512    # Adjust based on your system's performance

# Load and initialize the plugin
plugin = pyvst.Plugin(PLUGIN_PATH)
plugin.init(sample_rate=SAMPLE_RATE, buffer_size=BUFFER_SIZE)

# Optional: List all plugin parameters to find the key detection result
for i in range(plugin.num_parameters):
    info = plugin.get_parameter_info(i)
    print(f"Parameter {i}: {info.name} (range: {info.min_value} to {info.max_value})")

When you run this, you’ll see a list of parameters. Look for something like "Detected Key" or "Key Result"—this is the parameter you’ll read to get the detected key.

4. Process Real-Time Audio and Retrieve Key Detection #

Use sounddevice to create an audio stream, pass input to the plugin, and read the key parameter on each buffer:

import sys

def audio_callback(indata, outdata, frames, time, status):
    if status:
        print(status, file=sys.stderr)
    
    # Pass input audio to the plugin (VSTs expect stereo audio by default)
    plugin.process(indata)
    
    # Read the detected key parameter (replace 0 with the actual index you found)
    detected_key_value = plugin.get_parameter(0)
    # Map the float value to human-readable keys (adjust based on plugin behavior)
    # Example mapping—you may need to tweak this by testing known keys
    key_map = {
        0.0: "C Major", 0.05: "C Minor",
        0.1: "C# Major", 0.15: "C# Minor",
        # Add more entries as needed based on your plugin's output
    }
    detected_key = key_map.get(round(detected_key_value, 2), "Unknown")
    print(f"Detected Key: {detected_key}")

# Start the audio stream
with sd.Stream(samplerate=SAMPLE_RATE, blocksize=BUFFER_SIZE, callback=audio_callback):
    print("Listening for audio... Press Enter to stop.")
    input()

# Clean up resources
plugin.close()

Important Notes & Troubleshooting #

• VST Version Compatibility: Ensure your library supports the VST version of your plugin (VST2 vs VST3). If pyvst doesn’t work, try vsthost (another Python VST host library).
• Parameter Mapping: The plugin likely encodes key results as a float between 0 and 1. You’ll need to map this to actual key names—either check the plugin’s docs or experiment by playing known keys and recording parameter values.
• Audio Device Setup: Verify your default audio input is correct. Use sd.query_devices() to list devices and sd.default.device to set the right one.
• Alternative: Skip the VST Entirely: If loading the plugin feels too complex, use librosa—a Python audio library with built-in key detection. Install it with pip install librosa, then use:

  import librosa
  import sounddevice as sd
  
  def detect_key(audio_data, sr):
      y = audio_data.T[0]  # Use first channel for mono input
      key = librosa.feature.key_detector(y=y, sr=sr)
      print(f"Detected Key: {key}")
  
  def callback(indata, outdata, frames, time, status):
      detect_key(indata, SAMPLE_RATE)
  
  with sd.Stream(samplerate=SAMPLE_RATE, callback=callback):
      input()
  

  This uses a different algorithm than Hornet Songkey, but it’s a simpler pure-Python solution.

内容的提问来源于stack exchange，提问作者Julian