microphone recording working with FMOD!!!

This commit is contained in:
2026-01-14 18:02:41 +02:00
parent 96839e0e82
commit cb4d4036e7
12 changed files with 410 additions and 140 deletions

Binary file not shown.

View File

@@ -13,7 +13,6 @@ GameObject:
- component: {fileID: 7203143526414218131}
- component: {fileID: 2568780590385977406}
- component: {fileID: 1523408157143897080}
- component: {fileID: 2918895218826581840}
m_Layer: 0
m_Name: MicrophoneStand 1
m_TagString: Untagged
@@ -35,7 +34,7 @@ Transform:
m_Children:
- {fileID: 1850758373829337931}
m_Father: {fileID: 0}
m_RootOrder: 0
m_RootOrder: -1
m_LocalEulerAnglesHint: {x: 0, y: 0, z: 0}
--- !u!33 &4763701867718457502
MeshFilter:
@@ -120,41 +119,10 @@ MonoBehaviour:
m_Script: {fileID: 11500000, guid: 357101b2cface4943b04dfd25d4944e3, type: 3}
m_Name:
m_EditorClassIdentifier:
whisper: {fileID: 0}
microphoneRecord: {fileID: 2918895218826581840}
microphoneDevice:
outputText: {fileID: 0}
microphoneOffStatus: {fileID: 8977839985090371394}
microphoneOnStatus: {fileID: 6537061652288108950}
--- !u!114 &2918895218826581840
MonoBehaviour:
m_ObjectHideFlags: 0
m_CorrespondingSourceObject: {fileID: 0}
m_PrefabInstance: {fileID: 0}
m_PrefabAsset: {fileID: 0}
m_GameObject: {fileID: 4291579148315658230}
m_Enabled: 1
m_EditorHideFlags: 0
m_Script: {fileID: 11500000, guid: 3bc03a4c19604ea394e364f8fc632928, type: 3}
m_Name:
m_EditorClassIdentifier:
maxLengthSec: 60
loop: 0
frequency: 16000
chunksLengthSec: 0.5
echo: 1
useVad: 1
vadUpdateRateSec: 0.1
vadContextSec: 30
vadLastSec: 1.25
vadThd: 1
vadFreqThd: 100
vadIndicatorImage: {fileID: 0}
vadStop: 0
dropVadPart: 1
vadStopTime: 3
microphoneDropdown: {fileID: 0}
microphoneDefaultLabel: Default microphone
fmodWhisperBridge: {fileID: 0}
--- !u!1 &5819798980962142350
GameObject:
m_ObjectHideFlags: 0

View File

@@ -13,8 +13,7 @@ GameObject:
- component: {fileID: 3363453355800186393}
- component: {fileID: 2166102850012183631}
- component: {fileID: 8356602476881383464}
- component: {fileID: 4280129837482332964}
- component: {fileID: 8105025160921831064}
- component: {fileID: 5682436338090300270}
m_Layer: 0
m_Name: ModelGenerationManager
m_TagString: Untagged
@@ -64,6 +63,8 @@ MonoBehaviour:
INVOKEAI_BASE_URL: http://ltat-cgvr9.domenis.ut.ee:9090
DEFAULT_QUEUE_ID: default
MODEL_KEY: 81d45960-08a0-4b8c-a48b-e7d73b21bfe2
promptSuffix: ', single object, front and side fully visible, realistic style,
plain neutral background, clear details, soft studio lighting, true-to-scale'
--- !u!114 &2166102850012183631
MonoBehaviour:
m_ObjectHideFlags: 0
@@ -110,7 +111,7 @@ MonoBehaviour:
useVad: 1
tokensTimestamps: 0
audioCtx: 0
--- !u!114 &4280129837482332964
--- !u!114 &5682436338090300270
MonoBehaviour:
m_ObjectHideFlags: 0
m_CorrespondingSourceObject: {fileID: 0}
@@ -119,65 +120,14 @@ MonoBehaviour:
m_GameObject: {fileID: 154411548685861447}
m_Enabled: 1
m_EditorHideFlags: 0
m_Script: {fileID: 11500000, guid: a2836e36774ca1c4bbbee976e17b649c, type: 3}
m_Script: {fileID: 11500000, guid: a264cb5321d9f6741a2625b8c63c34a7, type: 3}
m_Name:
m_EditorClassIdentifier:
_componentIndexCache: 0
_addedNetworkObject: {fileID: 8105025160921831064}
_networkObjectCache: {fileID: 8105025160921831064}
_synchronizeParent: 0
_packing:
Position: 1
Rotation: 1
Scale: 0
_interpolation: 2
_extrapolation: 2
_enableTeleport: 0
_teleportThreshold: 1
_clientAuthoritative: 1
_sendToOwner: 1
_synchronizePosition: 1
_positionSnapping:
X: 0
Y: 0
Z: 0
_synchronizeRotation: 1
_rotationSnapping:
X: 0
Y: 0
Z: 0
_synchronizeScale: 1
_scaleSnapping:
X: 0
Y: 0
Z: 0
--- !u!114 &8105025160921831064
MonoBehaviour:
m_ObjectHideFlags: 0
m_CorrespondingSourceObject: {fileID: 0}
m_PrefabInstance: {fileID: 0}
m_PrefabAsset: {fileID: 0}
m_GameObject: {fileID: 154411548685861447}
m_Enabled: 1
m_EditorHideFlags: 0
m_Script: {fileID: 11500000, guid: 26b716c41e9b56b4baafaf13a523ba2e, type: 3}
m_Name:
m_EditorClassIdentifier:
<IsNested>k__BackingField: 0
<ComponentIndex>k__BackingField: 0
<PredictedSpawn>k__BackingField: {fileID: 0}
_networkBehaviours:
- {fileID: 4280129837482332964}
<ParentNetworkObject>k__BackingField: {fileID: 0}
<ChildNetworkObjects>k__BackingField: []
_isNetworked: 1
_isGlobal: 0
_initializeOrder: 0
_defaultDespawnType: 0
NetworkObserver: {fileID: 0}
<PrefabId>k__BackingField: 0
<SpawnableCollectionId>k__BackingField: 0
_scenePathHash: 0
<SceneId>k__BackingField: 0
<AssetPathHash>k__BackingField: 17646158845367820466
_sceneNetworkObjects: []
whisper: {fileID: 8356602476881383464}
useVadInStream: 0
recordDriverId: 0
desiredSampleRate: 48000
channels: 1
bufferLengthSec: 5
playLoopback: 0
loopbackVolume: 1

View File

@@ -1,36 +1,19 @@
using TMPro;
using UnityEngine;
using Whisper;
using Whisper.Utils;
public class MicrophoneStand : MonoBehaviour
{
public WhisperManager whisper;
public MicrophoneRecord microphoneRecord;
public string microphoneDevice;
public TextMeshProUGUI outputText;
private WhisperStream stream;
[SerializeField]
private string textOutput;
public GameObject microphoneOffStatus;
public GameObject microphoneOnStatus;
public FMODWhisperBridge fmodWhisperBridge;
// Start is called before the first frame update
async void Start()
void Start()
{
Debug.Log("Mic devices: " + string.Join(", ", Microphone.devices));
Debug.Log("Using mic device: " + microphoneDevice);
microphoneRecord.SelectedMicDevice = microphoneDevice;
// This causes about 1 sec long freeze, has to be done once at the start of the game
microphoneRecord.StartRecord();
stream = await whisper.CreateStream(microphoneRecord);
stream.OnResultUpdated += OnWhisperResult;
//stream.StartStream();
fmodWhisperBridge.OnWhisperResultProcessed += OnWhisperResult;
microphoneOffStatus.SetActive(true);
microphoneOnStatus.SetActive(false);
@@ -51,10 +34,7 @@ public class MicrophoneStand : MonoBehaviour
{
microphoneOffStatus.SetActive(false);
microphoneOnStatus.SetActive(true);
stream.StartStream();
//microphoneRecord.StartRecord();
Debug.Log("Whisper stream started.");
fmodWhisperBridge.ActivateRecording();
}
}
@@ -65,27 +45,17 @@ public class MicrophoneStand : MonoBehaviour
{
microphoneOffStatus.SetActive(true);
microphoneOnStatus.SetActive(false);
stream.StopStream();
//microphoneRecord.StopRecord();
textOutput = outputText.text;
fmodWhisperBridge.DeactivateRecording();
}
}
private void OnWhisperResult(string result)
{
Debug.Log("Whisper result processed: " + result);
outputText.text = result;
}
private void OnDestroy()
{
microphoneRecord.StopRecord();
Destroy(gameObject);
}
public string GetTextOutput()
{
return textOutput;
return outputText.text;
}
}

View File

@@ -0,0 +1,122 @@
using FMOD;
using FMODUnity;
using System.Runtime.InteropServices;
using UnityEngine;
public class FMODMicLoopback : MonoBehaviour
{
private uint LATENCY_MS = 50;
private uint DRIFT_MS = 1;
private uint samplesRecorded, samplesPlayed = 0;
private int nativeRate, nativeChannels = 0;
private uint recSoundLength = 0;
uint lastPlayPos = 0;
uint lastRecordPos = 0;
private uint driftThreshold = 0;
private uint desiredLatency = 0;
private uint adjustLatency = 0;
private int actualLatency = 0;
uint minRecordDelta = 0xFFFFFFFF;
private FMOD.CREATESOUNDEXINFO exInfo = new FMOD.CREATESOUNDEXINFO();
private FMOD.Sound recSound;
private FMOD.Channel channel;
// Start is called before the first frame update
void Start()
{
/*
Determine latency in samples.
*/
FMODUnity.RuntimeManager.CoreSystem.getRecordDriverInfo(0, out _, 0, out _, out nativeRate, out _, out nativeChannels, out _);
driftThreshold = (uint)(nativeRate * DRIFT_MS) / 1000;
desiredLatency = (uint)(nativeRate * LATENCY_MS) / 1000;
adjustLatency = desiredLatency;
actualLatency = (int)desiredLatency;
/*
Create user sound to record into, then start recording.
*/
exInfo.cbsize = Marshal.SizeOf(typeof(FMOD.CREATESOUNDEXINFO));
exInfo.numchannels = nativeChannels;
exInfo.format = FMOD.SOUND_FORMAT.PCM16;
exInfo.defaultfrequency = nativeRate;
exInfo.length = (uint)(nativeRate * sizeof(short) * nativeChannels);
FMODUnity.RuntimeManager.CoreSystem.createSound("", FMOD.MODE.LOOP_NORMAL | FMOD.MODE.OPENUSER, ref exInfo, out recSound);
FMODUnity.RuntimeManager.CoreSystem.recordStart(0, recSound, true);
recSound.getLength(out recSoundLength, FMOD.TIMEUNIT.PCM);
}
// Update is called once per frame
void Update()
{
/*
Determine how much has been recorded since we last checked
*/
uint recordPos = 0;
FMODUnity.RuntimeManager.CoreSystem.getRecordPosition(0, out recordPos);
uint recordDelta = (recordPos >= lastRecordPos) ? (recordPos - lastRecordPos) : (recordPos + recSoundLength - lastRecordPos);
lastRecordPos = recordPos;
samplesRecorded += recordDelta;
if (recordDelta != 0 && (recordDelta < minRecordDelta))
{
minRecordDelta = recordDelta; // Smallest driver granularity seen so far
adjustLatency = (recordDelta <= desiredLatency) ? desiredLatency : recordDelta; // Adjust our latency if driver granularity is high
}
/*
Delay playback until our desired latency is reached.
*/
if (!channel.hasHandle() && samplesRecorded >= adjustLatency)
{
FMODUnity.RuntimeManager.CoreSystem.getMasterChannelGroup(out FMOD.ChannelGroup mCG);
FMODUnity.RuntimeManager.CoreSystem.playSound(recSound, mCG, false, out channel);
}
/*
Determine how much has been played since we last checked.
*/
if (channel.hasHandle())
{
uint playPos = 0;
channel.getPosition(out playPos, FMOD.TIMEUNIT.PCM);
uint playDelta = (playPos >= lastPlayPos) ? (playPos - lastPlayPos) : (playPos + recSoundLength - lastPlayPos);
lastPlayPos = playPos;
samplesPlayed += playDelta;
// Compensate for any drift.
int latency = (int)(samplesRecorded - samplesPlayed);
actualLatency = (int)((0.97f * actualLatency) + (0.03f * latency));
int playbackRate = nativeRate;
if (actualLatency < (int)(adjustLatency - driftThreshold))
{
// Playback position is catching up to the record position, slow playback down by 2%
playbackRate = nativeRate - (nativeRate / 50);
}
else if (actualLatency > (int)(adjustLatency + driftThreshold))
{
// Playback is falling behind the record position, speed playback up by 2%
playbackRate = nativeRate + (nativeRate / 50);
}
channel.setFrequency((float)playbackRate);
}
}
private void OnDestroy()
{
recSound.release();
}
}

View File

@@ -0,0 +1,11 @@
fileFormatVersion: 2
guid: 12623af3ff8b7df4abbaacd753e43555
MonoImporter:
externalObjects: {}
serializedVersion: 2
defaultReferences: []
executionOrder: 0
icon: {instanceID: 0}
userData:
assetBundleName:
assetBundleVariant:

View File

@@ -0,0 +1,238 @@

using System;
using System.Runtime.InteropServices;
using UnityEngine;
using FMOD;
using FMODUnity;
using Whisper; // WhisperManager, WhisperStream, WhisperResult
using Whisper.Utils; // AudioChunk
/// <summary>
/// Capture microphone with FMOD and feed chunks to WhisperStream (no Unity Microphone).
/// Also (optionally) plays the recorded sound back via FMOD loopback.
/// </summary>
public class FMODWhisperBridge : MonoBehaviour
{
[Header("Whisper")]
[SerializeField] private WhisperManager whisper; // assign in Inspector
[SerializeField] private bool useVadInStream = false; // let WhisperStream do VAD or not
[Header("FMOD capture")]
[Tooltip("Recording device index (0 = default)")]
public int recordDriverId = 0;
[Tooltip("Set 48000 on Quest; falls back to device rate automatically")]
public int desiredSampleRate = 48000;
[Tooltip("Mono recommended for Whisper")]
public int channels = 1;
[Range(1, 10)] public int bufferLengthSec = 5;
[Header("Loopback (monitor your voice)")]
public bool playLoopback = true;
[Range(0f, 2f)] public float loopbackVolume = 1.0f;
public delegate void OnWhisperResultProcessedDelegate(string result);
public event OnWhisperResultProcessedDelegate OnWhisperResultProcessed;
// FMOD
private FMOD.System _core;
private Sound _recSound;
private Channel _playChannel;
private ChannelGroup _masterGroup;
private uint _soundPcmLength; // in samples
private int _nativeRate;
private int _nativeChannels;
// ring-buffer tracking
private uint _lastRecordPos = 0;
// Whisper
private WhisperStream _stream;
private bool _streamStarted;
// temp conversion buffer
private float[] _floatTmp = new float[0];
private bool isRecordingActivated = false;
private void Awake()
{
if (!whisper) whisper = FindObjectOfType<WhisperManager>();
_core = RuntimeManager.CoreSystem; // FMOD core system
}
private async void Start()
{
// Query device info to get native rate/channels.
// (FMOD: getRecordDriverInfo gives you system rate & speaker mode)
string name;
Guid guid;
SPEAKERMODE sm;
int smChannels;
DRIVER_STATE driverState;
// signature: getRecordDriverInfo(id, out name, nameLen, out guid, out systemrate, out speakermode, out speakermodechannels, out driverState)
_core.getRecordDriverInfo(recordDriverId, out name, 256, out guid, out _nativeRate, out sm, out smChannels, out driverState);
_nativeChannels = channels > 0 ? channels : smChannels;
UnityEngine.Debug.Log($"[FMOD→Whisper] Using input device #{recordDriverId}: \"{name}\" rate={_nativeRate} ch={_nativeChannels}");
// Build a user sound buffer that FMOD will fill (OPENUSER | LOOP_NORMAL).
CREATESOUNDEXINFO ex = new CREATESOUNDEXINFO
{
cbsize = Marshal.SizeOf(typeof(CREATESOUNDEXINFO)),
numchannels = _nativeChannels,
defaultfrequency = (_nativeRate > 0) ? _nativeRate : desiredSampleRate,
format = SOUND_FORMAT.PCM16,
length = (uint)(((_nativeRate > 0 ? _nativeRate : desiredSampleRate) * _nativeChannels) * sizeof(short)) // seconds=1 (we loop)
};
_core.createSound("", MODE.OPENUSER | MODE.LOOP_NORMAL | MODE.CREATESAMPLE, ref ex, out _recSound);
_recSound.getLength(out _soundPcmLength, TIMEUNIT.PCM);
// Start FMOD recording into that sound (looping ring buffer).
_core.recordStart(recordDriverId, _recSound, true);
UnityEngine.Debug.Log("[FMOD→Whisper] Recording started.");
// Optional loopback playback using FMOD (plays same sound ring buffer).
_core.getMasterChannelGroup(out _masterGroup);
if (playLoopback)
{
_core.playSound(_recSound, _masterGroup, false, out _playChannel);
_playChannel.setMode(MODE._2D);
_playChannel.setVolume(loopbackVolume);
UnityEngine.Debug.Log("[FMOD→Whisper] Loopback playback started.");
}
// Create Whisper stream WITHOUT MicrophoneRecord, just from (freq, channels).
// We'll push AudioChunk manually.
// NOTE: WhisperStreams sliding window is governed by managers stepSec/keepSec/lengthSec.
_stream = await whisper.CreateStream(ex.defaultfrequency, _nativeChannels);
_stream.OnResultUpdated += (txt) =>
{
//OnWhisperResultProcessed?.Invoke(txt);
//UnityEngine.Debug.Log($"[Whisper] result updated: {txt}");
};
_stream.OnSegmentUpdated += (seg) =>
{
OnWhisperResultProcessed?.Invoke(seg.Result);
//UnityEngine.Debug.Log($"[Whisper] Seg finished: {seg.Result}");
};
// If you want Whisper to respect VAD, enable in manager or set useVad (manager controls stream params).
whisper.useVad = useVadInStream;
_stream.StartStream();
_streamStarted = true;
// prepare temp arrays roughly 100ms of audio
EnsureTmpCapacity((ex.defaultfrequency / 10) * _nativeChannels);
}
private void Update()
{
if (!isRecordingActivated) return;
if (_core.handle != IntPtr.Zero) _core.update();
if (!_streamStarted || !_recSound.hasHandle()) return;
// How many samples recorded since last frame?
uint recPos;
_core.getRecordPosition(recordDriverId, out recPos);
uint deltaSamples = (recPos >= _lastRecordPos)
? (recPos - _lastRecordPos)
: (recPos + _soundPcmLength - _lastRecordPos);
if (deltaSamples == 0) return;
// Well read that region (16-bit) and convert to float[] [-1..1].
// Calculate byte range to lock in sound buffer
uint bytesToRead = deltaSamples * (uint)_nativeChannels * 2; // 16-bit = 2 bytes
uint startBytes = _lastRecordPos * (uint)_nativeChannels * 2;
IntPtr p1, p2;
uint len1, len2;
// Lock can wrap — FMOD splits into p1/p2.
_recSound.@lock(startBytes, bytesToRead, out p1, out p2, out len1, out len2);
try
{
// Convert both parts to float and push to Whisper
if (len1 > 0) CopyPcm16ToFloatAndFeed(p1, len1);
if (len2 > 0) CopyPcm16ToFloatAndFeed(p2, len2);
}
finally
{
_recSound.unlock(p1, p2, len1, len2);
}
_lastRecordPos = recPos;
}
public void ActivateRecording()
{
isRecordingActivated = true;
}
public void DeactivateRecording()
{
isRecordingActivated = false;
}
private void CopyPcm16ToFloatAndFeed(IntPtr src, uint byteLen)
{
int samples = (int)(byteLen / 2); // 2 bytes per sample
EnsureTmpCapacity(samples);
// Marshal the 16-bit PCM into managed space
// We pin a short[] overlay to avoid copying twice
int shorts = samples;
int byteCount = (int)byteLen;
// Use Marshal.Copy into a short[] then convert to float[-1..1]
// (You can also unsafe copy for speed if needed.)
EnsureShortOverlay(shorts, out short[] sBuf);
Marshal.Copy(src, sBuf, 0, shorts);
for (int i = 0; i < shorts; i++)
{
// 32768f avoids clipping at -32768
_floatTmp[i] = Mathf.Clamp(sBuf[i] / 32768f, -1f, 1f);
}
// Build a chunk for WhisperStream; with VAD off, IsVoiceDetected=true is fine.
var chunk = new AudioChunk
{
Data = _floatTmp.AsSpan(0, shorts).ToArray(),
Frequency = (_nativeRate > 0) ? _nativeRate : desiredSampleRate,
Channels = _nativeChannels,
IsVoiceDetected = true
};
_stream.AddToStream(chunk);
}
private short[] _shortOverlay;
private void EnsureShortOverlay(int samples, out short[] buf)
{
if (_shortOverlay == null || _shortOverlay.Length < samples)
_shortOverlay = new short[Mathf.NextPowerOfTwo(samples)];
buf = _shortOverlay;
}
private void EnsureTmpCapacity(int samples)
{
if (_floatTmp == null || _floatTmp.Length < samples)
_floatTmp = new float[Mathf.NextPowerOfTwo(samples)];
}
private void OnDisable()
{
if (_streamStarted)
{
_stream.StopStream();
_streamStarted = false;
}
if (_playChannel.hasHandle()) { _playChannel.stop(); _playChannel.clearHandle(); }
if (_recSound.hasHandle()) { _core.recordStop(recordDriverId); _recSound.release(); _recSound.clearHandle(); }
}
}

View File

@@ -0,0 +1,11 @@
fileFormatVersion: 2
guid: a264cb5321d9f6741a2625b8c63c34a7
MonoImporter:
externalObjects: {}
serializedVersion: 2
defaultReferences: []
executionOrder: 0
icon: {instanceID: 0}
userData:
assetBundleName:
assetBundleVariant:

View File

@@ -2,7 +2,7 @@
using UnityEngine;
[RequireComponent(typeof(AudioSource))]
public class MicrophoneTesting : MonoBehaviour
public class MicLoopback : MonoBehaviour
{
[Header("Mic settings")]
[Tooltip("Leave empty for default device")]

Binary file not shown.