jamosdev
/
rhubarb-lip-sync
mirror of https://github.com/DanielSWolf/rhubarb-lip-sync.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

VAD working

This commit is contained in:
Daniel Wolf 2019-10-04 14:04:31 +02:00
parent 295a80578d
commit 0321be6392
1 changed files with 18 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
rhubarb/src/main/kotlin/vad.kt
View File

@ -27,7 +27,8 @@ fun CountLeadingZeros32(n: UInt): Int {
// Normalize n by rounding up to the nearest number that is a sequence of 0
// bits followed by a sequence of 1 bits. This number has the same number of
// leading zeros as the original n. There are exactly 33 such values.
var normalized = n or (n shr 1)
var normalized = n
normalized = normalized or (normalized shr 1)
normalized = normalized or (normalized shr 2)
normalized = normalized or (normalized shr 4)
normalized = normalized or (normalized shr 8)
@ -318,7 +319,7 @@ class VadInstT(aggressiveness: Aggressiveness = Aggressiveness.Quality) {
// PDF parameters
var noise_means = kNoiseDataMeans.clone()
var speech_means = kSpeechDataStds.clone()
var speech_means = kSpeechDataMeans.clone()
var noise_stds = kNoiseDataStds.clone()
var speech_stds = kSpeechDataStds.clone()
@ -489,7 +490,7 @@ fun GmmProbability(self: VadInstT, features: List<Int>, total_power: Int, frame_
}
// Calculate local speech probabilities used later when updating the GMM.
val h1 = (h1_test shr 12) // Q15
val h1 = h1_test shr 12 // Q15
if (h1 > 0) {
// High probability of speech. Assign conditional probabilities for each
// Gaussian in the GMM. Otherwise use the initialized values, i.e., 0.
@ -530,7 +531,7 @@ fun GmmProbability(self: VadInstT, features: List<Int>, total_power: Int, frame_
// (Q14 * Q11 shr 11) = Q14.
val delt = (ngprvec[gaussian] * deltaN[gaussian]) shr 11
// Q7 + (Q14 * Q15 shr 22) = Q7.
nmk2 = nmk + (delt * kNoiseUpdateConst) shr 22
nmk2 = nmk + ((delt * kNoiseUpdateConst) shr 22)
}
// Long term correction of the noise mean.
@ -574,7 +575,7 @@ fun GmmProbability(self: VadInstT, features: List<Int>, total_power: Int, frame_
self.speech_means[gaussian] = smk2 // Q7.
// (Q7 shr 3) = Q4. With rounding.
tmp_s16 = ((smk + 4) shr 3)
tmp_s16 = (smk + 4) shr 3
tmp_s16 = features[channel] - tmp_s16 // Q4
// (Q11 * Q4 shr 3) = Q12.
@ -737,15 +738,15 @@ fun FindMinimum(self: VadInstT, feature_value: Int, channel: Int): Int {
val offset = channel * 16
// Accessor for the age of each value of the [channel]
val age = object {
inline operator fun get(i: Int) = self.index_vector[offset + i]
inline operator fun set(i: Int, value: Int) { self.index_vector[offset + i] = value }
val age = object { // TODO: Inline?
operator fun get(i: Int) = self.index_vector[offset + i]
operator fun set(i: Int, value: Int) { self.index_vector[offset + i] = value }
}
// Accessor for the 16 minimum values of the [channel]
val smallest_values = object {
inline operator fun get(i: Int) = self.low_value_vector[offset + i]
inline operator fun set(i: Int, value: Int) { self.low_value_vector[offset + i] = value }
val smallest_values = object { // TODO: Inline?
operator fun get(i: Int) = self.low_value_vector[offset + i]
operator fun set(i: Int, value: Int) { self.low_value_vector[offset + i] = value }
}
assert(channel < kNumChannels)
@ -757,7 +758,7 @@ fun FindMinimum(self: VadInstT, feature_value: Int, channel: Int): Int {
age[i]++
} else {
// Too old value. Remove from memory and shift larger values downwards.
for (j in i until 16) {
for (j in i until 15) {
smallest_values[j] = smallest_values[j + 1]
age[j] = age[j + 1]
}
@ -925,13 +926,13 @@ fun AllPassFilter(input: AudioBuffer, filter_coefficient: Int, filter_state: Mut
// First 6 taps of the impulse response:
// 0.6399 0.5905 -0.3779 0.2418 -0.1547 0.0990
val result = SampleArray(input.size / 2)
val result = SampleArray((input.size + 1) / 2)
var state32 = filter_state.toInt() * (1 shl 16) // Q15
for (i in 0 until input.size step 2) {
val tmp32 = state32 + filter_coefficient * input[i]
val tmp16 = tmp32 shr 16 // Q(-1)
result[i / 2] = tmp16.toShort()
state32 = (input[i] * (1 shl 14)) - filter_coefficient * tmp16 // Q14
state32 = input[i] * (1 shl 14) - filter_coefficient * tmp16 // Q14
state32 *= 2 // Q15.
}
filter_state.setValue(state32 shr 16) // Q(-1)
@ -1045,7 +1046,7 @@ fun LogOfEnergy(input: AudioBuffer, offset: Int, total_energy: MutableInt): Int
// [kLogConst] is in Q9, [log2_energy] in Q10 and [tot_rshifts] in Q0.
// Note that we in our derivation above have accounted for an output in Q4.
var log_energy = (((kLogConst * log2_energy) shr 19) + (tot_rshifts * kLogConst) shr 9)
var log_energy = ((kLogConst * log2_energy) shr 19) + ((tot_rshifts * kLogConst) shr 9)
if (log_energy < 0) {
log_energy = 0
@ -1203,5 +1204,6 @@ fun ProcessVad(self: VadInstT, fs: Int, audio_frame: AudioBuffer): Boolean {
assert(fs == 8000)
val vad = CalcVad8khz(self, audio_frame)
return vad != 0
// return vad != 0
return vad == 1
}