Spaces:
Running
on
Zero
Running
on
Zero
File size: 11,255 Bytes
7ee3434 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 |
/**
* Copyright (c) 2023 Amphion.
*
* This source code is licensed under the MIT license found in the
* LICENSE file in the root directory of this source tree.
*/
// ==== Event Listeners ====
let titleClickCount = 0;
$$("title").addEventListener("click", (e) => {
titleClickCount++;
if (titleClickCount > 5) {
titleClickCount = 0;
const newTitle = prompt("Please enter new title", document.title);
if (newTitle != null) {
$$("title").innerHTML = newTitle;
document.title = newTitle;
}
}
})
$$("mode_change").addEventListener("click", () => {
if (userMode === "basic") {
window.location.href = window.location.href.split("?")[0] + "?mode=advanced";
}
if (userMode === "advanced") {
window.location.href = window.location.href.split("?")[0] + "?mode=basic";
}
})
$$("mode_id").addEventListener("change", () => {
const selectedValue = $$("mode_id").value;
currentMode = selectedValue;
if (currentMode !== "Metric Comparison") {
if (currentHistogram.length > 0) d3.select(`#rect_${currentHistogram.shift().name}`).attr("stroke", "none")
unlockOptions();
}
refreshOptions();
resetDisplay();
});
$$("pic_id").addEventListener("change", () => {
const selectedValue = $$("pic_id").value;
currentShowingPic = selectedValue;
resetDisplay(false, false);
});
$$("range").addEventListener("input", lineChange);
$$("reset_map").addEventListener("click", () => {
resetDisplay(false, false);
})
$$("value").addEventListener("change", (e) => {
const value = e.target.value;
const lValue = 999 - value;
$$('range').value = lValue;
lineChange();
})
$$("components_pitch").addEventListener("change", (e) => {
hidePitch = !e.target.checked;
charts.forEach(c => c.reset())
// resetDisplay(false, false);
})
$$("components_frequncy").addEventListener("change", (e) => {
if (e.target.checked) {
showFrequency = [$$("inputs_min").value, $$("inputs_max").value]
} else {
showFrequency = false
}
charts.forEach(c => c.reset())
// resetDisplay(false, false);
})
$$("inputs_min").addEventListener("input", (e) => {
if (e.target.value > parseInt($$("inputs_max").value)) $$("inputs_min").value = $$("inputs_max").value
if (!$$("components_frequncy").checked) return
showFrequency = [e.target.value, $$("inputs_max").value]
charts.forEach(c => c.reset())
$$("inputs_left").innerHTML = e.target.value
})
$$("inputs_max").addEventListener("input", (e) => {
if (e.target.value < parseInt($$("inputs_min").value)) $$("inputs_max").value = $$("inputs_min").value
if (!$$("components_frequncy").checked) return
showFrequency = [$$("inputs_min").value, e.target.value]
charts.forEach(c => c.reset())
$$("inputs_right").innerHTML = e.target.value
})
const delayedLineChange = () => {
if (!sampledSteps || sampledSteps.length == 0) {
console.log('detect empty sampledSteps, waiting')
setTimeout(delayedLineChange, 100)
return
}
lineChange();
}
$$("sampling_steps").addEventListener("change", (e) => {
samplingSteps = e.target.checked;
sampledSteps = []
// TODO: update the display
resetDisplay(false, false);
setTimeout(delayedLineChange, 100)
})
$$("sampling_num").addEventListener("change", (e) => {
if (!e.target.value) {
alert('Please enter a number.')
return
}
if (parseInt(e.target.value) >= 1000) {
alert('Please enter a number less than 1000.')
return
}
if (parseInt(e.target.value) <= 0) {
alert('Please enter a number greater than 0.')
return
}
samplingNum = e.target.value;
sampledSteps = []
resetDisplay(false, false);
setTimeout(delayedLineChange, 100)
})
$$("step_axis").addEventListener('touchmove', (e) => {
e.stopPropagation();
}, { passive: false });
// $$("reset_preview").addEventListener("click", () => {
// charts.filter((c) => !c.sync).forEach((c) => c.reset());
// });
$$("add_preview").addEventListener("click", () => {
if (displaySteps.length >= 3 && userMode !== "basic") {
alert('Please remove step before pinning new one. Up to 3 steps can be pinned.')
return
}
const step = 999 - $$("range").value;
selectStep(step);
})
$$("help").addEventListener("click", () => {
localStorage.removeItem('GUIDED');
location.reload();
})
$$("metrics_help").addEventListener("click", () => {
alert("Metric Tips", "Scores: The higher the better. Log scores: The lower the better. \n" +
"The metrics are calculated by algorithm. Each metric is defined as follows:\n\n" +
"- <b>Dembed (Singer Similarity)</b>: This quantitatively assesses the similarity between the timbre of the original singer's voice and the converted voice. It's calculated using the cosine similarity between feature vectors representing the timbre characteristics of the two voices. A higher similarity score indicates the more timbre similarity. \n" +
"- <b>F0CORR (Pitch Correlation)</b>: This measures the Pearson Correlation Coefficient between the F0 values of the converted singing voice and the target voice. It assesses the linear relationship between the F0 contours of the two voices. A higher F0CORR indicates a stronger correlation and better F0 similarity. \n" +
"- <b>FAD (Fréchet Audio Distance)</b>: This is a reference-free evaluation metric to evaluate the quality of audio samples. FAD correlates more closely with human perception. A lower FAD score indicates a higher quality of the audio. \n" +
"- <b>F0RMSE (Pitch Accuracy)</b>: This measures the Root Mean Square Error of the Fundamental Frequency (F0) values between the converted singing voice and the target voice. It quantifies how accurately the F0 of the converted voice matches that of the target voice. A lower F0RMSE indicates better F0 accuracy. \n" +
"- <b>Mel-cepstral distortion (MCD)</b>: This assesses the quality of the generated speech by comparing the discrepancy between generated and ground-truth singing voice. It measures how different the two sequences of mel cepstra are. A lower MCD indicates better quality.");
})
$$("projection_help").addEventListener("click", () => {
alert("Projection Tips", "This view depicts the 128-dimensional diffusion steps in the diffusion model reduced to 2-dimensional embedding using t-SNE. Each point in the projection space represents a specific diffusion step. The arrangement of these points in 2D space illustrates how each step varies from others, providing perception of the progression and transformation occurring within the diffusion model.")
})
$$("control_help").addEventListener("click", () => {
alert("Control Tips", "<b>Projection Embedding:</b> \n" +
"- <b>Step (Diffusion Step)</b>: Output embedding of the diffusion step encoder, and it does not contain any other data. \n" +
"- <b>Step + Noise</b>: Output embedding of the diffusion step encoder added with the initial random noise. \n" +
"- <b>Step + Noise + Condition</b>: Output embedding of the diffusion step encoder added with the initial random noise and the model conditions, e.g., F0 and energy extracted from the source singing voice and speaker embedding corresponding to the target singer. (Advanced mode only) \n" +
"- <b>First/Middle/Last Layer</b>: The first, tenth, and final layer of the 20-layer residual blocks in the diffusion decoder. \n" +
"<b>Components:</b> \n" +
"- <b>F0 contour</b>: The fundamental frequency (F0) of speech refers to the primary frequency of vibration of the vocal cords during voiced sound production. Typically ranging from 80 to 450 Hz, F0 varies across individuals, influenced by the length of the vocal cords and factors like age and sex. Though it appears quasi-periodic due to natural variations in the voice called jitter and shimmer, F0 is dynamic and alters within sentences to convey nuances like emphasis or questioning.\n" +
"- <b>Frequency</b>: \"Frequency\" refers to the distribution of different sound frequencies represented across the Mel scale. Each Mel channel corresponds to a range of frequencies and essentially represents a bandpass filter output. These channels are designed to capture relevant features of sound for various audio processing tasks. \n"
)
})
$$("controls").addEventListener("click", () => {
playMode = !playMode;
updatePlayIcon(playMode)
localStorage.setItem('AUTO_PLAY', playMode ? "true" : "false");
if (parseInt($range.value) >= 999 || (samplingSteps && 999 - parseInt($range.value) == parseInt(sampledSteps[sampledSteps.length - 1]))) {
$range.value = 0;
}
})
const preferColor = window.matchMedia('(prefers-color-scheme: dark)')
if (preferColor.matches) {
darkMode = true;
} else {
darkMode = false;
}
preferColor.addEventListener("change", (e) => {
if (e.matches) {
darkMode = true;
} else {
darkMode = false;
}
console.log('darkMode', darkMode);
charts.forEach(c => c.reset())
d3.select("#histogram")
.attr("style", `color: ${darkMode ? 'white' : 'black'};`)
d3.select("#histogram2")
.attr("style", `color: ${darkMode ? 'white' : 'black'};`)
})
const autoPlay = () => {
let waitTime = 50;
if (!playMode) {
setTimeout(autoPlay, 50)
return
}
// auto play
if (samplingSteps) {
// get next step from samped steps
currentIndex = sampledSteps.indexOf(`${999 - $range.value}`); // input str
nextRange = 999 - sampledSteps[currentIndex + 1];
// get the gap and determine the waiting time
// const gap = nextRange - $range.value;
// console.log(gap)
waitTime = 500;
if (nextRange >= parseInt(sampledSteps[0]) || isNaN(nextRange)) {
playMode = false
updatePlayIcon(playMode)
setTimeout(autoPlay, waitTime)
return
}
$range.value = nextRange;
} else {
$range.value = jumpStep + parseInt($range.value);
if (parseInt($range.value) >= 999) {
playMode = false
updatePlayIcon(playMode)
}
}
lineChange(waitTime === 50);
setTimeout(autoPlay, waitTime)
}
setTimeout(autoPlay, 50)
// ==== Drag and Drop ====
const drag = (e) => {
e.preventDefault()
}
const dragEnter = (e) => {
e.preventDefault()
if (e.target.id.indexOf('display') === -1) return
if (!currentCard) return
let melArray = Array.from($mel.childNodes)
let currentIndex = melArray.indexOf(currentCard)
let targetindex = melArray.indexOf(e.target)
if (currentIndex < targetindex) {
$mel.insertBefore(currentCard, e.target.nextElementSibling)
} else {
$mel.insertBefore(currentCard, e.target)
}
}
const dragEnd = (e) => {
if (!currentCard) return;
currentCard.classList.remove('border-dashed', 'border-2', 'border-blue-500')
currentCard = null
}
document.addEventListener('dragover', drag);
document.addEventListener('dragenter', dragEnter);
document.addEventListener('dragend', dragEnd); |