flykhan/maths-cs-ai-compendium-zh
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Deployed 2536c93 with MkDocs version: 1.6.1

Browse Source
This commit is contained in:
flykhan
2026-05-03 11:47:30 +08:00
commit 1ff86b66fc
418 changed files with 617336 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
images/mel_filterbank.svg
+105
View File
@@ -0,0 +1,105 @@
<svg width="700" height="300" xmlns="http://www.w3.org/2000/svg">
<defs>
<marker id="mf-arrow" markerWidth="8" markerHeight="6" refX="8" refY="3" orient="auto">
<polygon points="0,0 8,3 0,6" fill="#666"/>
</marker>
</defs>
<text x="350" y="22" fill="#333" font-size="14" font-weight="bold" text-anchor="middle">Mel-Scale Triangular Filterbank</text>
<!-- Axes -->
<line x1="60" y1="220" x2="640" y2="220" stroke="#666" stroke-width="1" marker-end="url(#mf-arrow)"/>
<line x1="60" y1="220" x2="60" y2="40" stroke="#666" stroke-width="1" marker-end="url(#mf-arrow)"/>
<!-- X-axis label -->
<text x="350" y="245" fill="#666" font-size="10" text-anchor="middle">Frequency (Hz)</text>
<!-- Y-axis label -->
<text x="30" y="130" fill="#666" font-size="10" text-anchor="middle" transform="rotate(-90, 30, 130)">Filter Weight</text>
<!-- X-axis tick marks and labels -->
<line x1="60" y1="220" x2="60" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="60" y="236" fill="#666" font-size="8" text-anchor="middle">0</text>
<line x1="132" y1="220" x2="132" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="132" y="236" fill="#666" font-size="8" text-anchor="middle">1000</text>
<line x1="204" y1="220" x2="204" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="204" y="236" fill="#666" font-size="8" text-anchor="middle">2000</text>
<line x1="276" y1="220" x2="276" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="276" y="236" fill="#666" font-size="8" text-anchor="middle">3000</text>
<line x1="348" y1="220" x2="348" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="348" y="236" fill="#666" font-size="8" text-anchor="middle">4000</text>
<line x1="420" y1="220" x2="420" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="420" y="236" fill="#666" font-size="8" text-anchor="middle">5000</text>
<line x1="492" y1="220" x2="492" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="492" y="236" fill="#666" font-size="8" text-anchor="middle">6000</text>
<line x1="564" y1="220" x2="564" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="564" y="236" fill="#666" font-size="8" text-anchor="middle">7000</text>
<line x1="636" y1="220" x2="636" y2="225" stroke="#666" stroke-width="0.8"/>
<text x="636" y="236" fill="#666" font-size="8" text-anchor="middle">8000</text>
<!-- Y-axis ticks -->
<line x1="55" y1="220" x2="60" y2="220" stroke="#666" stroke-width="0.8"/>
<text x="50" y="224" fill="#666" font-size="8" text-anchor="end">0</text>
<line x1="55" y1="70" x2="60" y2="70" stroke="#666" stroke-width="0.8"/>
<text x="50" y="74" fill="#666" font-size="8" text-anchor="end">1.0</text>
<line x1="55" y1="145" x2="60" y2="145" stroke="#666" stroke-width="0.8"/>
<text x="50" y="149" fill="#666" font-size="8" text-anchor="end">0.5</text>
<!-- Mel-spaced triangular filters -->
<!-- Filters are narrower at low frequencies, wider at high frequencies -->
<!-- Each filter: triangle from base left to peak to base right -->
<!-- Filter 1: 60-100-140 Hz (very narrow, low freq) -->
<polygon points="60,220 67,70 78,220" fill="#3498db" opacity="0.15" stroke="#3498db" stroke-width="1.2"/>
<!-- Filter 2: 78-118-158 -->
<polygon points="78,220 96,70 118,220" fill="#e74c3c" opacity="0.15" stroke="#e74c3c" stroke-width="1.2"/>
<!-- Filter 3: 96-140-184 -->
<polygon points="96,220 118,70 147,220" fill="#27ae60" opacity="0.15" stroke="#27ae60" stroke-width="1.2"/>
<!-- Filter 4: 118-170-222 (wider) -->
<polygon points="118,220 147,70 183,220" fill="#f39c12" opacity="0.15" stroke="#f39c12" stroke-width="1.2"/>
<!-- Filter 5: 147-210-273 -->
<polygon points="147,220 183,70 228,220" fill="#9b59b6" opacity="0.15" stroke="#9b59b6" stroke-width="1.2"/>
<!-- Filter 6: 183-260-340 (getting wider) -->
<polygon points="183,220 228,70 285,220" fill="#3498db" opacity="0.15" stroke="#3498db" stroke-width="1.2"/>
<!-- Filter 7: 228-325-425 -->
<polygon points="228,220 285,70 355,220" fill="#e74c3c" opacity="0.15" stroke="#e74c3c" stroke-width="1.2"/>
<!-- Filter 8: 285-405-530 -->
<polygon points="285,220 355,70 442,220" fill="#27ae60" opacity="0.15" stroke="#27ae60" stroke-width="1.2"/>
<!-- Filter 9: 355-505-660 (much wider) -->
<polygon points="355,220 442,70 540,220" fill="#f39c12" opacity="0.15" stroke="#f39c12" stroke-width="1.2"/>
<!-- Filter 10: 442-600-760 -->
<polygon points="442,220 540,70 636,220" fill="#9b59b6" opacity="0.15" stroke="#9b59b6" stroke-width="1.2"/>
<!-- Annotations -->
<line x1="80" y1="50" x2="80" y2="58" stroke="#27ae60" stroke-width="0.8"/>
<line x1="80" y1="58" x2="140" y2="58" stroke="#27ae60" stroke-width="0.8" stroke-dasharray="3,2"/>
<text x="145" y="54" fill="#27ae60" font-size="9">Narrow filters: fine</text>
<text x="145" y="63" fill="#27ae60" font-size="9">resolution at low freq</text>
<line x1="540" y1="50" x2="540" y2="58" stroke="#e74c3c" stroke-width="0.8"/>
<line x1="540" y1="58" x2="460" y2="58" stroke="#e74c3c" stroke-width="0.8" stroke-dasharray="3,2"/>
<text x="410" y="54" fill="#e74c3c" font-size="9">Wide filters: coarse</text>
<text x="410" y="63" fill="#e74c3c" font-size="9">resolution at high freq</text>
<!-- Bottom note -->
<rect x="80" y="262" width="540" height="28" rx="4" fill="#f5f5f5" stroke="#333" stroke-width="0.8"/>
<text x="350" y="276" fill="#333" font-size="10" text-anchor="middle">Mel scale compresses high frequencies: mel(f) = 2595 . log10(1 + f/700)</text>
<text x="350" y="288" fill="#666" font-size="9" text-anchor="middle">Mimics human auditory perception -- we distinguish low frequencies better than high</text>
</svg>
Side by Side

After

Width:  |  Height:  |  Size: 5.6 KiB