Files
Obsidian-Main/00. Inbox/00.01. 雜/使用 librosa 做 mel spectrogram.md
Awin Huang 65df92aa08 vault backup: 2025-05-28 16:39:41
Affected files:
00. Inbox/00.01. 雜/AI Prompt.md
00. Inbox/00.01. 雜/ATP.md
00. Inbox/00.01. 雜/Brown noise.md
00. Inbox/00.01. 雜/Build xmrig with Visual Studio 2022.md
00. Inbox/00.01. 雜/GPU over IP.md
00. Inbox/00.01. 雜/Gitzo 腳架編號規則.md
00. Inbox/00.01. 雜/HEIF.md
00. Inbox/00.01. 雜/Home Project.md
00. Inbox/00.01. 雜/MPPT.md
00. Inbox/00.01. 雜/Mel spectrogram.md
00. Inbox/00.01. 雜/SSIM.md
00. Inbox/00.01. 雜/White noise.md
00. Inbox/00.01. 雜/flask.canvas
00. Inbox/00.01. 雜/三寶荒神.md
00. Inbox/00.01. 雜/不淨觀.md
00. Inbox/00.01. 雜/亞甲藍.md
00. Inbox/00.01. 雜/使用 librosa 做 FFT.md
00. Inbox/00.01. 雜/使用 librosa 做 mel spectrogram.md
00. Inbox/00.01. 雜/出離.md
00. Inbox/00.01. 雜/厭離.md
00. Inbox/00.01. 雜/台語諺語.md
00. Inbox/00.01. 雜/名言佳句.md
00. Inbox/00.01. 雜/固定型心態.md
00. Inbox/00.01. 雜/如何表達(How to Speak).md
00. Inbox/00.01. 雜/布萊茲‧帕斯卡(Blaise Pascal).md
00. Inbox/00.01. 雜/德國狼犬.md
00. Inbox/00.01. 雜/成長型心態.md
00. Inbox/00.01. 雜/挖礦.canvas
00. Inbox/00.01. 雜/時間不一致性.md
00. Inbox/00.01. 雜/皮質醇.md
00. Inbox/00.01. 雜/知識管理.md
00. Inbox/00.01. 雜/稼動率.md
00. Inbox/00.01. 雜/自我成長.canvas
00. Inbox/00.01. 雜/飼料.md
00. Inbox/00.01. 雜/魚病.canvas
00. Inbox/00.01. 雜/魚藥 - Levamisole(左旋咪唑、左美索、左美素).md
00. Inbox/00.01. 雜/魚藥 - Mebendazole(美鞭達唑).md
00. Inbox/00.01. 雜/魚藥 - Metronidazole(硝基嘧唑乙醇).md
01.00. Me/03. 🀄 Projects.md
10. 日記/2025-05-28(週三).md
2025-05-28 16:39:41 +08:00

13 lines
620 B
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
由於人類會對低頻低音高的片段更感興趣,所以會對通過 FFT 變換得到的 Amplitude 和 Frequency。
[[Mel spectrogram]] 和 spectrogram 的差別就是 mel spectrogram 的頻率是 mel scale 變換後的頻率你可以想像把Spectrogram整體往下壓
```python
mel_spect = librosa.feature.melspectrogram(y=y, sr=sr, n fft=2048, hop_Iength=1024)
mel_spect = librosa.power_to_db(mel_spect, ref=np.max)
librosa.display.specshow(mel_spect, y_axis='mel', fmax=8000, x_axis='time')
plt.title('Mel Spectrogram')
p1t.colorbar(format='%+2.0f dB')
```
Output:
![[Pasted image 20231212181946.png]]