Praat Introduction to Praat: Let's take a look at sounds : * WS Dec/01/'14 : ver. 1.1.4 1. Praat STEP 1: STEP 2: STEP 3: STEP 4: STEP 2 Editor STEP 3 Dynamic menu 2 FAQ: Pitch analysis http://goo.gl/r65la 2. : (1) Praat : ToBI: ToBI Annotation Environments ToBI Annotation Environments http://goo.gl/od3st Pros: currently the standard tool for speech processing better documented there are many useful script libraries has an embedded scripting language that facilitates many tasks Cons: both its user interface and its scripting language are non-standard, unintuitive, and often inconsistent script :
: Boersma, Paul & Weenink, David (2014). Praat: doing phonetics by computer [Computer program]. Version 5.4.01, retrieved 9 November 2014 from http://www.praat.org/ FAQ: How to cite Praat http://goo.gl/xeo0o 3. STEP 1: Praat (1) Praat : Praat : Praat Quit * confirm Quit: You have objects in your list! Do you still want to quit Praat? [OK] * [OK] (2) Praat STEP01 [Open] - [Read from file...] sample01_sentence.wav sample01_sentence_stereo.wav sample01_sentence.pitch sample01_sentence.intensity
*shift Ctrl Windows command Mac (3) Object Sound sample01_sentence Sound sample01_sentence_stereo Pitch sample01_sentence Intensity sample01_sentence 4 (4) Sound sample01_sentence_stereo : Sound sample01_sentence_stereo Dynamic menu [View & Edit] 2 Praat mono : Sound sample01_sentence_stereo Dynamic menu Convert - > Convert to mono Object Sound sample01_sentence_stereo_mono
(5) [Remove] Object STEP 2: (1) STEP02 [Open] - [Read from file...] sample02_words.wav sample02_words_sample.textgrid sample03_longsentences.wav sample03_longsentences.textgrid Object Sound sample02_words Sound sample03_longsentences TextGrid sample02_words_sample TextGrid sample03_longsentences 4 (2) Sound sample02_words Sound Annotate -> To TextGrid (3) All tier names Which of these are point tiers? All tier names segments words Which of these are point tiers? * phrases
* Pitch 1 tier pitch point tiers (4) Sound sample02_words Sound TextGrid sample02_words TextGrid 2 View & Edit (5) Annotation * segments, words tier(s) *Tier Tier [all], [in], [out], [sel], [bak] Zoom In Zoom Out Undo
[Total duration], [Visible Part] (6) View - Show analyses... Show spectrogram Show pitch Show intensity Show formants OK pit, bit, tit, dit, kit, git, fit, vit 8 (7) pit [in] [sel] (8) tier Boundary -> Add on all tiers * Boundary Boundary Boundary -> Remove (9) tier selected tier words tier TextGridEditor pit (10) segments tier [p] [i] [t] Boundary -> Add on selected tier *IPA Help Phonetic symbols * (a) Pitch - Pitch listing / Get pitch / Get minimum pitch / Get maximum pitch / Move cursor to minimum pitch / Move cursor to maximum pitch (b) Intensity - Intensity listing / Get intensity (c) Formant - Formant listing / Get first formant / Get second
formant / Get third formant / Get fourth formant / Get formant... ( X ) (d) Spectrum - (11) : File -> Save TextGrid as text file * 7 Object sound sample02_words, TextGrid sample02_words_sample 2 View & Edit WS annotation (12) [Remove] Object * TextGrid (13) STEP02 sample03_longsentences.wav sample03_longsentences.textgrid 2 Object Sound sample03_longsentences TextGrid sample03_longsentences (14) 2 View & Edit * [in] View -> Show analyses... Longest analysis (s): (15) [all] annotation 5 (16) 2 Extract - > Extract all intervals /
Extract non-empty intervals / Extract intervals where * Extract non-empty intervals... (17) Tier number tier Sentence tier 1 Preserve times * * Preserve times (18) Sound Sentence1, Sound Sentence2, Sound Sentence3, Sound Sentence4, Sound Sentence5 5 Object Sound sample03_longsentences, TextGrid sample03_longsentences 2 [Remove] STEP 3: STEP 2 Sound Sentence1 duration, pitch, intensity (1) Duration: Objects Sound Sentence1 Query - > Query time domain >
Get start tim e: セグメントの開始点 Get end tim e: セグメントの終了点 Get total duration: セグメントの長さ (2) Pitch: Objects ウィンドウに開かれた Sound Sentence1 を選択して Analyse periodicity - > To pitch... #オプション指定可 Time step (s)が 0.0 だと 10ms ごとに計算します Pitch floor, Pitch ceiling について Praat の Help では男性: 75-300 Hz 女性: 100-500 Hz を推奨していま す *ピッチ抽出のデフォルトは 10ms ごと 変更可 で 分析窓長は中心から前後に 20ms ず つの計 40ms ですから あまり短いセグメントに対する F0 抽出は効率が悪くなります src: Script for listing F0 statistics goo.gl/uffacn でき上がった Pitch Sentence1 を選択して Query - > 必要な情報を選択 Get m ean...: 平均ピッチ値 Get m inim um...: ピッチの最低値 Get m axim um...: ピッチの最高値 Get tim e of m inim um...: 最低値の時間情報 Get tim e of m axim um...: 最高値の時間情報 *周波数情報について Hz 以外の単位を設定することもできます
(3) Intensity: Objects Sound Sentence1 To Intensity # Intensity Sentence1 Query - > Get minimum...: Intensity Get maximum...: Intensity Get time of minimum...: Get time of maximum...: * 1 1 * WS : 0 sec : 8.897 sec : 8.897 sec F0 : 77 Hz F0 : 208 Hz F0 : 4.028 sec F0 : 4.823 sec * Sentence[2-5] Sentence[1-5], Pitch Sentence1, Intensity Sentence1 [Remove] STEP 4: Pitch (1) (2) STEP 3 Pitch Sentence1 Draw - Draw...
* object * Annotation STEP 2 Annotation File -> Draw visible sound and Textgrid (3) File -> Save as EPS file -> Save as EPS file... Mac File -> Save as Windows metafile... Windows (4) Word Tips: Sound 2 Pitch Intensity 1 : Sentence1 Pitch Intensity 1 (1) 1 Pitch (2) Pen > Dotted line (3) 2 Intensity Margins > Text right Intensity (db) Far (4) Intensity Garnish (5) Margins > Marks > Marks right : Number of marks: 6, Write numbers Draw ticks Draw dotted lines Tips: Sound Pitch 1 : Sentence1 Sentence3 1 (1) Object Sentence1, Sentence3 [Info] Total duration Sentence1: 8.897 sec, Sentence3: 14.011 sec (2) Sentence1, Sentence3 Pitch (3) Pitch Sentence1 Draw... Time range (s): 0.0 15.0 * (4) Pen > Dotted line (5) Pitch Sentence3 Draw... Time range (s): 0.0 15.0 Garnish 0-15 2 Before & After