To provide a method for easily or accurately detecting or predicting the occurrence timing of goby during roasting of coffee beans.SOLUTION: The coffee bean jelly detection method according to the present invention is a method for measuring coffee beans being roasted and a sound having a frequency within a range of 5 kHz or more and 80 kHz or less generated from a roasting device over time , An average value calculation step of calculating an average value of the sound pressure during the haze non-occurrence period, and an average value calculation step of calculating the average value of the sound pressure from the time when the measured sound pressure exceeds the threshold value and exceeds the threshold value, (N is greater than 1/6 but less than 1/3) of the threshold value in a time width of 0.1 milliseconds or more and 99 milliseconds or less selected from a period from 1 millisecond to 100 milliseconds before And a haze detection step of judging that a gob of coffee bean has been detected when the sound pressure is not detected.(FIG.【課題】コーヒー豆の焙煎に際し、ハゼの発生時期を簡易且つ正確に検知又は予測する方法を提供する。【解決手段】本発明に係るコーヒー豆のハゼ検知方法は、焙煎中のコーヒー豆及び焙煎装置から発生する5kHz以上80kHz以下の範囲に含まれる範囲の周波数の音を経時的に測定するに際し、ハゼ未発生期間中の音圧の平均値を算出する平均値算出工程と、音圧の平均値を閾値として用いて、測定した音圧が閾値を超え、且つ閾値を超えた時点より0.1ミリ秒前から100ミリ秒前までの期間から選択される0.1ミリ秒以上99ミリ秒以下の時間幅において、閾値のn倍以上(nは、1/6超1/3未満)の音圧を検知していない場合に、コーヒー豆のハゼを検知したと判断するハゼ検知工程と、を有する。【選択図】図4
