Q-value is a value that needs to be chosen when converting raw data to spectrograms. A low Q-value means we try to match short wave trains for a shorter amount of crests and troughs at various frequencies to the signal at various points in time, giving good time resolution and coarse frequency resolution; a high Q value uses much longer wave trains to give finer frequency resolution at the expense of coarser time resolution. Each spectrogram is computed at four values of Q, the one that gives the highest signal-to-noise ratio is used to create the image we get to see.
In gravitational wave detection and similar scientific fields, the Q-value plays a pivotal role in the analysis and comparison of subjects, referred to as glitches. This metric, along with time frames (Fk), is instrumental in examining the morphology of glitches across various subjects.
Q-values: Qn
Q1 = 5.65685424949238 = 2-1 x 11.31370849898476 = 22.5
Q2= 11.31370849898476 = 20x 11.31370849898476 = 23.5
Q3= 22.627416997969522 = 21 x 11.31370849898476 = 24.5
Q4= 45.25483399593904 = 22 x 11.31370849898476 = 25.52xQ1=Q2, 2xQ2=Q3, 2xQ3=Q4
Q4=2xQ3=4x2=8xQ1
Qn=2xQn-1Time frames: Fk
frame1 (F1): t = 0.5 sec = 2-1 sec
frame2 (F2): t = 1.0 sec = 20 sec
frame3 (F3): t = 2.0 sec = 21 sec
frame4 (F4): t = 4.0 sec = 22 sec2xF1=F2, 2xF2=F3, 2xF3=F4
F4=2xF3=4xF2=8xF1
Fk=2xF(k-1)
Generally the morphology of two subjects with different Q can be compared the best when their q-value divided by the time duration of the frame (q/t) are equal.
Q-value comparisons are essential for several reasons: