Table 6.

PD vs. HC classification EER (in %) obtained with different classifiers: MFCC-GMM baseline, and x-vectors combined either with cosine similarity (alone and with LDA) or with PLDA, with and without data augmentation.

	High-quality microphone				Telephone
	Males		Females		Males		Females
	Repet	Monol	Repet	Monol	Repet	Monol	Repet	Monol
MFCC-GMM	22	26	42	45	35	36	42	40
x-vec + cos	32	35	51	41	39	33	49	43
x-vec + LDA + cos	22	27	39	32	32	35	34	34
x-vec + augLDA + cos	24	25	34	30	33	33	39	33
x-vec + PLDA	24	28	39	35	33	36	34	36
x-vec + augPLDA	25	25	33	30	31	33	37	33

The datasets used are male and female high-quality microphone and telephone recordings. Analyzed tasks are free speech (monolog) and sentence repetitions (combined with readings for high-quality microphone recordings). Bold numbers indicate the best EERs for each dataset.