Help on Semantic Differential, PCA - Input values?

Dear all,

I am not good in statistics and I would be grateful if you could help me.

I conducted a study on the perception of 50 different sounds and used the semantic differential technique to measure the response of subjects. The scale used was a double sided 5-point scale.
(Example: Quiet 5 4 3 2 1 0 1 2 3 4 5 Loud)

Altogether I tested 10 different attribute pairs.
Now in order to find the main underlying dimensions I want to apply a PCA analysis using SPSS.

I am stuck at the very start: What kind of values do I have to input to SPSS to apply the PCA?

Do I have to calculate a mean value for each attribute scale and subject this 50 x 10 matrix to a PCA?

Or would it make more sense to first calculate the mean values for each scale (50 x 10 matrix), intercorrelate this matrix using Pearson (50 x 50 matrix) and then subject it to a PCA?

I've been reading lots of books about PCA but couldn't find any solution to my problem.

Thank you in advance.