Hello Kristyna,

I think you could present participants with a series of sad and happy music excerpts and then ask them to view photos of faces expressing either a happy or sad phase. I believe these two papers could be helpful in designing your studies.

http://research.gold.ac.uk/4213/1/NSL25933.pdf

http://daniellevitin.com/levitinlab/articles/2011-01119-Perception.pdf

You could use SEM to model the relationship between respondents'perception on music and their emotional effect. 

Methodology - you play the music to the participants for certain period of time (duration). After that you assess their perception concerning the music that they were listening to and also their state of emotion resulted from the music. You can employ the likert interval scale to measure the respondents assessment assessment for the constructs. The you model the data using Amos Graphic and analyze it using Structural Equation Modeling (SEM).

Why not just ask them to tell you how they feel.  I'm presuming it's their perception you want to capture not your guess at how they perceive it.

You could devise a series of questions for them to answer, but that puts much more of the researcher into the process.

Maybe one open question: How did that make you feel? or something like that, then let them construct their own narrative. You can deconstruct or analyze the text later using various analytical approaches.

You have an implicit hypothesis that music elicits emotions. Research on music perception has proliferated in recent years, but has gone well beyond the simple hypothesis of a specific link with emotion (this was the dominant theory back in the period 1650 to 1730 or so). 

Music is a 'hard case' because verbal descriptions of music are almost always inadequate to describe the effects. Some theorists have argued that this is because music is inherently meaningless (the critic Hanslick is associated with this position) but others, including Mendelssohn, have argued that the meaning of music is far more specific than linguistic meaning. There are many phenomena that we can call 'pizza', but each piece of music is unique. 

Listen to the opening of the sixth Bartók quartet. What is that emotion? The viola plays a melody that will dominate the whole quartet, at once utterly precise and utterly elusive. To reduce the experience to words is to betray it.

I recommend you read up on contemporary music research, which is really, really interesting, and develop your ideas based on what we know (or theorise!) about the relationship between music and syntactic processing (or even prosodic processing)

Link to the literature: http://www.frontiersin.org/SearchData.aspx?sq=music&ht=1

Note a particularly relevant article here: http://journal.frontiersin.org/article/10.3389/fpsyg.2014.00327/abstract

It is an excellent introduction to contemporary thinking (and debate) on the subject of music and mood.

https://www.youtube.com/watch?v=6ya_J--MwZU

You can play a music for duration of about 1 minute to 3 minutes and ask listeners to rate emotions on the scale 1 to 10.  

Listener's familiarity or liking about a particular genre or artist can effect the results. Music learners can also have a different view. The questionnaire should have musical background and likes/dislikes of listeners.  

Same music listening experience can be different for same listeners at different times depending on their mind set, time and environment etc. 

Which 6 basic emotions you are looking for? 

I think that presenting people with 'happy' and 'sad' music isn't good enough any more. Music is neither happy nor sad, and most music does not produce an emotional response that is simply classifiable or is invariable from one person to another. 

My advice would remain the same: read up on the contemporary literature, which is far more sophisticated in its theoretical perspectives, and far more sophisticated in its research designs than the contributions here on RG would suggest.