Having a budget means avoiding expensive mistakes that result in dissatisfaction with the result. A strategy or methodology can be very helpful.
The ideal situation is that you hear someone else's system in a room like your own that absolutely pleases you, and when you ask how much it costs it comes under your budget limit... you simply go shopping, get it set up, and enjoy it.
Less ideal is not knowing what the various component options will do, how they'll do it together, and whether that will do it for you. The analytical approach can help narrow the field.
Here is my thinking on how to cut through the combinations and zoom in on solutions. There will still be multiple solutions, but I think this approach clears a good path. Of course there are many other things involved!
1] The air in the room
Determine how loud you want to listen to music. Be honest with yourself and decide on an average level that will make you happy. As a guide, consider that the standard reference level at which recording engineers monitor for mastering has historically been in the 86-87dB sound pressure level in the studio control room; this is the level at which they expect the general listener to play music. This is a pretty loud level at which there is no feeling that anything in the music is not present. Some will want to listen louder at 90dB average, and many will be more than happy with 80dB which is a medium level.
Note:
The tonal balance of the recording was mastered while being monitored at studio level 86-87dB, so it is at that level that the bass and treble balance will approach most what the engineer was hearing. If you play louder or softer, you may need to compensate with tone controls to recover that same tonal balance.
The room is going to have an influence on the sound, specifically in the range from about 250Hz on down, and this influence will increase with higher levels of listening. If your preference is louder listening, room treatments are going to become increasingly necessary to preserve the quality of the sound.
2] The music in the air
Music has peaks above its average level. For pop and rock this peak will be up to 12dB for most recordings; for jazz and classical it will be up to 15dB. There are a few specialized recordings that go up to 18dB. These peak levels above the average levels are crest factors.
Note:
Adding the crest factor to your average listening level determines the peak window at which transient will play in the room. For example, if you take a conservative approach and choose a crest factor of 20dB in order to cover all possible music, and choose an average listening level of 85dB, then the top transient peaks window will extend to 105dB.
The peaks window is going to be used to select speakers and amplification in order that the system plays cleanly up through the selected crest factor.
3] The speaker
The speaker must play clean up through the chosen peaks window. If your peaks window tops out at 105dB as in the example, then your speaker must be able to play clean up to 105dB. Many speakers do not sound clean at high levels, so it is important to match the speaker's characteristic in this regard to your needs. This is also one of the dark areas because speaker distortion specs are rarely admitted (because many speakers distort up to 30% when pressing toward the higher levels).
Note:
Some speaker specifications include maximum sound level in dB. Other do not, but provide a maximum average or peak power level. If the dB sound level maximum is provided, this may exclude the speaker depending on your required level. If the maximum power level is provided, the speaker may be excluded from the results due to the required amp power to meet required level (see below).
Listening tests may also exclude speakers auditioned at required listening level.
4] The amplification
The amp needs to be able to power the speaker through its peak window level cleanly. This relationship between the amp and the speaker depends on the amp's rated power and the speaker's rated sensitivity and many combinations will satisfy this, and some will not. Part of the synergy of a system is getting this matching of the amp and speaker correct.
For example, if one's peak window tops at 105dB, and a speaker has a rated sensitivity of 85dB@1W@1m, then the calculation for required amp rated power goes like this:
Subtract speaker sensitivity rating from peak window top
105-85=20
divided by three
20/3=6 2/3 (call it 7)
double one that result's number of times
1, 2, 4, 8, 16, 32, 64, 128 (seven doublings of 1)
So about 128W needed to peak the speaker to 105dB. But since this is a peak measure, divide by the square root of 2 (about 1.4) to get average rated power.
128/1.4= about 90w
Note:
The above is an approximate figure, which is fine; in general all values are being pushed to the conservative advantage so as to always have a strong margin. Since 20dB is a linear factor of 100, the math is convenient... 20dB is a jump from 1 to 100, so 20dB above the 1W sensitivity rating is 100W peak and about 70W average.
The speaker's maximum power capabilities need to be compared to the result of this calculation of required power. Some combinations of speaker and amp will not result in a speaker that can withstand the power required to meet the level desired, or to do so cleanly.
5] The source
Note:
All I know is PHONO, and all I know of that is B&O, so I'll just offer that one must do the best they can and defer to others' wisdom.