I don't. It may be easier to start with a speaker that is reasonably flat but a speaker that is flat will sound bright and bass shy under most circumstances. Most people have no idea what their system is doing and their opinion is based on assumptions that are usually dead wrong.
My system is measured and graphed on a computer. Filters are generated that make the systems performance perfectly flat as a starting point. Then response curves of my design are overlaid to make the system do what I want. The basic curve is boosted 6 dB at 20 Hz and down 12 dB at 20 kHz. On top of this there is dynamic loudness compensation which follows the Fletcher Munson curves based on volume. This assures that the tonal balance I like in my room is the same regardless of volume. After this is the subwoofer crossover which I can modify on the fly if I desire.
The problem is that the right tonal balance changes with volume. In a standard set up each record or file has a "right" volume. A record that was mastered for high volume will sound dull and bassless at low levels prompting the listener to think it was a bad recording. Records that were mastered for lower volumes will be shrill at higher volumes making all the women in the room extremely angry. Tone controls help a little. But they and standard loudness compensation are just not l flexible enough on top of the fact that the listener really has no idea what is going on. When you see what your system is doing on the computer and integrate that to what you are hearing you learn pretty quick what making specific changes will do. This is an aid of immeasurable benefit. This alone justifies digital audio to the point where I gladly digitize my phono amp so it can play on the same ball field.
I have measured several systems. The reason Wilson's are so... comfortable is that they are engineered with a notch at about 3 kHz which takes the sting out of female voices and violins. I suspect other companies do this as well.
My system is measured and graphed on a computer. Filters are generated that make the systems performance perfectly flat as a starting point. Then response curves of my design are overlaid to make the system do what I want. The basic curve is boosted 6 dB at 20 Hz and down 12 dB at 20 kHz. On top of this there is dynamic loudness compensation which follows the Fletcher Munson curves based on volume. This assures that the tonal balance I like in my room is the same regardless of volume. After this is the subwoofer crossover which I can modify on the fly if I desire.
The problem is that the right tonal balance changes with volume. In a standard set up each record or file has a "right" volume. A record that was mastered for high volume will sound dull and bassless at low levels prompting the listener to think it was a bad recording. Records that were mastered for lower volumes will be shrill at higher volumes making all the women in the room extremely angry. Tone controls help a little. But they and standard loudness compensation are just not l flexible enough on top of the fact that the listener really has no idea what is going on. When you see what your system is doing on the computer and integrate that to what you are hearing you learn pretty quick what making specific changes will do. This is an aid of immeasurable benefit. This alone justifies digital audio to the point where I gladly digitize my phono amp so it can play on the same ball field.
I have measured several systems. The reason Wilson's are so... comfortable is that they are engineered with a notch at about 3 kHz which takes the sting out of female voices and violins. I suspect other companies do this as well.