Originally Posted by robobob
One word: transients!
How many have used a simple oscilloscope (with high bandwidth) to observe an audio signal?
Most audiophile thinking is stuck on the concept of average or RMS power requirements. Real life and real music (using the dynamic range of a live, uncompressed performance as the standard for discussion) is very spiky and unpredictable in levels.
Proper handling of transients requires sufficient headroom! ANY hard clipping or even soft clipping from an inadequate slew rate destroys the dynamics and timbre in short order and takes the ease out of the listening experience. Our ear/brain systems can be easily fatigued by simply placing a low pass filter in the signal path, forcing the brain to try to compensate by "calculating" the missing upper harmonics.
Here is another factor: how dynamic or compressed is the source?
The goal of audio engineering is to reproduce, as closely as possible, the source as captured. That is a clear goal when recording natural acoustic sounds, classical music, jazz, folk etc., which our ear/brains have specifically evolved to process and for which we have massive experience for comparison.
However, when the source is synthetic, as in electonica and the musical genres dependent on electronic processing, the standard can become arbitrary. When the majority of audio sources are A) artificial electronic creations and B) consist of compressed dynamics/transients, the standard for reproduction reduces to: which system flatters the source, in the ear of the beholder!
Also, as pop music has been mastered louder and louder, in order to catch the ear's attention when played through low fidelity sources, there is a more restricted dynamic range inherent in the source. Playing a source such as this through a system with inadequate headroom, MAY not be as noticeable because the system has less dynamic range to reproduce from the source.
As the listening level increases toward reference, going up the power doubling curve from a higher initial (average) level, requires more doubling, with transient levels possibly requiring 10 times the power over the average.
Discarding the need to justify compromising due to budget, what remains is a practical general principle that supplying sufficient power to the speaker to avoid ANY clipping distortion under all conceivable operating conditions
, should be the foundation for powering audio systems. Even if that clipping is hard to identify and of short enough duration that the drivers are not immediately degraded, the negative effect on the sonic illusion is real and discernible.
High efficiency speakers at a given SPL start lower on the power doubling curve but clean speakers also encourage louder levels since they sound effortless!
While most home theaters may not require all of the wattage which Jeff specifies in his literature to reach reference levels, with the bounty of dedicated power amplifiers on the market, why cripple your quality experience with designed-to-minimal-cost receiver amplifiers?
Those claiming that they can't hear the differences: our perceptions of audio and video can be trained.... or not.
Some people never perceived the rainbow effect inherent in single-chip DLP projectors until the unfortunate day when someone showed them how to see it! As a result, many became candidates for different projector technologies.
Similarly, as the quality of your listening room improves with masking, time smearing, room resonances removed, your ear begins to unravel the sonic illusion, perceiving layers previously unheard.