by Pablo Bellinghausen –
With so many small venues and bands buying their own public address systems nowadays (PA systems for short), loudspeakers and the power amplifiers required to drive them are one of the more common products sold in pro audio. However, power ratings are also one of the most commonly misunderstood subjects, with people quoting and basing their purchases on single ratings without further specification; this can easily translate into broken equipment if one is not careful.
Not all watts are alike
Although the power rating (measured in watts, or W) is of course the most important figure when looking at power amps, the power figure by itself tells you little about the real sound level the system can safely achieve. Power ratings will change enormously depending on the impedance and the calculations used to achieve them.
For example, the Behringer iNuke NU3000 claims to be a “3000-watt amp”, but if one reads the fine print in the product specs, these are the real specifications:
2 × 1500 W peak power at 2Ω
2 × 440 W peak power at 8Ω
Without going much into detail, this means that said “3000-watt”amp is actually ideally paired up with a couple of standard PA speakers having peak ratings around 440 W; if we wanted to guarantee clean power for a pair of speakers (whilst being careful not to overdrive them) we could use this amp on something like the RCF M601, which is a small PA speaker quoted as a peak power handling of 400 W. The M601 is meanwhile quoted as a “100-watt speaker”because RCF uses what is called the “AES power”measurement to advertise their speakers – we’ll go over what that means in a bit.
Understanding Audio Power
An electrical audio signal is a voltage oscillation; one can refer to the maximum instantaneous value that the signal can be, called peak voltage, or to the average area occupied by the signal, derived from the “root mean square”voltage calculation (or RMS). For the simplest of audio signals, a sine wave, the RMS voltage equates roughly to 70% of the peak voltage:
Going from voltage (in volts) to power (in watts) to decibels (just a value ratio) to actual physical loudness (in dB SPL) involves logarithmic operations and will depend on many different factors (from the sensitivity and resistive qualities of a speaker to the acoustics of the room), but there are a few online calculators that can help out, such as the one by Crown Audio:
Without going into details, the peak voltage is used to calculate the peak power and the RMS voltage is used to calculate the continuous average power (“RMS power”is a very common but incorrect way to call it). For a sine wave, the peak power is twice the continuous average power.
The RMS value is the one that would relate more with the actual perceived loudness of the PA system; however, depending on the type of sound being played for a given peak power rating, certain signals can be played louder without distorting than others. Here is an example of a commercially mastered pop song, as measured close up to the speaker inside a live venue (played back very loudly!):
Because of the natural ups and downs in a musical signal, the difference here is about 9 dB between RMS and peak voltage. This translates into a peak power that is on average about 8 times the continuous average power.
This is a close-up recording of a live performance that includes very loud drums, electric bass and guitar through their respective instrument amps:
Uncompressed live drums have very loud transients (short level peaks when the stick hits the drum or cymbal). Here, the loudest one is about 14 dB louder than the average sound level. This translates into a whopping 25 times more peak power than the continuous average, and about 3 times more power than required for pre-recorded music at the same average volume.
The difference between the average level and the “safety buffer”left over for transients in a system is called headroom, and will make a lot of difference in live music. Headroom is expensive (twice the power will only give you about a 3 dB increase in level) but is the only way to guarantee a clean reproduction of the full signal.
This is why live venues must have a more powerful system if they are miking up live drums than a basic pre-recorded music playback system, which won’t have those transients. Many live venues that skimp out on their sound will just let those spikes overdrive the PA; a few clipped transients here and there will barely be noticeable and won’t damage any equipment, but this is of course not ideal. When amplifying live music including drums, for the best possible sound the power ratings should be about 3 times higher than for pre-recorded music at the same perceived volume.
Additionally, a high-quality limiter can shave off some of the larger spikes and squeeze in a few extra decibels. Limiters are extremely fast-acting compressors that can be very effective at taming transients, therefore removing the need for extreme headroom in live music situations, but they will almost invariably harm the sound unless used sparingly. Unlike in live sound, limiters are not very useful for pre-recorded music since it is already heavily compressed, and will sound dreadful as soon as the signal reaches the limiter’s threshold – they should only be used as a last-resort safety measure for careless DJs or unmonitored, unexperienced users.
Types of power specifications
Although there are several other factors, the main killer for both speakers and amps in normal conditions is actually heat; any signal going through circuitry will heat up the components somewhat, and both amplifiers and speakers are built to withstand thermal levels higher than their ideal operating value, but once you’ve reached that level the sound will already be heavily distorted. There will always be a certain amount of harmonic distortion on the output no matter the volume, but the total harmonic distortion (THD) will slowly start to creep up the louder the sound, until the signal reaches a maximum level (called clipping point in amps), after which distortion becomes incredibly obvious.
The peak power rating of a speaker or amp is the loudest possible sound it can reproduce, for very short periods of time only. This measurement will usually be made with a 1 kHz sine wave for a few milliseconds and will involve a certain amount of distortion, which better manufacturers will quote in their full specs. For example, a peak power measurement at 0.1% THD will be lower than a 10% THD measurement for the same amp. In live events the ear can hardly detect a THD lower than about 3%, but with mastered material in carefully controlled quiet environments like music studios it can be lower than 1%.
The important thing to take into account is that the peak power is the highest guaranteed power for (mostly) clean transients and is therefore the measurement to use for headroom rather than for constant power; the amp or speaker will not be able to withstand the peak power for more than a few milliseconds.
Continuous average power (“RMS Power”)
This is the real “audio power” in the sense of the amount of energy being produced. Since at its most basic level the amount of energy is what gives you the amount of heat, the RMS value on a speaker or amp is really a sort of thermal safety rating. This means that a 250 W RMS rating on a product means it is guaranteed to play back a signal of about 250 W on average for a long time without overheating.
For the best sound quality, the safety RMS value should seldom or never be reached: if the signal is very dynamic and has a lot of transients that go above the peak power, then those will be clipped even if the signal is below the RMS value and therefore the components themselves will (within reason) handle the distortion without breaking. Conversely, when playing back very compressed signals like dance music, reaching audible distortion in pro equipment means that some of the components are most likely already working past the safe RMS rating and therefore overheating somewhat.
The RMS rating can be measured in many different ways; for example, some companies will measure it by using a 1 kHz sine wave for a few seconds. In the case of an amp (in which the circuitry will work fine almost until the clipping point), the “safe” RMS rating would be about 1.5 times smaller than the peak power. The best companies will do their tests using pink or white noise (which has more dynamics, more bass and more treble than just a sine wave) for a couple of hours; in this case, the properly-measured RMS value in speakers is usually about 4 times smaller than the peak value.
Power rating standards
Because of the large variation in all of these measurement types, there are several “official” standard measurements that can often be seen in spec sheets nowadays, the most common ones for speakers being IEC268-5 and AES2-2012, sometimes abbreviated to IEC and AES. Both tests involve playing back pink noise for a while and checking that there is no damage to the speaker. The tests are slightly different, with the AES one sending separate signals to sub and tweeter, and using a more dynamic signal than the IEC test, but for a shorter period of time (12 dB between RMS and peak instead of 6 dB, but 2 hours instead of 100 hours); in PA speakers both will most often be about 4 times smaller than the quoted peak rating.
The AES test is the one RCF uses for their speakers, which explains why the M601 mentioned above has an “AES power” rating of 100 W and a peak power of 400 W.
In our next instalment we will take a look at how to safely choose and match equipment by using power, sensitivity and impedance ratings for the best possible results, and to what happens when the basic wiring and operating precautions aren’t taken (hint: nothing good!)