Speaker efficiency vs. power requirements


Recently someone gave me the "math" behind speaker efficiency ratings and power requirements. Although I am not sure if the information below is 100% accurate, it is what I have been told. Can we lay this on the table for discussion and try to resolve this confusing issue once and for all?

0 db is a power ratio of 1. Records and tape have dynamic ranges of 30-40db. To achieve a 30db dynamic range requires a power ratio of 1,000:1 and 40 db requires 10,000:1. So if you assign 1 watt of power to a speaker yielding 90db SPL, you need 1000 watts to deliver a true 30db dynamic range. With digital material we find dynamic ranges of 60-70db requiring power ratios of 1,000,000:1 & 10,000,000:1 respectively. Power amps of 1-10 million watts are not feasible today but the point is, more power offers more dynamic realism. Forget power vs. loudness because that really is not a factor in the overall scheme.
bwhite
I DON'T KNOW WHO YOU WERE TALKING TO BUT BOY ARE THEY CONFUSED. Speaker sensitivity is simply the amount of db's produced as a reference (usually 1 watt at 1 meter). What your amp kicks out are electrical watts (you know the RMS variety) while what's (sorry for the pun) coming out of your speaker are acoustic watts. Even a very powerful 200 watt amp may in fact only have a watt or two coming from a speaker, after the inevitable heat loss from electrical to mechanical. The wattage of your amp must be taken in context with your speaker sensitivity, how far you listen from, what kind of music and how loud? You are also confusing dynamic range with the decibel scale. The decibel scale is a fixed number, 70 db is always 70 db. Dynamic range is just the ratio of a quietest sound to a loudest sound (also called signal to noise ratio). Don't forget that power ratings double with a corresponding 3 db change. Therefore a 200 watt amp has twice the wattage of a 100 watt amp and is only 3 db louder, which is about the limits that you can detect by ear. I have thrown out some basic stuff out there as I am mainly a biology major. I'm sure some physics major or enginneer can add plenty more.
Your math is Ok, but 90 dB is not the starting point for your dynamic range. Dynamic range is the difference between the softest and the loudest, and 90 dB is pretty loud to start with. 120 dB is EXTREMELY LOUD, equivalent to a jet taking off. 140 dB is the threshold of pain. Your million watt amplifier producing 160 dB would destroy your hearing. A more reasonable 60 dB range would be 40 to 100 dB, well within the capabilities of most hi end amplifiers and speakers.
Your conclusion about amplifier requirements is not valid. Your dynamic range numbers for both analog and digital are off. Analog tape can have approximately a 60dB range. I've seen several different numbers regarding digital, but the effective range seems to be 80-85dB. The listening room also comes into play. A typical residential listening room has a background sound level of 50dB (if you live in a rural area far from trains, air routes and highways you could get down to the mid 30s). So if you set you system gain such that the softest sound level is equal to the listening room's background noise level, then you would need to produce a maximum sound pressure of 135dB to fully reproduce the digital medium's dynamic range. Assuming it could go that loud, which is a very big assumption, a 90dB sensitive speaker would require a mere 32-33,000 watts of amplifier power. With its smaller dynamic range, analog sources would only require 128 watts of amplifier power.

In the real world the above number are not particularly relevant. The measured dynamic range of recorded music rarely exceeds 40dB. Pop/rock music is typically in the 10-15dB range. Any intelligent listner rarely would listen at room sound levels above 105dB. A 90dB sensitive speaker only requires 32 watts of amplifier power to produce 105dB levels.
Yes, I think the flaw is starting with the 'efficiency' of 90db, and thinking of this as the base number for dynamic range.

So, a 90db speaker, with 100 watts will do 100 db, and 110 db with 1000 watts. Dynamic range will be from 0 (no current to speaker) to 100 or 110, respectively.

With the logic in the quote above, speaker efficiency is taken out of the equation, as it uses the speaker's efficency as the dynamic range 'floor.'
Your source is ok, I guess, its just said things in a kinda stupid way. This topic has been covered before. Every 10db increase is a doubling in SPL (sound pressure level, and I would call this loudness, colloqually (I know its spelled wrong)) and perceived loudness. 80db is twice as loud as 70db. 90db is 4x as loud as 70db (2x2=4). It takes ten times (10x) the power to double the SPL. So if you can play 90db with 1 watt, then 100db will take 10 watts,110db will need 100 watts, and 120db will need 1,000 watts. Got it. OK. So the first 4 sentences of your source are OK; its just said really weird.

He gets confused on the 5th sentence, or at least I am. If we need 60-70db dynamic range, well, to my knowledge, a bit limited, a 90db efficient speaker already has that dynamic range. It can play 90db with 1 watt so it can do 70db with even less power. What he's done is added the 60-70db onto the 90db to give a dynamic range of 150 to 160db. And from there it appears to be ok mathematically, yes you'll need 1,000,000 watts into a 90db efficient speaker if you expect it to play 150db loud. Thankfully, there are speakers w/ 100db efficiencies and higher, so we can make up for power with sensitivity/efficiency. And there's little reason to go beyond 120db because you'll start getting hearing damage. The 6th sentence is a no-brainer and his first five sentences are a little dumb, it he's just trying to support #6. Just disregard the last sentence (#7). As far as I know, loudness, i.e. how loud the system can play, is dynamic realism, the system can go from really really loud to quite. And the amount of power is in direct relation to how loud it plays once the system efficiency is taken into account. Hopefully, I'm not wrong on anything or I'll have confused you more trying to unlearn, relearn, unlearn, etc.