In Reply to: Re: Excellent document, thanks Michael. Further questions. posted by Martin419 on December 4, 2005 at 01:56:48:
It's quite hard to do an actual power plot by frequency as there is a risk that the tweeters can get damaged by clipping. Most people who do these measurements tend to stay at low frequencies only.Hi Fi Choice is right - in fact most of the time, less than one watt is required. If you assume a speaker sensitivity of 85 dB/1w/1m@8ohm, it means that speaker is able to generate 85 dB SPL if you supply just 1 watt of amp power at 8 ohm and you listen 1m away. 85 dB is pretty loud, most people would listen below 80dB on average.
But we are talking about peaks, not averages.
Basically a doubling of power is required for every 6dB increase in volume. so, whilst 1 watt is required for 85dB, 10 watts are required for 91 dB, 100 watts for 97 and 1000 watts for 103dB, and 10000 watts is required for 109dB!
Uncompressed music can have peaks up to 25dB above "average level".
So if you assume an average listening level of 85 dB, then you need 85+25=110 for peak.
Thus, in this particular situation, you do need more than 10000 watts for 110dB peaks! but that's not taking into account the effect of wall reflections in a contained space.
However, since you have previously stated that your room do not have strong reflections, presumably you cannot then take advantage of room reinforcement :-)
Of course, that 10,000 watts is for ALL channels. if you have multiple channels, then that amount of power needs to be distributed across the channels depending on relative levels. But be careful when doing decibel arithmetic, for example two speakers generating 50dB each does not give a total sound pressure of 100dB, the actual figure is 53dB (50dB+50dB = a 3dB increase)
On the flip side, most of us cannot detect minor instances of clipping, so the reality is provided you are happy that your peaks will occasionally clip, you can make do with a lot less watts. But it's a compromise that you have chosen.
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Re: Excellent document, thanks Michael. Further questions. - Christine Tham 22:41:47 12/04/05 (10)
- Sorry - Christine Tham 22:45:31 12/04/05 (9)
- Re: Sorry - Dave Kingsland 09:33:54 12/05/05 (2)
- Oops :-) you're right - it's watts rather than power (nt) - Christine Tham 11:11:59 12/05/05 (1)
- You don't need 10000W, just 256W - Jack100 16:05:30 12/16/05 (0)
- "Root-Mean-Square" (RMS) versus "peak-to-peak" wattage . . . - Martin419 08:15:29 12/05/05 (5)
- You're partially right - Christine Tham 11:38:07 12/05/05 (4)
- Re: You're partially right - Martin419 12:31:30 12/05/05 (3)
- Re: You're partially right - Christine Tham 15:04:45 12/05/05 (2)
- Re: You're partially right - Martin419 02:31:46 12/06/05 (1)
- Re: You're partially right - Christine Tham 15:39:16 12/06/05 (0)