Well I own two signal generators because I didn't know that ordinary DVMs like my old Fluke 77 don't read high frequency voltages accurately. My thrift store FETVM, which has a 3V scale, tells me my ancient Lafayette tube piece is dead flat to 20K. Redundancy counts when results matter; assumptions not so much.
I built a bass only speaker and amplifier to match up with my full range satellites (AE1s).
I set up the bass levels to suit the relative sensitivities of the speakers and such that music and spoken voice and singing sounded "normal".
I was surprised to find that the bass only speaker crossed in at 150Hz was seeing an average voltage ~20dB below the average level at the satellites. Yes, 180mVac to 1.8Vac for a total stereo (electrical) output power of ~0.8W average. It was pretty loud. I would not be surprised if someone were to tell me the regular transient peaks exceeded 50Wpk, But even that would not be a problem, I had 4 amps running in the stereo setup with a potential 540W, or approximately +10dB of further overhead for the extreme (and rare) transient peaks.
BTW,
It did not sound loud, until one tried to speak !!!
I set up the bass levels to suit the relative sensitivities of the speakers and such that music and spoken voice and singing sounded "normal".
I was surprised to find that the bass only speaker crossed in at 150Hz was seeing an average voltage ~20dB below the average level at the satellites. Yes, 180mVac to 1.8Vac for a total stereo (electrical) output power of ~0.8W average. It was pretty loud. I would not be surprised if someone were to tell me the regular transient peaks exceeded 50Wpk, But even that would not be a problem, I had 4 amps running in the stereo setup with a potential 540W, or approximately +10dB of further overhead for the extreme (and rare) transient peaks.
BTW,
It did not sound loud, until one tried to speak !!!
I haven't gotten around to re-testing with the 120hz tone that you posted. Perhaps I will have time for that on the weekend. In the meantime, I have setup my bedroom system which uses 8ohm full range drivers with a sensitivity of 84.5dB and it is driven by the Lepai TP2020A+ (which is, realistically, a 5 w/chan. into 8ohms amplifier).
I must say that this is more than enough power for well beyond "reasonable bedroom SPL", and I doubt that I am using a whole watt. I will test this with the 120hz tone to verify.
Regardless of how the test turns out, I can only conclude that with very high efficiency drivers, even including overhead for transients (the meaning of which is debatable, depending on what type of music you are into) this little "5 watt" amp is more than enough power for the vast majority of home audio applications. I could certainly ditch my 50w/chan. integrated from my main living room sound system and replace it with this little Lepai and I wouldn't be lacking watts. Not even close.
I use speakers with a sensitivity of about 96db/w/m . The room is 13x17x8 ft. with carpeting, curtains and a few stuffed chairs. I listen to a wide range of music types (Jazz, various types of rock, swing bands, country and clasical) at levels that are generally well below 'live' spls. I measured the power delivered to the speakers at typical listening spls. These were typically in the 70-80 db range at about 12 feet from the speakers, measured with a Radio Shack meter set to 'fast' and 'C' weighted. The meter measurements indicate only about 3 db of loss from 1m (in front of the speakers) to the listening position. I assume that the meter generally underestimates short term, high amplitude peaks by 10+ db, so that peaks were actually in the 85-95 db range. Power was monitored with a 2-channel oscilloscope. I did not bother to correlate the meter readings to the scope measurements, as the goal was to measure levels and to check for clipping while enjoying musical favorites at typical listening spl's. Scope readings indicated average power levels in the range of 0.05-0.1W. Peak levels were in the 1.2-1.6W range. No clipping was observed, and none was audible. I sleep comfortably knowing that 2wpc adequately meet my needs.
Regards,
Bob
Regards,
Bob
Would try those measurements... but baby girl is going at about 105dB at the moment... will need to figure out why she is crying then only what the dmm is reading...
Oon
LOL
Girls are loud! For fun the other day I asked my daughter to scream as loud as she can while I monitored with my SPL meter from about 2 feet away. She hit 114dB!!
My ears were ringing more than usual after that 🙂
We've got the crying one next door.........We also have the many barking dogs here..Will I be measuring Woofer output???
________________________________________________________Rick........
________________________________________________________Rick........
nope - is that yet another credential we need to justify our enjoyment? sorry for the pi$$y attitude, but to paraphrase an old Linn advert line "if it sounds good (to you), it is"
no audio reproduction system is (will ever be in our lifetimes?) "perfect" - if the inevitable compressions, distortions, etc created by theoretically inadequate match of sensitivity and power don't detract from one's musical enjoyment, WTF does it matter?
cranky rant off
Hmmm... so both dads are in danger of developing Tinnitus if daughters take up the habit of screaming at them... 🙂
Last edited:
To me that's a classic sign or test of a system that is "good" to listen too.
That tallies well with my observations.
Quoted from other thread
"I recently got my hands on a sound level meter (C weighted) and was (pleasantly) surprised at just how low the levels were that I normally listen too. At the listening position around 70db (average level) is plenty for normal listening. The meter shows around 75db peaks (125ms response)"
I already have tinnitus, but it was worse for about 1/2-hour after she hit 114dB 🙂
Luckily, she doesn't make a habit of screaming at me. Not yet, anyway. 🙂
then there'll be the stage when any communication would be appreciated (say from 12-20 or so)
To fuel the debate, I have a link to the scan of an old article : Puissance quand tu nous tiens !
Sorry it's in French. Pano can go through. For the others I will try to help :
they say that many guys are very happy to put a multimeter across a speaker, measure ridiculous tensions when they listen the thing already loudly (must remind you something in this thread). Then they jump to the conclusion that in most of the cases 3 watts are more than enough...
But this simplistic procedure ignores the transient nature of music. A 30 W amp (they don't even dare to measure with say a 300B) will be compared with a 75 w, both at an average listening level of 85 dB provided by a comfortable 100 dB/1W/1m VOT.
But this shows clipping and compression already at this level, the small amp delivers 15 V when the bigger one reaches 20 V on the same peak. They evoke also the responsability of the PSU that is not fast enough in many cases. . For them all this results in the analytic quality of the amp. Bigger and faster is better.
The author was a good friend of Hiraga and a big contributor of the "L'Audiophile". For the lazy guys, I join a picture : with a same basis level, the peaks are indeed much higher with the big amp. Of course it's not a girl and guitar extract.
Now, we can also argue that their "small amp", a Radford, was not top class, but this study goes in the same direction than a.wayne posts.
Sorry it's in French. Pano can go through. For the others I will try to help :
they say that many guys are very happy to put a multimeter across a speaker, measure ridiculous tensions when they listen the thing already loudly (must remind you something in this thread). Then they jump to the conclusion that in most of the cases 3 watts are more than enough...
But this simplistic procedure ignores the transient nature of music. A 30 W amp (they don't even dare to measure with say a 300B) will be compared with a 75 w, both at an average listening level of 85 dB provided by a comfortable 100 dB/1W/1m VOT.
But this shows clipping and compression already at this level, the small amp delivers 15 V when the bigger one reaches 20 V on the same peak. They evoke also the responsability of the PSU that is not fast enough in many cases. . For them all this results in the analytic quality of the amp. Bigger and faster is better.
The author was a good friend of Hiraga and a big contributor of the "L'Audiophile". For the lazy guys, I join a picture : with a same basis level, the peaks are indeed much higher with the big amp. Of course it's not a girl and guitar extract.
Now, we can also argue that their "small amp", a Radford, was not top class, but this study goes in the same direction than a.wayne posts.
Attachments
Last edited:
"But this shows clipping and compression already at this level, the small amp delivers 15 V when the bigger one reaches 20 V on the same peak. They evoke also the responsability of the PSU that is not fast enough in many cases."
The images do indeed show higher amplitude peaks for the more powerful amplifier, even though the smaller amp is not clipping. IME, some amps do sound more dynamic and this can be due to power supply differences. This could just as well have turned out opposite, with the smaller amp having higher amplitude peaks than trhe more powerful amp, both below clipping. My SET sounds obviously more dynamic then my Luxman ss, though it has less than 1/10th the power.
Power needs scale up rapidly (log function!) as the listening spl goes up. And it's always possible to seek out the most challenging music for this testing. Many examples can be found published or on the web. I have no trouble believing them, like the one above. Nonetheless, I need only about 2wpc for unclipped power to my speakers for the vast majority of music that I listen to, and at the spls I employ. YMMV.
There are many other system performance variables that can be as important sonically as accurate reproduction of spl peaks, like the dynamic aspects described above, tonal balance, phase, amplitude response and many others. People tend to place more or less emphasis on these. For some, the uncompressed reproduction of the rarest, fastest, highest amplitude peaks is of paramont importance. For me, the dynamic behavior of the amp below clipping levels has a substantial influence on musical enjoyment.
What is important is that we all enjoy music through our sound systems.
Bob
The images do indeed show higher amplitude peaks for the more powerful amplifier, even though the smaller amp is not clipping. IME, some amps do sound more dynamic and this can be due to power supply differences. This could just as well have turned out opposite, with the smaller amp having higher amplitude peaks than trhe more powerful amp, both below clipping. My SET sounds obviously more dynamic then my Luxman ss, though it has less than 1/10th the power.
Power needs scale up rapidly (log function!) as the listening spl goes up. And it's always possible to seek out the most challenging music for this testing. Many examples can be found published or on the web. I have no trouble believing them, like the one above. Nonetheless, I need only about 2wpc for unclipped power to my speakers for the vast majority of music that I listen to, and at the spls I employ. YMMV.
There are many other system performance variables that can be as important sonically as accurate reproduction of spl peaks, like the dynamic aspects described above, tonal balance, phase, amplitude response and many others. People tend to place more or less emphasis on these. For some, the uncompressed reproduction of the rarest, fastest, highest amplitude peaks is of paramont importance. For me, the dynamic behavior of the amp below clipping levels has a substantial influence on musical enjoyment.
What is important is that we all enjoy music through our sound systems.
Bob
Exactly, the subject is vast and the example above is not innocent in music or amps choice.
But the mathematics behind have no feeling. With a limited voltage output say of 5 volts, for reproducing without compression a 40 dB peak as above, it will be difficult to listen to this kind of music at a comfortable level (I don't even speak of realistic level).
But the mathematics behind have no feeling. With a limited voltage output say of 5 volts, for reproducing without compression a 40 dB peak as above, it will be difficult to listen to this kind of music at a comfortable level (I don't even speak of realistic level).
I bet your 2wpc SET amp can actually produce at least 30wpc for a few milliseconds.
You are right about not needing much AVERAGE power. I have a maybe 8W guitar amp with a 105dB speaker. If I turn it up I can be heard inside the house across the street.
- Status
- Not open for further replies.