Hi Pano
Unfortunately I have no markings on my gain knob to refer to and it has no digital readout.
Also, some, most of my sources are analogue and or at hifi level so they make the transition to pro signal level and balanced.
I am not sure where the knob is relative to 0dB, I do know that the .5 V would be about -42 dB from the onset of Voltage limiting at the power amp.
The speakers begin to become dynamically nonlinear at around 56 VRMS using pink noise (requires an amplifier than can swing 112VRMS).
That is in a test where the RMS power is raised every 15 min until the first frequency falls -3dB relative to the response curve shape at 1W.
I can’t say how loud I use them, usually not very loud at all but there are times when I turn things up, for example, when the mood strikes, (for me) few things compare to Deep Purples BBC live concert from the old days, played louder and cleaner than I could ever dreamed of most of my life with headroom to spare or to play things like the ozone percussion group at a level that sounds like live drums.
You have Altec’s eh….. well I guess I should not be surprised your curious about this stuff, cool!. I grew up using Altec horns in pa in the way old days .
That also reminds me of an important thing, some folks have tube amplifiers and these (many) reach the limit much more gracefully (sonically speaking) than most all SS stuff, some tube stuff even has a little headroom.
Shortly after Highschool I worked at an old tube amplifier company called Grommes Precision, back then, In seriously casual listening tests, i had concluded that a tube amp that had graceful clipping might sound like it was 2-4 or more times larger than the ss amps of the day (as in our 30W tube amp sounded more powerful or nearly as the 125W ss and didn’t go crazy bad when the soa protection engaged with woofers).
Anyway it didn’t dawn on me until now to mention there can be a big difference between how these types act once one reaches the end of the paved road, for one, it’s an ugly sounding brick wall, the other more like bungee chords haha.
I see your in NC, not that far from Gainesville Ga. If you’re interested, I’d like to invite you to the shop to “hear some horns” next time I am down there.
Best,
Tom
Unfortunately I have no markings on my gain knob to refer to and it has no digital readout.
Also, some, most of my sources are analogue and or at hifi level so they make the transition to pro signal level and balanced.
I am not sure where the knob is relative to 0dB, I do know that the .5 V would be about -42 dB from the onset of Voltage limiting at the power amp.
The speakers begin to become dynamically nonlinear at around 56 VRMS using pink noise (requires an amplifier than can swing 112VRMS).
That is in a test where the RMS power is raised every 15 min until the first frequency falls -3dB relative to the response curve shape at 1W.
I can’t say how loud I use them, usually not very loud at all but there are times when I turn things up, for example, when the mood strikes, (for me) few things compare to Deep Purples BBC live concert from the old days, played louder and cleaner than I could ever dreamed of most of my life with headroom to spare or to play things like the ozone percussion group at a level that sounds like live drums.
You have Altec’s eh….. well I guess I should not be surprised your curious about this stuff, cool!. I grew up using Altec horns in pa in the way old days .
That also reminds me of an important thing, some folks have tube amplifiers and these (many) reach the limit much more gracefully (sonically speaking) than most all SS stuff, some tube stuff even has a little headroom.
Shortly after Highschool I worked at an old tube amplifier company called Grommes Precision, back then, In seriously casual listening tests, i had concluded that a tube amp that had graceful clipping might sound like it was 2-4 or more times larger than the ss amps of the day (as in our 30W tube amp sounded more powerful or nearly as the 125W ss and didn’t go crazy bad when the soa protection engaged with woofers).
Anyway it didn’t dawn on me until now to mention there can be a big difference between how these types act once one reaches the end of the paved road, for one, it’s an ugly sounding brick wall, the other more like bungee chords haha.
I see your in NC, not that far from Gainesville Ga. If you’re interested, I’d like to invite you to the shop to “hear some horns” next time I am down there.
Best,
Tom
More data points?
I swapped out my 91dB efficient Snell Type J/III's for 95dB efficient Klipsch KG4.5's. At darned near "make the neighbors complain" level I got 2.45V.
That correlates pretty well with the speakers' published sensitivity specs.
If we accept that 2.45V^2/8 ohms is the power I need for that, then we're talking only 0.75 watt into 8 ohms.
If we accept that 2.45V times 4 is what the 0dBFS peaks will require, then that equals 12 watts into 8 ohms.
I can handle that!
Remember that I have a 12' x 24' x8' room in an apartment building, so "loud" for me isn't very.
--
I swapped out my 91dB efficient Snell Type J/III's for 95dB efficient Klipsch KG4.5's. At darned near "make the neighbors complain" level I got 2.45V.
That correlates pretty well with the speakers' published sensitivity specs.
If we accept that 2.45V^2/8 ohms is the power I need for that, then we're talking only 0.75 watt into 8 ohms.
If we accept that 2.45V times 4 is what the 0dBFS peaks will require, then that equals 12 watts into 8 ohms.
I can handle that!
Remember that I have a 12' x 24' x8' room in an apartment building, so "loud" for me isn't very.
--
Ron: Thanks for the data points. Seems like a 6WPC amp would be the smallest you would want not to clip the 12 watt peaks. Not hard to do. You may like the sound of a bigger amp for reasons that are beyond the scope of this thread (power supplies and such) but an amp that can do a clean 6 watts RMS should keep you out of clipping.
Tom: Thanks for the further info. I'll cogitate on it and see what I can figure.
Thanks also for the invitation. I've actually been wanting to come down there for some time to hear your rig and meet you. Now that I'm back to work full time, it won't be easy, but I'll try!
Tom: Thanks for the further info. I'll cogitate on it and see what I can figure.
Thanks also for the invitation. I've actually been wanting to come down there for some time to hear your rig and meet you. Now that I'm back to work full time, it won't be easy, but I'll try!
Funny you should say that. I happen to have a 6Wpc amp using PP 2A3s. It is the minimum amount of power I've been able to live with. But that's with the Klipsch speakers. The amp does feel a little underpowered into the Snell or Tannoy speakers I have (both rated at 91dB/1w/1m). But they still work well enough for me to enjoy listening to just about all of my music.
To get that figure of 6W per side, did you figure on peaks being -3dB down from 0dBFS? I thought the numbers meant that I'd want a 12W per ch amp, to be able to deliver 12 watt peaks...
--
So Tom gets away with 12 watts and i need 650 !!!!!.....
Tom misunderstood the test. The sine wave played at -12dB isn't supposed to represent anything whatsoever, it isn't supposed to be loud at a given volume setting or anything at all.
Providing he set the volume control using music in his test with the 101dB loudspeakers, then when he played back Pano's test tone of -12dB he ended up with 0.5Vrms, then at that volume setting the amplifier would be called upon to produce 2Vrms given a 0dB digital signal at the source.
In this case he would require (2^2)/8 = 0.5 watts into a resistive 8 ohm load to ensure that his amplifier wouldn't clip due to voltage limitations when called upon to reproduce a 0dB signal at his source for his given volume setting.
I am still waiting for a response from Tom to post 338. But I will expand upon it myself anyway.
Here we're using Pano's sine wave test of -12dB. If under the circumstances set by your volume control the amplifier gives out 2 volts, then to see how many volts the amplifier would give out for a signal 12dB greater in magnitude, or otherwise at digital zero, we have to multiply the signal by 4 and we end up with a value of 8 volts.
Lets say Pano chose to use a different sine wave level initially and now we're using -40dB. The volume control is in the same position as above because this is set initially when listening to music and this hasn't changed. Now we play Pano's test signal through the hifi and we end up with 0.08volts and it's pretty damned quiet, but this isn't a problem. What we want to know is how many volts the amplifier will give out when called on to replay digital zero. In this case however it is 40dB higher and to get a voltage that is 40dB higher we must multiply the measured voltage by 100.
0.08v * 100 = 8 volts again.
The end point of the test hasn't changed except in the way we measure the test tones voltage to arrive at the end point. The -12dB test does not represent a 4x difference between any average to peak level, in the same way that the -40dB test doesn't represent a 100x difference between the average to peak level. What we're interested in with the test is only the absolute maximum that the peak level can achieve for the given volume setting (when the source hits 0dB) and then to see if your amplifier has the required voltage swing to reproduce it.
Edit - I see you edited your previous post and were in fact referring to Ron >.< Either way the rest of my post still stands.
I guess you like to listen at very loud levels a.wayne or have incredibly insensitive loudspeakers!
Last edited:
Is that for me or Tom , we already did our 1812 thingy ...
That was for Tom, once explained you seem to have completely grasped the concept.
4Vrms on a -20dB reference. For a 20+dB gain to bring us up to full scale digital, on a voltage = multiply by a factor of 10 = 40Vrms = (40^2)/8 = 200 watts into an 8 ohm resistive load.
As you know your loudspeakers seem to be a pain to drive however (1ohm above 300hz!
) and are active by the sounds of things. If you're active and have variable gain vs frequency in the way then this complicates things. You must run the test at the point of maximum gain of your active network as this represents the maximum signal level that will get through to the amplifier. You will need to do this for each (different) active channel in your system too.If one uses a single amplifier and has a passive loudspeaker then the gain through the entire reproduction chain (not counting the intrinsic high pass and low passes built into the DAC/amplifiers at the frequency extremes) will be perfectly flat. That is the amplifier will give the same voltage at 100hz as it will at 5khz as it will at 10khz for a given digital reference level. But if you have active gain that varies with frequency you naturally have to pick the frequency of maximum gain for the given amplifier in the system.
Edit - as your system is a pain to drive, then obviously instead of dividing by 8 as above, you will divide by the impedance of your system. You use 4 ohms below 300hz, which would imply 400 watts are required there, but for further up the chain where you've got a 1 ohm load (is this really true?) you're looking at watts into the ridiculous.
Last edited:
@5th Element ...
Currently passive , being driven by one amplifier . As to the concept , I never had a problem grasping it you had an issue grasping that program material will vary and as such i thought it was better for Pano to give both recording and test tone .
Thats all .......
Yes this particlular speaker is an direct drive 4 way ribbon hybrid ( 6 ft 7" tall ) impedance is 1 ohm from 300 ( approx) and .8 @ 20K . The subwoofer system is active below 80 hz, but i did not include it in the test.
I have another version building which i had calculated to be 4 db more efficient if so then we can half that. Admit the imp is redonculous, but so is the sound of ordinary speakers. Actually i can drive them with quite a few amps, of course pretenders fall away , most require fan cooling , a few don't...
Krells love them , Eagles run away and Thresholds sing .....
Currently passive , being driven by one amplifier . As to the concept , I never had a problem grasping it
you had an issue grasping that program material will vary and as such i thought it was better for Pano to give both recording and test tone .Thats all .......
Yes this particlular speaker is an direct drive 4 way ribbon hybrid ( 6 ft 7" tall ) impedance is 1 ohm from 300 ( approx) and .8 @ 20K . The subwoofer system is active below 80 hz, but i did not include it in the test.
I have another version building which i had calculated to be 4 db more efficient if so then we can half that. Admit the imp is redonculous, but so is the sound of ordinary speakers. Actually i can drive them with quite a few amps, of course pretenders fall away , most require fan cooling , a few don't...
Krells love them , Eagles run away and Thresholds sing .....
Last edited:
Tom said...
Now I am going to be very pedantic about this because it's necessary, but Tom says he needed 0.5Vrms with a mid band sine wave to reach a very decent loudness. Where I'm going to get nit picky is this...
1) Did Tom play his most dynamic piece of music at the loudest volume he would ever play it at, THEN, keeping the volume control in the same place, play Pano's -12dB sine wave test and get 0.5Vrms or...
2) Did Tom play only a sine wave and set the level based off of how loud the sine wave was?
If he did 2, which is what he's implied by saying he needed 0.5Vrms with a mid band sine wave to get a decent loudness. Then he's not done the test correctly.
Last edited:
Tom,
what I'm still trying to understand is why are you trying to make this seem so complicated. let me try another way of putting it.
I found that using 0 dBfs sines, my amp is just below clipping at an output voltage of 22 volts, into a 4 ohm resistive load at any frequency (20Hz-18kHz). hopefully it behaves well into my speakers which don't go below 4.4 ohm at any frequency (obviously I won't test that with sines into my speakers).
with the volume pot set at the position that gives 22V at speakers inputs, I found that 99% percent of my recordings sound loud enough for my taste (some even too loud). my normal listening level is at or below 9V.
now, assuming that you're not telling me that I should be listening at higher SPLs (we're talking about my room, my stereo, my ears and my taste after all) what part of the above you disagree with and why?
also, I'd like to add that I still believe that real-time FFT analysis is even more telling of clipping, compared to a temporal view. at the onset of clipping high order harmonics suddenly show and begin to rise at a high rate compared to the fundamental as the clipping advances. for instance when using a 1k tone, with my amp the "just clipping" state is barely visible on a waveform while the FFT suddenly starts to resemble a comb.
and to everyone.
I find what one previous poster said very true. after I got my first serious speakers I noticed that I started to listen at higher SPLs. the first time I became aware of it was when one friend of mine was at my home. he was sitting maybe one meter away from me and I needed to raise my voice (almost shout) so he could understand what I was saying. until I did that neither of us noticed how loud the sound was.
what I'm still trying to understand is why are you trying to make this seem so complicated. let me try another way of putting it.
I found that using 0 dBfs sines, my amp is just below clipping at an output voltage of 22 volts, into a 4 ohm resistive load at any frequency (20Hz-18kHz). hopefully it behaves well into my speakers which don't go below 4.4 ohm at any frequency (obviously I won't test that with sines into my speakers).
with the volume pot set at the position that gives 22V at speakers inputs, I found that 99% percent of my recordings sound loud enough for my taste (some even too loud). my normal listening level is at or below 9V.
now, assuming that you're not telling me that I should be listening at higher SPLs (we're talking about my room, my stereo, my ears and my taste after all) what part of the above you disagree with and why?
also, I'd like to add that I still believe that real-time FFT analysis is even more telling of clipping, compared to a temporal view. at the onset of clipping high order harmonics suddenly show and begin to rise at a high rate compared to the fundamental as the clipping advances. for instance when using a 1k tone, with my amp the "just clipping" state is barely visible on a waveform while the FFT suddenly starts to resemble a comb.
and to everyone.
I find what one previous poster said very true. after I got my first serious speakers I noticed that I started to listen at higher SPLs. the first time I became aware of it was when one friend of mine was at my home. he was sitting maybe one meter away from me and I needed to raise my voice (almost shout) so he could understand what I was saying. until I did that neither of us noticed how loud the sound was.
Last edited:
I think that we really need to be nit picking on this because some readers might get the wrong idea and still be convinced that they need kilowatt amps even with well behaved speakers with medium sensitivities and small/medium listening rooms.
I reread the first posts and I think I know what's going on. in the first post from this thread Pano wrote:
From that you can determine how much power you need.
I think Tom (and maybe others) stopped reading there and hence the confusion. there's simply no other explanation.
Make that three of us.......... I size amps based on dynamic headroom requirements and Pano's way is basically the way I would size a 70V distributed system where fast transients are moot, so with nothing functioning around here to measure, all I can use is a thought experiment.
I'm sitting 1 m away from a 1% eff., nominally 8 ohm speaker [~92 dB/W/m] and my loudest comfortable SPL is 85 dB and I want an amp that will ensure it won't clip on 105 dB peaks.
What do you calculate doing it your way and what formula did you use to calculate the theoretical voltage level at 85 dB?
TIA,
GM
obviously a totally unrelated number because no-one said that the test should be done at an average level of 85 dB. maybe I did mine at 80, maybe I did it at 95, I don't know because I don't have a SPL meter.
I think you misunderstood the test too as it's not about what should be going on but about what is actually going on. all you're saying is "your test is wrong because your listening level is too low".
I've said nothing of the sort! Please don't put words in my mouth!
Regardless, it shouldn't be unrelated! That's the point! Instead of a voltage reading, I've given what it represents. I've already calculated what I believe are close enough values, but I want to see what Pano's way comes up with.
GM
Regardless, it shouldn't be unrelated! That's the point! Instead of a voltage reading, I've given what it represents. I've already calculated what I believe are close enough values, but I want to see what Pano's way comes up with.
GM
Thanks for your response GM this is actually a very good example of what this test isn't about and can hopefully serve as a good example for those who perhaps don't quite get it.
This test doesn't take into consideration your theoretical max SPL requirements, ie requirements written on paper. It also doesn't take into consideration the sensitivity of your loudspeakers as per their specification.
What we're doing here is taking the most dynamic piece of music you have and playing it at the maximum volume you would ever play it at. This basically determines the maximum volume you would ever want your hifi to reproduce given your most demanding piece of music and now we know what that volume is as per the setting on your volume control.
This is rather basic in its idea, no we don't know what that volume is in dB, but we know that with your musical tastes and listening habits that you will never turn your hifi up beyond that setting on the volume control.
Now the one assumption that we are making, for the rest of this test to be valid, is that most people will do their listening on a CD player, SACD player, DVD audio player, PC + external DAC etc. That is a digital source which has a max possible signal level fundamentally baked into its specification, which is 0dBfs. The output of digital sources is usually set to = 2Vrms as per a single ended output when the recording reaches 0dBfs but this isn't necessary for the test to work. What's important is that there's a fixed defined absolute ceiling baked into digital reproduction that you cannot exceed.
Now, we've set the volume control to the maximum that the listener will ever listen to their system at and we've also got a source that has a fundamental maximum amplitude that it can reproduce baked into its specification. This in itself allows us to determine a maximum voltage output on the amplifier, if you will, for their maximum volume setting.
After having set the maximum volume, Pano has provided a sine wave with an output equal to -12dBfs at a frequency that most standard digital volt meters will be capable of measuring accurately. You now play this through your system and measure the voltage that your amplifier outputs with your multimeter. Lets say it equals 5 volts rms.
Pano chose a -12dBfs signal over using a 0dBfs signal to protect your speakers and to ensure that your amplifier isn't going to clip when playing the test signal. We are really interested in the voltage the amplifier outputs with a sine wave played at 0dBfs but this is a tall order for anyone's system when its set to the 'loudest you will ever use it at'. So we use a -12dB signal.
Now this isn't a problem of course because scaling that measured 5Vrms up by 12 decibels, to show what the amplifier would really output given a 0dBfs signal, is a trivial task and is accomplished by multiplying the voltage level by 4. In this case we will end up with 20Vrms.
After having done this we now know, that given your previously set maximum volume setting on your hifi, that when your digital source is replaying the maximum signal that it can ever output, that this translates to the amplifier having to swing 20Vrms.
Most decent hifi amplifiers come with an 8 ohm resistive load specification for the amount of watts they can output. Lets say the amplifier is specified at being able to produce 100 watts into 8 ohms. Knowing this we can back calculate what voltage the amplifier is capable of swinging to produce those 100 watts. In this case it's 28Vrms or 40 volts peak/80 volts peak to peak.
Knowing this we can now see that (providing the amplifier doesn't go into current limiting) the amplifier is perfectly capable of reproducing the 20Vrms maximum that it would ever be called on to reproduce.
From this value you are now of course free to calculate, given the max calculated voltage drop that the amplifier will supply to their terminals, the maximum peak SPL that your loudspeakers are likely to produce, but it is only necessary should you want to know.
The point here is that we've taken the persons system to the max it will ever be taken to and then shown, that when called upon, that their amplifier will not clip when their DAC outputs its maximum signal level equal to 0dBfs in a digital recording.
The reason why Pano created this thread and the poll was to see how many volts (and then potentially watts) people were actually needing their amplifiers to be able to swing to reproduce their most demanding pieces of music at their loudest volume setting.
You see the target figure of 105dB for peaks batted around all over the place as if it's a goal you should want to arrive at, as can be seen however, most people who've done the test are probably not reaching anywhere near 105dB on their peaks. I know I don't even though the system is capable of doing it and capable of doing it cleanly too.
This test doesn't take into consideration your theoretical max SPL requirements, ie requirements written on paper. It also doesn't take into consideration the sensitivity of your loudspeakers as per their specification.
What we're doing here is taking the most dynamic piece of music you have and playing it at the maximum volume you would ever play it at. This basically determines the maximum volume you would ever want your hifi to reproduce given your most demanding piece of music and now we know what that volume is as per the setting on your volume control.
This is rather basic in its idea, no we don't know what that volume is in dB, but we know that with your musical tastes and listening habits that you will never turn your hifi up beyond that setting on the volume control.
Now the one assumption that we are making, for the rest of this test to be valid, is that most people will do their listening on a CD player, SACD player, DVD audio player, PC + external DAC etc. That is a digital source which has a max possible signal level fundamentally baked into its specification, which is 0dBfs. The output of digital sources is usually set to = 2Vrms as per a single ended output when the recording reaches 0dBfs but this isn't necessary for the test to work. What's important is that there's a fixed defined absolute ceiling baked into digital reproduction that you cannot exceed.
Now, we've set the volume control to the maximum that the listener will ever listen to their system at and we've also got a source that has a fundamental maximum amplitude that it can reproduce baked into its specification. This in itself allows us to determine a maximum voltage output on the amplifier, if you will, for their maximum volume setting.
After having set the maximum volume, Pano has provided a sine wave with an output equal to -12dBfs at a frequency that most standard digital volt meters will be capable of measuring accurately. You now play this through your system and measure the voltage that your amplifier outputs with your multimeter. Lets say it equals 5 volts rms.
Pano chose a -12dBfs signal over using a 0dBfs signal to protect your speakers and to ensure that your amplifier isn't going to clip when playing the test signal. We are really interested in the voltage the amplifier outputs with a sine wave played at 0dBfs but this is a tall order for anyone's system when its set to the 'loudest you will ever use it at'. So we use a -12dB signal.
Now this isn't a problem of course because scaling that measured 5Vrms up by 12 decibels, to show what the amplifier would really output given a 0dBfs signal, is a trivial task and is accomplished by multiplying the voltage level by 4. In this case we will end up with 20Vrms.
After having done this we now know, that given your previously set maximum volume setting on your hifi, that when your digital source is replaying the maximum signal that it can ever output, that this translates to the amplifier having to swing 20Vrms.
Most decent hifi amplifiers come with an 8 ohm resistive load specification for the amount of watts they can output. Lets say the amplifier is specified at being able to produce 100 watts into 8 ohms. Knowing this we can back calculate what voltage the amplifier is capable of swinging to produce those 100 watts. In this case it's 28Vrms or 40 volts peak/80 volts peak to peak.
Knowing this we can now see that (providing the amplifier doesn't go into current limiting) the amplifier is perfectly capable of reproducing the 20Vrms maximum that it would ever be called on to reproduce.
From this value you are now of course free to calculate, given the max calculated voltage drop that the amplifier will supply to their terminals, the maximum peak SPL that your loudspeakers are likely to produce, but it is only necessary should you want to know.
The point here is that we've taken the persons system to the max it will ever be taken to and then shown, that when called upon, that their amplifier will not clip when their DAC outputs its maximum signal level equal to 0dBfs in a digital recording.
The reason why Pano created this thread and the poll was to see how many volts (and then potentially watts) people were actually needing their amplifiers to be able to swing to reproduce their most demanding pieces of music at their loudest volume setting.
You see the target figure of 105dB for peaks batted around all over the place as if it's a goal you should want to arrive at, as can be seen however, most people who've done the test are probably not reaching anywhere near 105dB on their peaks. I know I don't even though the system is capable of doing it and capable of doing it cleanly too.
I used 0dB as peak. If you need 12 watts peak, then an amp that can do 6 watts RMS without clipping or bad distortion can give you those 12 peak watts without clipping or bad distortion.
Remember the peaks of a sine wave are 3dB higher than its RMS voltage. That means twice the power (1.42 x voltage).
Oh hang on.....I thought this thread about a simple test everyone with a VM could carry out.
What a fool I am
The OP must have been asking if this was a valid test and how it could be improved
Pano, 5th Element, I thank you for your efforts and, to be honest, cannot believe the keystrokes you have put in just trying to keep this simple test thread on track. Its pretty dead though chaps. Any earnest mind has to battle through 30 odd pages of turd for a just a handful of measurement posts.
To the obtuse, argumentative & downright wilful, your job is done (the thread is basically a wreck for the rest of us) and if you answer this to defend yourself, you merely pin the above to your chest as a badge.
The main measurements provided here have been those of character and I would say Pano & 5th Element come up shiny
What a fool I am
The OP must have been asking if this was a valid test and how it could be improved
Pano, 5th Element, I thank you for your efforts and, to be honest, cannot believe the keystrokes you have put in just trying to keep this simple test thread on track. Its pretty dead though chaps. Any earnest mind has to battle through 30 odd pages of turd for a just a handful of measurement posts.
To the obtuse, argumentative & downright wilful, your job is done (the thread is basically a wreck for the rest of us
) and if you answer this to defend yourself, you merely pin the above to your chest as a badge.The main measurements provided here have been those of character and I would say Pano & 5th Element come up shiny
Last edited: