Hi Pano, all
I tried to explain why the sine test has a tenuous connection to what is required for some kinds of program material but fine for others.
The issue was supposed to be what is required to drive the speakers, that includes what is hard for them, not just compressed music or tones ..
I tried to explain that one’s impression of loudness is not necessarily an indicator of producing all the signal or strongly tied to what one measures while what one measures IS what the system is doing.
I tried to explain (since some folks are tempted to try to measure ) that the meters people are likely to used if they don’t have real test equipment HAVE a frequency response which may or may not allow them to be accurate other than around 60Hz or with changing signals.
Also, your ears have nothing like flat response so for a given pitch and actual loudness, the apparent loudness can vary tremendously (over a million to one) and only sounds as loud as it measures in the upper mid band (2-4KHz, look at the equal loudness curves).
I have suggested that in order to tell if one has enough power, that the BEST TEST is one examining the actual Voltage signal going to the speakers at a level your happy with, with the program material you like to listen to and look for signs of limiting with an oscilloscope..
Honestly, I don’ see why this is a bad idea here, this kind of test has revealed many problems in larger sound systems too where gain structure (not just amp limits) can be an issue too..
In a effort to illustrate that the peak electrical power or Voltage required and the apparent loudness ARE NOT strongly tied, I suggested two kinds of artificial test signals which would clearly illustrate that, in the case of the pink noise signals with various music like peak to average ratios, these if played back at the same apparent loudness, requires as much as tens thousands of times more peak power to be the same subjective loudness .
A sine wave, requires even less peak power for a given apparent loudness than the 6dB crest pink noise and that more compressed than any music one can find..
All this may seems like an excessive focus on the dynamics of the signal but that’s what I wish to preserve /reproduce.
While it was a mistake to contribute to this thread, I am not asking people to believe what I am suggesting, just to look using tools that will show the engineering reality at the root of this.
Remember as in politics and many other areas, It is often the
“Dunning – Krueger effect” that makes some things seem simple that are actually not.
Best,
Tom
I tried to explain why the sine test has a tenuous connection to what is required for some kinds of program material but fine for others.
The issue was supposed to be what is required to drive the speakers, that includes what is hard for them, not just compressed music or tones ..
I tried to explain that one’s impression of loudness is not necessarily an indicator of producing all the signal or strongly tied to what one measures while what one measures IS what the system is doing.
I tried to explain (since some folks are tempted to try to measure ) that the meters people are likely to used if they don’t have real test equipment HAVE a frequency response which may or may not allow them to be accurate other than around 60Hz or with changing signals.
Also, your ears have nothing like flat response so for a given pitch and actual loudness, the apparent loudness can vary tremendously (over a million to one) and only sounds as loud as it measures in the upper mid band (2-4KHz, look at the equal loudness curves).
I have suggested that in order to tell if one has enough power, that the BEST TEST is one examining the actual Voltage signal going to the speakers at a level your happy with, with the program material you like to listen to and look for signs of limiting with an oscilloscope..
Honestly, I don’ see why this is a bad idea here, this kind of test has revealed many problems in larger sound systems too where gain structure (not just amp limits) can be an issue too..
In a effort to illustrate that the peak electrical power or Voltage required and the apparent loudness ARE NOT strongly tied, I suggested two kinds of artificial test signals which would clearly illustrate that, in the case of the pink noise signals with various music like peak to average ratios, these if played back at the same apparent loudness, requires as much as tens thousands of times more peak power to be the same subjective loudness .
A sine wave, requires even less peak power for a given apparent loudness than the 6dB crest pink noise and that more compressed than any music one can find..
All this may seems like an excessive focus on the dynamics of the signal but that’s what I wish to preserve /reproduce.
While it was a mistake to contribute to this thread, I am not asking people to believe what I am suggesting, just to look using tools that will show the engineering reality at the root of this.
Remember as in politics and many other areas, It is often the
“Dunning – Krueger effect” that makes some things seem simple that are actually not.
Best,
Tom
I thought that was pretty clear, maybe it wasn't.
You were clear and when I originally read the thread starting posts I understood it. After all of this discussion though I forgot about that and started to think that the voltage in the poll was the true maximum 0dBfs figure. It's entirely my own fault I will admit.
Tom will you please stop and READ the thread until you understand what Pano's test is actually doing. The sine wave has absolutely no connection to music and isn't trying to either, it is not trying to represent some kind of average that you will find in program material.
retrograde discussion
It does appear that the thread is moving inexorably towards invoking Godwin's law...
seems to me, the uncontrolled variable here being contested is "as loud as you'll ever listen"... to what???
Personally, I don't often listen to sine waves, so the application of the test may have limited merit towards drawing the stated conclusions.
John L.
It does appear that the thread is moving inexorably towards invoking Godwin's law...
seems to me, the uncontrolled variable here being contested is "as loud as you'll ever listen"... to what???
Personally, I don't often listen to sine waves, so the application of the test may have limited merit towards drawing the stated conclusions.
John L.
Again read the thread -.- This isn't uncontrolled, you actually listen to your most dynamic and system challenging piece of music as loud as you would ever listen to it.
Again you need to read the thread, no one is asked to listen to a sine wave, not ever, not even for a pico second.
no misunderstanding
That's "sort of a control" I guess... having a poll with individuals' "scores" somehow to me implies that conclusions can be drawn across all those participating about the adequacy of amplification at realistic levels w/o distortion based on the results. This seems unwarranted, since the "control" is relative to only the system under test, not across systems. Each persons "loud as you'll ever listen" and "most demanding music" hardly qualify as absolute references across all those taking the test. I agree, in a simplistic view, it tells you something, I'm just not sure about some of the conclusions being drawn.
But the test tone to peak values stated imply this is representative of normal listening venues and media... my belief is that others posting here are trying to elucidate the limitations of the test and expand on them... to no avail I guess.
Btw: I've read the thread, thank you, several times. Even took the test, by golly!!
John L.
That's "sort of a control" I guess... having a poll with individuals' "scores" somehow to me implies that conclusions can be drawn across all those participating about the adequacy of amplification at realistic levels w/o distortion based on the results. This seems unwarranted, since the "control" is relative to only the system under test, not across systems. Each persons "loud as you'll ever listen" and "most demanding music" hardly qualify as absolute references across all those taking the test. I agree, in a simplistic view, it tells you something, I'm just not sure about some of the conclusions being drawn.
But the test tone to peak values stated imply this is representative of normal listening venues and media... my belief is that others posting here are trying to elucidate the limitations of the test and expand on them... to no avail I guess.
Btw: I've read the thread, thank you, several times. Even took the test, by golly!!
John L.
Nope!
You turn up a favourite track 'till it's loud enough, then don't touch the volume again. Replace the music cd with the test tone, then hit play and measure the voltage output. You can even unplug the speakers at that point, it doesn't matter. There is nothing implied in the different levels, just a mathematical convenience.
Which is the entire point of this test. It is supposed to give the individual a fairly accurate representation for their amplifiers output voltage requirements in their system with their music under their listening condition - this is all. It isn't supposed to work across systems.
Although if one did to this test and found out their voltage requirements they could use their speakers sensitivity rating to calculate what their rough peak levels are and then could size an amplifier for a system with different sensitivity speaks, but once again, this would only be relevant to that individual. This is very important to that individual mind you, but not at all important to someone else.
No. Just no. There is no other way to put it. The test tone to 'peak' value means nothing whatsoever. Pano chose to use a -12dBfs test tone. He could have chosen to use a 0dBfs test tone which would have given zero room for any 'peaks' he could have also chosen to use a test tone that was -40dB, the end result would have been exactly the same. Sure the numbers entered into the poll would have been different, but once you perform the necessary calculations to figure out how much voltage swing your amplifier requires for 0dBfs (which is what the ultimate point of the test is), then the end results would be identical.
Then you need to read it again because you don't get it.
My perception is that there are no limitation to the test
It is easy to be carried out
the low frequency sine wave are ideal for DM
Wayne so far the only one that may have walid point of using a -20 db and multiply by 10 but that would take reading to the millivolt range and not everibody got a 5 digit meter.
I know for shure now that with only 2 dB head room my amp will probably clip
regretably we got draged in to I know best argument.
Could please moderator delete all my post here
What else do you want?
Imo this is spoiling Pano briliant IDEA for all
What many seem to miss is this... and I'm just trying to think of a different way to put the main point across here.
If your amp clips at 10.00000 volts rms output and you play and adjust the volume control on your amp such that a 0db sine wave from the CD player gives 9.99999 volts rms then no signal from that CD player into that amp at that volume setting you have just set can ever drive it into clipping.
That in essense is what this whole thread is about.
Dynamic range of the signal/music off CD doesn't come into it.
So turning that on its head, the quietest bits on the CD can be anything you like below 0db. We can have a dynamic range of max output (0db) down to the noise floor (say -90db). It's not relevant
Or, and again this is trying to give some other slant on this that puts the point over... think of it as a DC question. 0 db (bit "65536" of our 16 bit CD spec) equals some maximum voltage. Every other "bit" is a value (voltage) less than that maximum. So when you have amplified that maximum "bit 65536" (its voltage) to the most you want it then "it aint gonna get no bigger folks" no matter what CD you put in.
If your amp clips at 10.00000 volts rms output and you play and adjust the volume control on your amp such that a 0db sine wave from the CD player gives 9.99999 volts rms then no signal from that CD player into that amp at that volume setting you have just set can ever drive it into clipping.
That in essense is what this whole thread is about.
Dynamic range of the signal/music off CD doesn't come into it.
So turning that on its head, the quietest bits on the CD can be anything you like below 0db. We can have a dynamic range of max output (0db) down to the noise floor (say -90db). It's not relevant
Or, and again this is trying to give some other slant on this that puts the point over... think of it as a DC question. 0 db (bit "65536" of our 16 bit CD spec) equals some maximum voltage. Every other "bit" is a value (voltage) less than that maximum. So when you have amplified that maximum "bit 65536" (its voltage) to the most you want it then "it aint gonna get no bigger folks" no matter what CD you put in.
With all due respect, if you completely understood how the test works then you wouldn't have made the comments you made in post 424.
reading trist thread reminds me of something that an older friend told me a long while ago. it is a truism, but it's representative to what's going on here. it goes something like "you can try and force someone to see the truth, but in the end the actual process is a thing oneself does alone. and there is no other way".
you can try and explain over and over and over and in the end you find that some didn't even read the thread because they saw the words "sine wave" being used in the context of audio. which is the big no-no of the subjectivist audiophile.
that is fine with me but I see Pano's and 5th's attempts to explain this failing over and over because people simply refuse to try and understand what this is about.
I can't help but notice the psychological side of this
auplater, 5th explained it very well many times. he didn't simply tell you "no, you're wrong". you need to try to understand that the reason that -12dB sine was chosen was ONLY to protect your speakers/woofers and the reason a low freq was used was ONLY because cheaper DMM's give a precise reading there (I guess they're optimized to measure mains-derived voltages). those -12dB are NOT an estimation of peak-to-average ratio found in music! it's a coincidence, nothing more. also, you need to try and understand that what you are measuring with the DMM is simply the highest voltage your digital source (DAC, CD, DVD, SACD, MP3 player etc) will ever output when the amp is set to the volume level determined using music. and you also need to keep in mind that your digital source can NEVER output a voltage higher that the value you measured x 4 (usually around 2V RMS). if you don't understand or agree the above please tell us at least what it is.
you can try and explain over and over and over and in the end you find that some didn't even read the thread because they saw the words "sine wave" being used in the context of audio. which is the big no-no of the subjectivist audiophile.
that is fine with me but I see Pano's and 5th's attempts to explain this failing over and over because people simply refuse to try and understand what this is about.
I can't help but notice the psychological side of this
auplater, 5th explained it very well many times. he didn't simply tell you "no, you're wrong". you need to try to understand that the reason that -12dB sine was chosen was ONLY to protect your speakers/woofers and the reason a low freq was used was ONLY because cheaper DMM's give a precise reading there (I guess they're optimized to measure mains-derived voltages). those -12dB are NOT an estimation of peak-to-average ratio found in music! it's a coincidence, nothing more. also, you need to try and understand that what you are measuring with the DMM is simply the highest voltage your digital source (DAC, CD, DVD, SACD, MP3 player etc) will ever output when the amp is set to the volume level determined using music. and you also need to keep in mind that your digital source can NEVER output a voltage higher that the value you measured x 4 (usually around 2V RMS). if you don't understand or agree the above please tell us at least what it is.
Last edited:
Ok, busy weekend, just ran your -18db pink test. Got .44VAC at 85dbc.
My brain is pretty tired but calculating a 6x multiplier I get 2.64v at 0dB. into 8ohms is .87w... that dosen't sound right.
math check please
Typical pink noise has a 12dB crest factor, the VOM will read the (far lower) average, not peak power.Ok, busy weekend, just ran your -18db pink test. Got .44VAC at 85dbc.
My brain is pretty tired but calculating a 6x multiplier I get 2.64v at 0dB. into 8ohms is .87w... that dosen't sound right.
math check please
A sine wave is the only way to get an accurate voltage reading, and even that will start to vary when you get more than an octave or two from 60Hz, unless you have a really good meter.
Right! Do NOT measure the pink noise with the volt meter. Measure the pink noise with the SPL meter, then run the test tone and measure it. But beware. The test tone is going to be LOUD once you've set the pink noise to 85dB. How loud? Should be 6dB louder, or 91dB. Don't try this at 3:00 AM while the rest of the house is asleep. 
Maybe I should make a -18dB sine wave?
EDIT: And here they are. 120Hz and 220Hz at -18dBfs MP3 inside a Zip
These are 6dB lower than the original tones, or 1/2 the voltage.
Maybe I should make a -18dB sine wave?
EDIT: And here they are. 120Hz and 220Hz at -18dBfs MP3 inside a Zip
These are 6dB lower than the original tones, or 1/2 the voltage.
Last edited: