John, I don't want to quote the whole thing, but wanted to thank you for your note. Appreciated for sure. And that really, really, REALLY sucks about being part of the firestorms.
The trick is, really, that working more on the DSP/crossover side of the spectrum is really where I need to improve my system the most. Everything else is orders of magnitude below that. Fortunately, you can do a decent job of single-blind (aka computer randomized ABX) testing of different configurations to find out what works best.
Not that I won't continue to play around with hardware, just that's going to reap the most fruit.
The trick is, really, that working more on the DSP/crossover side of the spectrum is really where I need to improve my system the most. Everything else is orders of magnitude below that. Fortunately, you can do a decent job of single-blind (aka computer randomized ABX) testing of different configurations to find out what works best.
Not that I won't continue to play around with hardware, just that's going to reap the most fruit.
Lots of humor and little communication.
First CDs have an absolute maximum level. So if your audio power amplifier can handle that without clipping into the loudspeaker of your dreams you have eliminated one issue. Now as many amplifiers have a gain between Av of 20 to 30 a peak voltage of 60 volts RMS should be sufficient. With some loudspeakers actually dropping to as low as 2 ohms that requires a peak current capability of around 50 amps with rail voltages of 90ish volts DC.
Now using an amplifier with around 3 nV/rt Hz noise that will get you about 160 dB signal to noise for the generally recognized audio bandwidth
That brings up the issue of distortion level. The big issue is that loudspeakers do not mask distortion they increase the order of distortion from the 2nd or 3rd harmonic to the 4th through 9th albeit at a lower level.
If you have the amplifier that never clips it must be low distortion at all playback levels. That is where amplifiers usually exhibit different characteristics.
In my work a dynamic range into the system can go from an announcer turning his head away from a cardioid microphone while reading statistics to screaming into the same microphone with lips touching it. A range of more than 100 dB. Now the goal of faithful reproduction of that would create significant and accurate complaints as the goal is actually to maximize communication.
Next comes ear training. When correctly administered speech intelligibilty tests are done you use a trained listener bank. There are clearly better and more consistent results with training. It also turns out once trained the acuity does note seem to fade with time.
Some folks seem to think expert level training takes between 2,000 to 10,000 hours. Fortunately for speech intelligibilty training results seem to converge around 20-30 hours.
As folks here have noted on large scale systems the atmospheric absorption requires significantly more power for the high frequencies. I might be sorry to have mentioned that as I still see designs where the amplifiers scale down as frequency increases.
There are measurable frequencies from musical instruments well above 20,000 hertz. Some have speculated these frequencies need to be included in reproduction to get more accuracy. However if measured at a distance of 20 feet these higher frequencies also have significant atmospheric absorption.
Then we get into the significant issue of listening acoustics. RNM has presented a very interesting non symmetrical room.
I recently signed the longest non disclosure agreement probably ever produced, there are some areas that I cannot add anything new to the conversation. But I will privately laugh at some of the assumptions presented here.
The limit to sound reproduction may actually be the standard adopted for the recording, distribution and reproduction. That makes introducing really better systems quite difficult.
In the larger systems of mine what are issues that are ignored as being insignificant become elephants in the room. So the real question becomes at what scale do they become significant?
One of the important issues is Maslow's observation of tools limiting results. (If all you have is a hammer, everything looks like a nail.). If you only look at RLC you are quite incomplete. Adding I&E helps but it still is quite incomplete. Using FFTs as the measurement standard also tilts the observations. The expectation that measurement is the standard is also very biased. Picasso is the best example of electrical accuracy may not be the goal.
George should be back on track tomorrow.
First CDs have an absolute maximum level. So if your audio power amplifier can handle that without clipping into the loudspeaker of your dreams you have eliminated one issue. Now as many amplifiers have a gain between Av of 20 to 30 a peak voltage of 60 volts RMS should be sufficient. With some loudspeakers actually dropping to as low as 2 ohms that requires a peak current capability of around 50 amps with rail voltages of 90ish volts DC.
Now using an amplifier with around 3 nV/rt Hz noise that will get you about 160 dB signal to noise for the generally recognized audio bandwidth
That brings up the issue of distortion level. The big issue is that loudspeakers do not mask distortion they increase the order of distortion from the 2nd or 3rd harmonic to the 4th through 9th albeit at a lower level.
If you have the amplifier that never clips it must be low distortion at all playback levels. That is where amplifiers usually exhibit different characteristics.
In my work a dynamic range into the system can go from an announcer turning his head away from a cardioid microphone while reading statistics to screaming into the same microphone with lips touching it. A range of more than 100 dB. Now the goal of faithful reproduction of that would create significant and accurate complaints as the goal is actually to maximize communication.
Next comes ear training. When correctly administered speech intelligibilty tests are done you use a trained listener bank. There are clearly better and more consistent results with training. It also turns out once trained the acuity does note seem to fade with time.
Some folks seem to think expert level training takes between 2,000 to 10,000 hours. Fortunately for speech intelligibilty training results seem to converge around 20-30 hours.
As folks here have noted on large scale systems the atmospheric absorption requires significantly more power for the high frequencies. I might be sorry to have mentioned that as I still see designs where the amplifiers scale down as frequency increases.
There are measurable frequencies from musical instruments well above 20,000 hertz. Some have speculated these frequencies need to be included in reproduction to get more accuracy. However if measured at a distance of 20 feet these higher frequencies also have significant atmospheric absorption.
Then we get into the significant issue of listening acoustics. RNM has presented a very interesting non symmetrical room.
I recently signed the longest non disclosure agreement probably ever produced, there are some areas that I cannot add anything new to the conversation. But I will privately laugh at some of the assumptions presented here.
The limit to sound reproduction may actually be the standard adopted for the recording, distribution and reproduction. That makes introducing really better systems quite difficult.
In the larger systems of mine what are issues that are ignored as being insignificant become elephants in the room. So the real question becomes at what scale do they become significant?
One of the important issues is Maslow's observation of tools limiting results. (If all you have is a hammer, everything looks like a nail.). If you only look at RLC you are quite incomplete. Adding I&E helps but it still is quite incomplete. Using FFTs as the measurement standard also tilts the observations. The expectation that measurement is the standard is also very biased. Picasso is the best example of electrical accuracy may not be the goal.
George should be back on track tomorrow.
Brass harmonics often have lots of energy at audible high frequencies. See figure 1(a)
There's life above 20 kilohertz! A survey of musical instrument spectra to 102.4 kHz
With wide dispersion speakers, there is a problem with the asym... esp that glass door/wall nearby.
However, that is all the more reason to use speakers with narrow dispersion like the JBL M2. +/- 30 degrees. The speakers actually will need to be toed in a little to be on-axis at the listening position I use. This will further reduce side reflection. Then there is a large sound absorbing panel to be put up on the glass wall/door in critical imaging and listening... no reflection. Very symet sound field and excellent imaging is the end result.
THx-RNMarsh
Last edited:
Due to the discussion we have had. I think a few of the true believers will like this article.
Rogue Audio RP-1 preamplifier | Stereophile.com
Ok that's all requirements to understand the performance neatly side-stepped
Rogue Audio RP-1 preamplifier | Stereophile.com
Ok that's all requirements to understand the performance neatly side-stepped
Thank you Wahab for almost forcing me into wrongness, because usually when I am wrong, I learn a lot. Being almost wrong (I did fall for the average power of the 20 Hz square wave as the relevant measurement for a moment) is almost as good for learning, but much easier on the soul.
In search for hard data, I found the following:
Musical Waveform Peak and Average Power Frequency Distributions
In table 1, 1/3rd octave band passed peak values are shown for a piece of classical music.
There are two local peaks, one at 315 Hz @ 500 mV, and one at 12.5 kHz @ 350 mV.
The ratio between these two peaks is 1.42. Does mother nature throw another sqrt(2) at our feet here? Who knows, it may be coincidence, but at least it makes it easy to see that in power terms, if your high amp has only half the power of the low amp, all voltage requirements can be easily met.
Another consideration in this regard is that some clipping at 12.5kHz may send some bats into a spin, but will not be audible to humans. Half the power for the highs is more than enough.
In search for hard data, I found the following:
Musical Waveform Peak and Average Power Frequency Distributions
In table 1, 1/3rd octave band passed peak values are shown for a piece of classical music.
There are two local peaks, one at 315 Hz @ 500 mV, and one at 12.5 kHz @ 350 mV.
The ratio between these two peaks is 1.42. Does mother nature throw another sqrt(2) at our feet here? Who knows, it may be coincidence, but at least it makes it easy to see that in power terms, if your high amp has only half the power of the low amp, all voltage requirements can be easily met.
Another consideration in this regard is that some clipping at 12.5kHz may send some bats into a spin, but will not be audible to humans. Half the power for the highs is more than enough.
Ok that's all requirements to understand the performance neatly side-stepped
Unlike that silly battery-operated unit discussed elsewhere on the forum, the performance here is OK. The tube stage needs some work to get the distortion down (it may or may not be audible, but it's easy to get it an order of magnitude or two better to assure that it's inaudible), and the headphone performance is dreadful, but the phono looks fine.
$1700 for "OK" performance seems a bit steep, but chacun a son gout.
edit: I misread the distortion plot. That looks OK, low enough. My apologies, need moar coffee.
Due to the discussion we have had. I think a few of the true believers will like this article.
Rogue Audio RP-1 preamplifier | Stereophile.com
Quote:
I believe the most important thing an audio component can do is disappear into the one-song-after-another matrix of the reviewer's journey.
Ok that's all requirements to understand the performance neatly side-stepped
Wow... I need to get hold of his buzzword generator, it's a real good one!
Those numbers are not an absolute limit for all kinds of music but only for those four pieces, so half the power is more than enough only for those exemples that are in no way the end of all in music variability.
To summarize a passive 2 way system with a 2 x 100W amp will have better dynamic that a multi amped system that would be say 2 x 100W + 2 x 50W...
http://atcloudspeakers.co.uk/hi-fi/loudspeakers/tower-series/scm300aslt/
550W to 300Hz
200W to 3KHz
100W above 3KHz
Rated for continuous SPL of 121dB. If you believe the numbers FANE calculated which is that the 50/50 split in power occurs at 350Hz these numbers seem very sensible when you account for driver sensitivity differences and baffle step correction. I (and the active studio monitor industry) are clearly missing something. As an aside I would have though equal power amps would be cheaper as only a single PSU...
550W to 300Hz
200W to 3KHz
100W above 3KHz
Rated for continuous SPL of 121dB. If you believe the numbers FANE calculated which is that the 50/50 split in power occurs at 350Hz these numbers seem very sensible when you account for driver sensitivity differences and baffle step correction. I (and the active studio monitor industry) are clearly missing something. As an aside I would have though equal power amps would be cheaper as only a single PSU...
Actually your approach works well for the direct or near field. The asymmetry enhances the reverberant or far field. So you are getting the best of both.
Many if not almost all modern recordings are done near field and require room reverberation to enhance them. Looks like you have that.
That is why I use interesting as the descriptor. Much to be learned from your approach.
I will have to let it process in the distant regions of my mind as it seems the technique should be applicable to stadium sky box style suites.
You seems to miss the fact that peaks ramps are reproduced exclusively by the tweeter, so if there s a peak of 100W that is initiated by a 0.5ms rise time ramp that reach full amplitude all the power will be dispatched to the tweeter, if this latter use a 50W amp there will be saturation while with a 100W passive system there will be no saturation.
Last edited:
but outside German industrial noise music where does that appear? and for a (say) 90dB sensitive tweeter that's 110dB@1m which is above THX levels, so batsrad loud for a domestic setting. In reality most of us listen at an average between 60 and 80dB with a crest factor of 12-14dB.*
Not saying your requirements would not be valid for a domestic build, just a bit of an edge case or a love of very strange and very loud music.
*Next preamp I build will have the volume dial calibrated against THx to show SPL at -20dBFS.
Not saying your requirements would not be valid for a domestic build, just a bit of an edge case or a love of very strange and very loud music.
*Next preamp I build will have the volume dial calibrated against THx to show SPL at -20dBFS.
That has nothing to do with the style of music, anytime there s a rising hedge whose duration is shorter than the high pass filter time constant all the energy will be dispatched to the tweeter, or treble amp in a multiway system.
If the cut off frequency is say 1KHz then any rising hedge whose duration is somewhat lower than 1ms will be entirely treated by the tweeter, if the rising hedge reach full amplitude then a multiway system that use a treble amps of lower power than for the bass amp it wont be capable to reproduce the attack of the signal accordingly if it s a fortissimo that imply instruments of various register like a classic orchestra, the low frequencies will be reproduced accurately but not the sum of the attack of all instruments.
Attachments
Ed's back.
We did cover the last one at length (so to speak) and there was an amazing amount of disagreement from all sides.
Those who are knowledgeable in the art, do tend to spend JS on reality (all kids and no meat).
In that sense, Mr. Curl does have a point.
On the other hand, the overzealous spenders appear to lack any theoretical backbone, which leads to examples as tasting insulator sleeves of multibuck power/interconnect wires.
(and then there's the legion of those who fit the FIQ42 bill, pretending to reach aural bliss by tweaking the power cord of an Aldi clearance sale offer, due to zero combined cerebral and wallet content)
Last edited:
- Status
- Not open for further replies.