Output impedance results in modulated frequency response at amplifier output terminals. See the measured influence of real 3-way speaker to amplitude response of the amp with only 0.25 Ohm output impedance. You can easily imagine what happens with tube amp with >1 Ohm output imp.
Blue .... card loopback
Red .... amplitude response into 6.8 Ohm load
Green ... amplitude response with 3-way speaker
The rest are speculations.
Blue .... card loopback
Red .... amplitude response into 6.8 Ohm load
Green ... amplitude response with 3-way speaker
The rest are speculations.
Attachments
What are these tests supposed to be showing? It would seem that one channel is driving a signal into the room. Another idle channel is being watched as the speaker connected to it picks this up. There's no surprise in that.
If the amp were to make a contribution I could hardly see it being greater than modifying a reflection picked up at the speaker, one which would be reflected acoustically at any rate.
If there were to be reverb, wouldn't that require a delay loop somewhere?
The loop delay comes from the room itself: acoustic delay of the sound entering the room, being reflected locally, and returning via the loudspeakers' microphonic effect.
Mike (mikebse2a3 on the K-forum) tried to recreate the effect that Bob Carver described using two different amplifiers that he had on hand and a Jubilee bass bin, which has a minimum impedance of about 3 ohms at 60 and 120 Hz. The HF horn/driver combination of that loudspeaker has a minimum Z of 11 ohms at about 600 Hz, so its contribution to the effect was ruled out due to its high "damping factor" (including the amplifiers chosen) of at least 1. I don't believe that Mike showed the smoking gun levels of reverb, but he did verify that it is occurring, and it can be measured.
Carver described a reverb effect that was at least 20 dB higher in the quoted Enjoy the Music article, which would be audible during the decays. He claims that his amplifiers are in fact designed to amplify that effect--how he did that, he didn't discuss (as is usual for Mr. Carver). Perhaps he is assuming a crest factor of 20 (i.e., dynamic range of the recordings used) during transients to arrive at his -30 dB reverberation levels due to loudspeaker microphonics and high output Z amplifiers.
The reason for the tests by Mike was due to the effect that I put into that FAQ. It is the only effect that I can find that explains the differences in sound heard--typically called "tube magic" which is a function of its output impedance relative to the loudspeaker's input impedance with very efficient loudspeakers, i.e., horn loaded loudspeakers located in the corners of a room where the room acoustic reinforcement is maximized. This effect typically adds to the depth of the soundstage, hence the "reverb effect", like dipole loudspeakers produce in-room to create an artificial depth of soundstage image even when there is none in the recording.
The last of my responses to Mike is that there is a freeware utility provided by Bill Waslo (Liberty Instruments) called "Audio Diffmaker" that could be used to capture the differences between amplifiers - with and without high output impedance using horn-loaded loudspeakers located in corners of the room, especially for the transients, where the effect of reverberations induced would be most audible. Neither he nor I have run those tests yet.
You could--if you're running horn loaded with tube/valve amplifiers like a SET or OTL without any kind of feedback to lower its output impedance.
Chris
Mike (mikebse2a3 on the K-forum) tried to recreate the effect that Bob Carver described using two different amplifiers that he had on hand and a Jubilee bass bin, which has a minimum impedance of about 3 ohms at 60 and 120 Hz. The HF horn/driver combination of that loudspeaker has a minimum Z of 11 ohms at about 600 Hz, so its contribution to the effect was ruled out due to its high "damping factor" (including the amplifiers chosen) of at least 1. I don't believe that Mike showed the smoking gun levels of reverb, but he did verify that it is occurring, and it can be measured.
Carver described a reverb effect that was at least 20 dB higher in the quoted Enjoy the Music article, which would be audible during the decays. He claims that his amplifiers are in fact designed to amplify that effect--how he did that, he didn't discuss (as is usual for Mr. Carver). Perhaps he is assuming a crest factor of 20 (i.e., dynamic range of the recordings used) during transients to arrive at his -30 dB reverberation levels due to loudspeaker microphonics and high output Z amplifiers.
The reason for the tests by Mike was due to the effect that I put into that FAQ. It is the only effect that I can find that explains the differences in sound heard--typically called "tube magic" which is a function of its output impedance relative to the loudspeaker's input impedance with very efficient loudspeakers, i.e., horn loaded loudspeakers located in the corners of a room where the room acoustic reinforcement is maximized. This effect typically adds to the depth of the soundstage, hence the "reverb effect", like dipole loudspeakers produce in-room to create an artificial depth of soundstage image even when there is none in the recording.
The last of my responses to Mike is that there is a freeware utility provided by Bill Waslo (Liberty Instruments) called "Audio Diffmaker" that could be used to capture the differences between amplifiers - with and without high output impedance using horn-loaded loudspeakers located in corners of the room, especially for the transients, where the effect of reverberations induced would be most audible. Neither he nor I have run those tests yet.
You could--if you're running horn loaded with tube/valve amplifiers like a SET or OTL without any kind of feedback to lower its output impedance.
Chris
Last edited:
Hi Allen,
If I understand Bob Carver's description of the "reverb effect" correctly, he's saying that the loudspeaker is providing a signal to the feedback loop of an amplifier that then gets amplified at its output. My problem with this is that the feedback network is applied to the inverting input and the resulting inverted output will appose the applied input signal !
So instead of the applied signal increasing in amplitude, surely it must get smaller, because the signal has gone through an inversion. Sounds like emperors new cloths to me!
Peter
If I understand Bob Carver's description of the "reverb effect" correctly, he's saying that the loudspeaker is providing a signal to the feedback loop of an amplifier that then gets amplified at its output. My problem with this is that the feedback network is applied to the inverting input and the resulting inverted output will appose the applied input signal !
So instead of the applied signal increasing in amplitude, surely it must get smaller, because the signal has gone through an inversion. Sounds like emperors new cloths to me!
Peter
That's exactly why Mike B. did the tests, so that "pundits" can accept the fact that what Carver is describing is actually happening. I don't know about you, but I believe Carver, and not the pundits that say that it "can't happen". Triodes have internal feedback that respond to the effects of changing load conditions.
Testing is the reason to believe - not mental models and naysayers that don't want to believe it because it's inconvenient that it's happening with their favorite low power SETs, and that it isn't some "magic" that is occurring that makes the sound "better than recorded".
YMMV.
Chris
Testing is the reason to believe - not mental models and naysayers that don't want to believe it because it's inconvenient that it's happening with their favorite low power SETs, and that it isn't some "magic" that is occurring that makes the sound "better than recorded".
YMMV.
Chris
Last edited:
Where are the measurements? Didn't see them.
I can understand how this might happen in corner horn woofers with a very low damping factor amp, but don't know how it could happen with typical mid or high horns. They usually have a resistive attenuation network that would isolate the amp from the microphonics.
I can understand how this might happen in corner horn woofers with a very low damping factor amp, but don't know how it could happen with typical mid or high horns. They usually have a resistive attenuation network that would isolate the amp from the microphonics.
You didn't see Mike's measurements in this thread?
https://community.klipsch.com/index.php?/topic/157466-speaker-microphone-and-room-reverb-effects/
"Big as Dallas" as they say. The bass bin that was used is typically crossed over at about 425 Hz. I'm not sure, but I don't believe that Mike low-passed the bass bin in his measurements.
To some degree, you're right, but not by more than the relative efficiency of the midrange horn/driver or tweeter vs. the bass bin. Usually there is about a 3-6 dB difference with the fully horn-loaded loudspeakers that I've dealt with (the subject of the FAQ).
If you're taking issue with only a part of the argument--as it appears that you are--I recommend that you use your Altec A5s to measure it yourself. I'm not the one that disbelieves.
Chris
Cask, I've read you FAQ with interest a few times, and thanks. It's just this one thing I'm stuck on.
If you were to fill the throat of the receiving horn with concrete you'd get a reflection, probably as if there was only a bare corner.
If you short the driver terminals with a corner horn there will be some give, and the movement will damp through the driver Re, while being released per the transient response.
If you include Zo it adds with Re to reduce the damping from one finite level to another.
Feedback through an amplifier is for all present intents and purposes, immediate. It would give the effect of reducing Zo. It would reduce the cone movement increasing damping.
Could amplification of the received feedback (to take it beyond the level of a bare corner reflection) be achieved without positive feedback?
If you were to fill the throat of the receiving horn with concrete you'd get a reflection, probably as if there was only a bare corner.
If you short the driver terminals with a corner horn there will be some give, and the movement will damp through the driver Re, while being released per the transient response.
If you include Zo it adds with Re to reduce the damping from one finite level to another.
Feedback through an amplifier is for all present intents and purposes, immediate. It would give the effect of reducing Zo. It would reduce the cone movement increasing damping.
Could amplification of the received feedback (to take it beyond the level of a bare corner reflection) be achieved without positive feedback?
Last edited:
I think of it this way:
With a corner horn loudspeaker acting as a microphone, sending a time-delayed acoustic reflection signal back to the amplifier output terminals, anything in the electric circuit at the amplifier output terminals that looks like impedance will generate a negative response from the amplifier, if the amplifier has feedback properties (like all triode-based amplifiers will have).
If the high output impedance amplifier is designed to amplify that "time-delayed over-running signal", like Carver claims his new amplifier designs do, the effect can be magnified. I assume he is using global feedback from the output stage to the input, maybe even with more electrical delay added for dramatic effect.
If the output impedance of the amplifier is effectively zero (like most SS amplifiers and tube/valve amplifiers with feedback that drives the output impedance to well below 1 ohm), there is nothing that the amplifier sees to push back against, even if large quantities of negative feedback are used at the output of the last amplifier gain stage.
The key to all this is the time delay of that signal coming back - it's like a purely over-running load. The servo-mechanism (electric feedback potential, in this case) will move to null that response even if there is no command input. All this is time delayed, so if you get a big impulse through the forward loop of the amplifier, the returning "feedback" signal will be delayed and attenuated by a great amount, However, the ear is a logarithmic device in terms of perceiving relative loudness, and if that feedback signal is part of a sudden transient (spike or impulse), the amplifier will see a returning time-delayed signal that it will response to, even if nothing else is currently playing (forward loop) at that time. Humans in the room will hear a series of echo pulses, just like being in a cave, all delayed in time to each other by the characteristic time delay of the returning acoustic reflections from the room.
Chris
With a corner horn loudspeaker acting as a microphone, sending a time-delayed acoustic reflection signal back to the amplifier output terminals, anything in the electric circuit at the amplifier output terminals that looks like impedance will generate a negative response from the amplifier, if the amplifier has feedback properties (like all triode-based amplifiers will have).
If the high output impedance amplifier is designed to amplify that "time-delayed over-running signal", like Carver claims his new amplifier designs do, the effect can be magnified. I assume he is using global feedback from the output stage to the input, maybe even with more electrical delay added for dramatic effect.
If the output impedance of the amplifier is effectively zero (like most SS amplifiers and tube/valve amplifiers with feedback that drives the output impedance to well below 1 ohm), there is nothing that the amplifier sees to push back against, even if large quantities of negative feedback are used at the output of the last amplifier gain stage.
The key to all this is the time delay of that signal coming back - it's like a purely over-running load. The servo-mechanism (electric feedback potential, in this case) will move to null that response even if there is no command input. All this is time delayed, so if you get a big impulse through the forward loop of the amplifier, the returning "feedback" signal will be delayed and attenuated by a great amount, However, the ear is a logarithmic device in terms of perceiving relative loudness, and if that feedback signal is part of a sudden transient (spike or impulse), the amplifier will see a returning time-delayed signal that it will response to, even if nothing else is currently playing (forward loop) at that time. Humans in the room will hear a series of echo pulses, just like being in a cave, all delayed in time to each other by the characteristic time delay of the returning acoustic reflections from the room.
Chris
One thing is that I don't see the acoustically fed back signal as correlating to the original stimulus. I could be wrong but not entirely, I think.
The cone will 'push back' if the coil is shorted, passively via damping, though only partially compared to the fed back stimulus.
In order for it to 'amplify' this I would think that Zo should have negative resistance, to first negate Re, and then some.
The cone will 'push back' if the coil is shorted, passively via damping, though only partially compared to the fed back stimulus.
In order for it to 'amplify' this I would think that Zo should have negative resistance, to first negate Re, and then some.
I don't believe that I'd want the "speaker-microphone effect", to be clear. I'd select an amplifier with low enough output impedance such that the effect isn't audible.
The type of "enhancement feedback" that you may be talking about is something that Carver might be implementing in order to enhance the reverb effect from in-room reflections while the amplifier is connected to a high efficiency speaker in a small room--having relatively higher amplitude returning acoustic reflections than in a larger room. I'm not sure how he would do that, and frankly, I don't have a lot of interest in knowing what he does other than perhaps using straight global feedback or the combination of global and gain-stage feedback schemas used together-but no further tricks like delays or positive feedback schemas.
Any tube/valve type amplifier design would start to get expensive due to the needed open-loop gain to support those levels of feedback that might enhance the "speaker-microphone effect" feedback levels.
Everything that has been discussed on this subject has included milliseconds of delay of the feedback due to the acoustic transit of energy into the room and then reflected (with losses) back to the loudspeaker again. The idea of "correlation" doesn't even come into the picture in this view. What kind of correlation are you referring to?
You can think of this as a linear superposition of two processes happening in parallel: forward-loop signal being injected into the room, and microphonic acoustic return back from the room after milliseconds of delay, much lessened in amplitude and subject to the nonlinear effects of room acoustic absorption and diffraction, as well as varying impedance coupling back to the amplifier output terminals via the frequency-dependent input impedance of the loudspeaker.
Any amplifier feedback effects sum these two processes together in somewhat linear fashion at the output stage. The amplifier takes as its input the complex sum of those two processes at any instant in time.
If you short a voice coil of a driver while it is receiving the room's acoustic energy, the only damping force would be the diaphragm velocity-based resistance of current flowing through whatever impedance exists in the coils. The only restoring force would be the driver's diaphragm surround suspension force.
Again, I'm guessing what your question is here, but I'd say that the effect that Carver is trying to achieve is moderate levels of room reverb effect in order to achieve "pleasing" levels of apparent room spaciousness via reverb, not to increase the reflectivity of the corner. Acoustic feedback through a microphone isn't pleasant to hear in a live venue. Doing that same thing (i.e., "greater than 1") using canned music tracks and some esoteric room acoustic feedback scheme in the amplifier would sound awful indeed.
OK, I think that's clear.
I'm not sure I see this as the case. The speaker has mechanical resistance but the primary action is reactive. On the other hand the voice coil impedance would primarily behave resistive. That is to say I'd expect voice coil heat to be the main damping outlet and the restorative force (transient response) acting upon the balance.
I would, but they have "Gone to Texas", as they say.
The author (Mike) does say he believes the effect to be negligible when music is playing from the speakers.
Mike and I have discussed this. Just because Mike says it's inaudible doesn't make it so. Carver says just the opposite is true. I still believe Carver, especially when you think about real music, not steady state sine waves.
One way to cut through the opinions is to use Bill Waslo's Audio DiffMaker to hear the differences only.
Sorry to hear of the demise of your A5's. Did you replace them with something better?
Chris
One way to cut through the opinions is to use Bill Waslo's Audio DiffMaker to hear the differences only.
Sorry to hear of the demise of your A5's. Did you replace them with something better?
Chris
Last edited:
I just looked up the input impedance of the two horn/driver units that Mike used for those tests. The level of microphonic behavior is related to both the efficiency of the loudspeaker and the damping factor - with a threshold of approximately 1, below which the microphonic behavior increases significantly.
The damping factor appears to have been at or higher than 1 for both conditions of the tests. Both amplifiers used for the tests have output impedances of something around 1 ohm (plus or minus). The input impedance of the first tested loudspeaker horn/drivers--the bass bin--at 315 Hz is 7 ohms. The input impedance of the second tested unit at 2kHz--the K-402 with TAD TD-4002 driver--is almost 20 ohms.
By inspection, the HF horn should exhibit lower microphonic signal levels--and it does.
Chris
The damping factor appears to have been at or higher than 1 for both conditions of the tests. Both amplifiers used for the tests have output impedances of something around 1 ohm (plus or minus). The input impedance of the first tested loudspeaker horn/drivers--the bass bin--at 315 Hz is 7 ohms. The input impedance of the second tested unit at 2kHz--the K-402 with TAD TD-4002 driver--is almost 20 ohms.
By inspection, the HF horn should exhibit lower microphonic signal levels--and it does.
Chris
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.