With FB the sound sticks to the speakers, less air, the highs & upper mids gets detached from the rest and there is a strong feeling of rhythmical inaccuracy. Oops, did I just say inaccurate?
Possibly as for the technical background of the terms "soundstage" and "imaging" as well as 3-D sensation in general, one might have a look at what this goes back to regarding the underlaying pattern of *all* of such sensations:
http://www.diyaudio.com/forums/ever...-giant-leap-maturizing-audio.html#post2616725
The point to keep in mind is those ASAR pattern encoding is present in mono as well as in stereo.
Taking what is said there one step further - its easy to see / extrapolate that such effects also occur from electronic circuits.
Take a feedback loop for example which - lets just assume for a moment - may have internal group delay of 1ms.
What will happen is that we will face more or less the same looped echo effect as from any other ASAR pattern - just that the "echo" after each 1ms is altered according to the FB properties.
Of course a 1ms is not any realistic - but I guess you get the picture.
Of course such looped echo effects / ASAR patterns are not restricted to electronic circuits that have implemented feed back loops. Just any cross talk issue - be it power supply lines or signal lines will result in such effects embraced by ASAR patterns - *if* there is delay time (group delay) in between
Best
Michael
http://www.diyaudio.com/forums/ever...-giant-leap-maturizing-audio.html#post2616725
The point to keep in mind is those ASAR pattern encoding is present in mono as well as in stereo.
Taking what is said there one step further - its easy to see / extrapolate that such effects also occur from electronic circuits.
Take a feedback loop for example which - lets just assume for a moment - may have internal group delay of 1ms.
What will happen is that we will face more or less the same looped echo effect as from any other ASAR pattern - just that the "echo" after each 1ms is altered according to the FB properties.
Of course a 1ms is not any realistic - but I guess you get the picture.
Of course such looped echo effects / ASAR patterns are not restricted to electronic circuits that have implemented feed back loops. Just any cross talk issue - be it power supply lines or signal lines will result in such effects embraced by ASAR patterns - *if* there is delay time (group delay) in between
Best
Michael
Last edited:
Using unrealistic numbers , one can only get unrealistic pictures...
Using real numbers , group delay in a circuit will inevitably appear
as a negligible quantity.
As for NFB affecting transients , well , that might be right for very badly designed circuits,
wich i guess are the convenient basis upon all the vilifying of NFB and SS devices are built...
As for soundstage and imaging , exluding the misdesigned amps,
they are only relevant when comparing speakers and their optimal
posistioning in the room...
Actually no, this is just a common mis belief - you have to keep apart quality and quantity.
*If* there is a pattern at all, it can be identified way below what people think might be "relevant" magnitudes.
Thats the magic about "pattern recognition" so few people are aware of.
Believe me - mankind are extremely trained to detect such ASAR patterns for ages
Again no.
You would have to dig way deeper as to why.
The point is not "bad designs" - at least thats not what I'm discussing here - Bode diagrams and its proper interpretation are around for a loooong time - its all about "delay time" involved.
Fundamental difference !
Michael
Last edited:
Yes, gootee you are right. But we don't have to stick with feedback. (Nonetheless it is ok if you want to go on with feedback.)
We already found out that a unnatural not real wide soundstage can be created with negative crosstalk.
Now we should go on and understand
how the NJM 4565D creates a front at the front and a distant back more than other.
Last edited:
As far as I know, no-one has ever demonstrated that. Before trying to find an explanation for a phenomenon, it's important to first show that there IS a phenomenon to be explained.
That's what I have heard! A lot of people have experienced that the NJMs have a little bit ringing decay. For me it would only explain why the back is so distant but not why at the same time the front is at the front.
Sorry, I can't proove that there IS a phenomeon. I have already suggested that everyone can test it with his cd-player.
Then, absent any evidence or plausibility, one must fall back on the most likely explanation.
In my humble opinion I think room acoustics has far more influence than any of the phenomenon that is being proven here. Unless you are listening in a free space environment with no reflections.
Therefor I do not believe that amplifiers alone can account for these ghostly images that is bening investigated.
If you were to connect your amplifier directly to the recording microphone and listen with an in-ear device, I am almost certain that you will only hear what the microphone hears.
I cannot see how an amplifier would selectively place some instruments/voices or sounds nearer to the listener and others behind, how does the amplifier know what to do? What has been recorder is time coherent with what is produced and you make decisions of relative placement in space based on time of arrival.
Therefor I do not believe that amplifiers alone can account for these ghostly images that is bening investigated.
If you were to connect your amplifier directly to the recording microphone and listen with an in-ear device, I am almost certain that you will only hear what the microphone hears.
I cannot see how an amplifier would selectively place some instruments/voices or sounds nearer to the listener and others behind, how does the amplifier know what to do? What has been recorder is time coherent with what is produced and you make decisions of relative placement in space based on time of arrival.
Last edited:
Hi all,
While not exactly aimed at the topic, some time ago I conducted an experiment with adding controlled amounts of distortion via a DSP-Plugin (Electri-Q, a VST-Plug, in its 'analog' mode) in my fully DSP controlled active speaker setup.
The speakers (15" pro coax driver running free air without any baffle) had quite simple passive XO but were fully DSP-corrected to achieve excellent magnitude and phase response and precise L/R-matching at the listening pos.
Amps and DAC used were pretty mid-fi stuff, at best. Still I have the highest level of spacial transparency and richness ever achieved at my home, with good recordings (or when "beefing up" the soundstage which is quite possible and really works out for me).
The plugin creates ~1%THD at 0dBFS, no 2nd, progressively falling thereafter. There is a side effect typical of simple digital distortion, aliasing of components which happen to fall near fs/2. Also, the RMS level is slightly louder by a fraction of a dB. The distortion mechanism does not obey energy conservation, same is true for a real amplifier, therefore I choose to NOT correct for that.
When switching the distortion in/out (blind/randomized, and sighted), changes in 'air', 'attack', 'timbre' were noticable, the distorted version sounding 'fresher' overall and quite preferable in my book. But soundstage collapsed a bit (notably when the polarity which gave better subjective imaging was used), speakers become more noticable and phantom sources more blurred in the treble region. Part of this might also have been induced by the aliasing products which are known to disturb HF imaging.
If I could dial in nice 'euphonic' distortion which does not affect soundstaging to much and sounds subjectivly better overall (I don't listen to flute ensemble music and the like) I'd take that any day.
---------:---------
Yet a more interesting point in the discussion about soundstaging vs feedback, I found that many drivers do profit much when their internal feedback mechanism, with typically a low feedback factor, is taken out of effect by high impedance drive (feasible for FR and active multiway). That did clear up imaging properties (seemingly related to distortion and time-smear) every time I tried, even with drivers that would not seem to apply for candidates... I had some surprises there. Keeping some damping at the fundamental system resonance is desireable when in the passband, and strong electromagnetic damping as in established practice seems still the best way to run a driver in a reflex box and other complex load environments.
With arbitrary variable drive impedance a high feedback amp will try to minimize the distortion of a linear (vector) sum of weighted voltage and current components at the driver, yet the current and the voltage at the driver will be distorted when looking at them in isolation. IHMO the first task would be to find the best impedance vs frequency profile for any given driver in an active multiway (designed for precise imaging to start with) as a baseline.
The next step would be so see if the sound changes when the amps are constructed differently as long it is still possible to dial in the profiles (which will need available negative feedback for realisation, unless in very simple cases).
That would be pretty high effort and time comsuming expreriments... not easily handled by any skilled individual alone...
My 2ct, Klaus
While not exactly aimed at the topic, some time ago I conducted an experiment with adding controlled amounts of distortion via a DSP-Plugin (Electri-Q, a VST-Plug, in its 'analog' mode) in my fully DSP controlled active speaker setup.
The speakers (15" pro coax driver running free air without any baffle) had quite simple passive XO but were fully DSP-corrected to achieve excellent magnitude and phase response and precise L/R-matching at the listening pos.
Amps and DAC used were pretty mid-fi stuff, at best. Still I have the highest level of spacial transparency and richness ever achieved at my home, with good recordings (or when "beefing up" the soundstage which is quite possible and really works out for me).
The plugin creates ~1%THD at 0dBFS, no 2nd, progressively falling thereafter. There is a side effect typical of simple digital distortion, aliasing of components which happen to fall near fs/2. Also, the RMS level is slightly louder by a fraction of a dB. The distortion mechanism does not obey energy conservation, same is true for a real amplifier, therefore I choose to NOT correct for that.
When switching the distortion in/out (blind/randomized, and sighted), changes in 'air', 'attack', 'timbre' were noticable, the distorted version sounding 'fresher' overall and quite preferable in my book. But soundstage collapsed a bit (notably when the polarity which gave better subjective imaging was used), speakers become more noticable and phantom sources more blurred in the treble region. Part of this might also have been induced by the aliasing products which are known to disturb HF imaging.
If I could dial in nice 'euphonic' distortion which does not affect soundstaging to much and sounds subjectivly better overall (I don't listen to flute ensemble music and the like) I'd take that any day.
---------:---------
Yet a more interesting point in the discussion about soundstaging vs feedback, I found that many drivers do profit much when their internal feedback mechanism, with typically a low feedback factor, is taken out of effect by high impedance drive (feasible for FR and active multiway). That did clear up imaging properties (seemingly related to distortion and time-smear) every time I tried, even with drivers that would not seem to apply for candidates... I had some surprises there. Keeping some damping at the fundamental system resonance is desireable when in the passband, and strong electromagnetic damping as in established practice seems still the best way to run a driver in a reflex box and other complex load environments.
With arbitrary variable drive impedance a high feedback amp will try to minimize the distortion of a linear (vector) sum of weighted voltage and current components at the driver, yet the current and the voltage at the driver will be distorted when looking at them in isolation. IHMO the first task would be to find the best impedance vs frequency profile for any given driver in an active multiway (designed for precise imaging to start with) as a baseline.
The next step would be so see if the sound changes when the amps are constructed differently as long it is still possible to dial in the profiles (which will need available negative feedback for realisation, unless in very simple cases).
That would be pretty high effort and time comsuming expreriments... not easily handled by any skilled individual alone...
My 2ct, Klaus
Last edited:
I have only changed the opas in the cd-player (for the critics, everything with powersupply etc. is fine). My speaker are still at the same place in my room.
I want to give something to think about:
I also mentioned that the NJMs have also a wider soundstage.
I couldn't see how this can be created just as a lot of critics and as you. That is why I asked how this could be. Fortunately, someone showed us negative crosstalk creates a wide soundstage.
But I think the effect of near and distant is something complete different than wide.
On the record are informations about the soundstage and imaging. Distant sounds have a decay and tone darkening. Maybe the NJMs untruly emphasize, make the decay louder etc. that is on the record.
PS: According to feedback and the title of the topic. I think I draw a false conclusion with the feedback. So don't argue to much about feedback. But you are free to go on to discuss the effects of feedback.
Last edited:
I did a test this afternoon. Marantz PM7000 Integrated vs my Ovation 250.
The Marantz does not image as well on width and depth.
Definitely, the recording and speakers play a big part, but there is still an element that is attributable to the preamp/amp.
I don't think we have found the holy grail yet.
The Marantz does not image as well on width and depth.
Definitely, the recording and speakers play a big part, but there is still an element that is attributable to the preamp/amp.
I don't think we have found the holy grail yet.
How does the amp know what to emphasize or not or where to place certain sounds? First off all, this is a classical case of anthropomorphism - the attribution of human characteristics to dead objects. Like 'the amp thinks' or 'maybe the amp untruly emphasizes'.
The amp is a dead object that gets fed with an electrical signal (see if you can spot the anthropomorphism in this last scentence?).
Electrically there is no difference between a near sound or a far sound - it's just a continuously varying voltage level, that is interpreted BY US as a far or near sound.
The real weak point here is that the perception of an individual listener is taken as reality - I put in a different opamp, I 'perceive' a greater soundstage, therefor the opamp causes a greater soundstage.
The first order of business would be to make sure that there is indeed a difference in sound stage, and for that you need to do some kind of controlled, repeatable listening test that clearly shows that this is the case. Otherwise, we're all chasing ghosts.
jan didden
Last edited:
You misunderstand me. The amp don't think!
A lot of people heard a difference in soundstage but very often they mentioned it with decay. But only decay would put everything to the back.
The NJM has a bad settling time. It could be the reason for artifical decay.
I am guessing much:
Let's say the NJM adds a certain amount of decay to the music because of the bad settling time.
On the record original there is a front with no decay and sounds in the back with decay.
The NJM now adds the same amount of decay to both front and back and its decay.
Now the front has only some decay.
The back originally already has some decay. The original decay itself gets some decay.
Maybe it makes psychoacoustical a difference.
The front is heard now only 1 metre set back.
The back appears 10 metres put back.
Do you understand what I mean?
Jan, i think when a designer leaves the straight way to fidelity he can design in certain characters. This is more extreme in mechanical devices like turntables. I measured the original Linn and it was running a bit to fast. When i dropped the cartridge or was throwing a metal ball on a string to the chassis i could measure considerable ringing. When you add reverb ( ringing tubes ? ) and make the speed a bit higher you can get better speech recognition and a sense of pace. This is all wrong and tiring in the end of cause because you eat the same sausage each day.
Y'all are picking at the argument for the sake of argument. I don't believe anyone here thinks an amp "knows" or "chooses."
But amps can make the soundstage sound different thru slightly different FR or harmonic content, to name but two. The electrical effect is simple, the interpretation of it is not.
Just as an example - MP3s always sound more "forward" to me. Perhaps because of the loss of some subtle details in the signal. It can be a pleasant effect, actually - in the short term.
But amps can make the soundstage sound different thru slightly different FR or harmonic content, to name but two. The electrical effect is simple, the interpretation of it is not.
Just as an example - MP3s always sound more "forward" to me. Perhaps because of the loss of some subtle details in the signal. It can be a pleasant effect, actually - in the short term.
Pano,
FR and different harmonic content are measureable so that can be easily confirmed. But any modern opamp like the 4562 has a ruler flat FR out to 100's of KHz, and harmonics below at least -120dB. Nobody can hear any sound stage difference from that in controlled listening.
MP3 is a deliberate modification of the source with the aim to take out as much of the music and still make it reasonable for the casual listener. A whole different kettle of fish.
jan didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.