Hello everyone,
and I would like to discuss a topic that is often mistakenly associated with the dynamic range of loudspeakers in my opinion. It is about the subjectively perceived by the listener decay time of instruments. In some loudspeakers, for example, a hi-hat decays for a very long time, and in others, although the strike itself is not less loud, the reverberation disappears very quickly and a clean, empty background appears. I thought that these loudspeakers might differ in the SPL characteristic as a function of the supplied electrical power - a loudspeaker that produces a long reverberation would have this characteristic more linear, and the one that produces less reverberation, falling more abruptly towards the muting of the sound level.
Since I could not find such graphs on the Internet, only a typical SPL graph in the frequency domain, I decided to make simple measurements myself. I connected the amplifier to a computer that played white and then pink noise, and I measured the sound intensity from the near field to eliminate the influence of the room. I increased the volume from zero to a specific level, then repeated one of the measurements with reduced amplifier gain, i.e. the volume intervals were smaller. And what?
These graphs show that the faster decay of instrument reverberation in some loudspeakers and slower in others does not result from the slope of their measured SPL characteristics vs. the supplied electrical power. So it is not any dynamic range of the loudspeakers or their SPL given in catalogs that decides the phase of the instruments' decay, because even if some loudspeakers play louder overall, the reverberation is received relatively to the impact and also decays relatively to the level of the impact (attack phase).
And my question: what, in your opinion, can decide that some loudspeakers beautifully convey the nuances of sound with a long phase of instrument decay, while others leave this information "to themselves"? PS. I used the Revox BX-350 and Wega Direct 200 speakers for the measurements. Both have dome tweeters, with the Wega being 3-way and the Revox being 2-way. The midrange measurement in the Revox speakers concerned one of the 4 parallel mid-woofers, while in the Wega - one midrange dome. The Revox BX-350 have 4 paper speakers and one tweeter, and the crossover frequency is 3200 Hz. As you can see in the graphs, the course of the SPL curves is almost identical, apart from the difference in levels for the midrange, resulting from the measurement of only one of the four speakers in the Revox speakers.
The question returns: what does the length of the instrument decay phase depend on? Is it about, for example, the damping of the speaker membrane? I almost forgot: out of the tested speakers, the Revox are the ones that "eat" the details of the sound. In addition, they have their own character, which is always audible, and the Wega speakers "disappear", they are almost transparent to the sound, they do not add a strong signature to it. Both play very pleasantly, but the difference in sound detail is huge in Wega's favor. And this despite the fact that the Revox-s have undergone recapping to MKP capacitors, and the Wega still have factory-fitted Wego electrolytic capacitors in the entire crossover. Any hint will be valuable - besides, this is a topic for an interesting discussion 🙂 And now I will try to paste photos 😉
For tweeters (X axis - PC volume sent to DAC and amp):
For the midrange:
And now midrange with reduced volume increments and reduced room noise floor.
...and well, above graph says that Revox is a speaker that should transmit more silent details to the listener as the difference between quiet and very quiet sounds is lower than in Wega. The reality shows the other way 😉 And what could I do to encourage Revox four tiny cones to show more silent nuances of the sound - no idea as this is not about sensitivity of a speaker. That must be something else - more sensitive speaker of course would play the decay phase louder but that is all relative to the attack phase which would be also louder played.
Regards,
Mike
and I would like to discuss a topic that is often mistakenly associated with the dynamic range of loudspeakers in my opinion. It is about the subjectively perceived by the listener decay time of instruments. In some loudspeakers, for example, a hi-hat decays for a very long time, and in others, although the strike itself is not less loud, the reverberation disappears very quickly and a clean, empty background appears. I thought that these loudspeakers might differ in the SPL characteristic as a function of the supplied electrical power - a loudspeaker that produces a long reverberation would have this characteristic more linear, and the one that produces less reverberation, falling more abruptly towards the muting of the sound level.
Since I could not find such graphs on the Internet, only a typical SPL graph in the frequency domain, I decided to make simple measurements myself. I connected the amplifier to a computer that played white and then pink noise, and I measured the sound intensity from the near field to eliminate the influence of the room. I increased the volume from zero to a specific level, then repeated one of the measurements with reduced amplifier gain, i.e. the volume intervals were smaller. And what?
These graphs show that the faster decay of instrument reverberation in some loudspeakers and slower in others does not result from the slope of their measured SPL characteristics vs. the supplied electrical power. So it is not any dynamic range of the loudspeakers or their SPL given in catalogs that decides the phase of the instruments' decay, because even if some loudspeakers play louder overall, the reverberation is received relatively to the impact and also decays relatively to the level of the impact (attack phase).
And my question: what, in your opinion, can decide that some loudspeakers beautifully convey the nuances of sound with a long phase of instrument decay, while others leave this information "to themselves"? PS. I used the Revox BX-350 and Wega Direct 200 speakers for the measurements. Both have dome tweeters, with the Wega being 3-way and the Revox being 2-way. The midrange measurement in the Revox speakers concerned one of the 4 parallel mid-woofers, while in the Wega - one midrange dome. The Revox BX-350 have 4 paper speakers and one tweeter, and the crossover frequency is 3200 Hz. As you can see in the graphs, the course of the SPL curves is almost identical, apart from the difference in levels for the midrange, resulting from the measurement of only one of the four speakers in the Revox speakers.
The question returns: what does the length of the instrument decay phase depend on? Is it about, for example, the damping of the speaker membrane? I almost forgot: out of the tested speakers, the Revox are the ones that "eat" the details of the sound. In addition, they have their own character, which is always audible, and the Wega speakers "disappear", they are almost transparent to the sound, they do not add a strong signature to it. Both play very pleasantly, but the difference in sound detail is huge in Wega's favor. And this despite the fact that the Revox-s have undergone recapping to MKP capacitors, and the Wega still have factory-fitted Wego electrolytic capacitors in the entire crossover. Any hint will be valuable - besides, this is a topic for an interesting discussion 🙂 And now I will try to paste photos 😉
For tweeters (X axis - PC volume sent to DAC and amp):
For the midrange:
And now midrange with reduced volume increments and reduced room noise floor.
...and well, above graph says that Revox is a speaker that should transmit more silent details to the listener as the difference between quiet and very quiet sounds is lower than in Wega. The reality shows the other way 😉 And what could I do to encourage Revox four tiny cones to show more silent nuances of the sound - no idea as this is not about sensitivity of a speaker. That must be something else - more sensitive speaker of course would play the decay phase louder but that is all relative to the attack phase which would be also louder played.
Regards,
Mike
Mike,
Damping would be one aspect of many that affect a loudspeaker's transient response, and the nature of it's decay, known as "impulse response".
The ratio between TS parameters "Bl" (motor strength) and "Mms" total mass of the moving parts of a speaker driver, including the air load) can be used to give a rough idea of transient response of drivers with similar impedance, materials and design.
Higher Bl to Mms ratios generally result in shorter decay times, but different materials and construction result in different stiffness to weight ratios also affecting decay time and resonant frequencies.
This paper goes into some of the aspects affecting transient response:
https://www.bksv.com/media/doc/17-198.pdf
This paper goes into how transient problems can be addressed electronically:
https://www.fulcrum-acoustic.com/wp...dspeaker-transient-response-with-dsp-2005.pdf
There are some mechanical things that could be done to improve your speakers transient response, but more measurement detail would be required to determine what would be appropriate.
REW (Room EQ Wizard) is a free software that makes phase and impulse response measurements possible:
https://www.roomeqwizard.com/
The time decay differences between your Revox BX-350 and Wega Direct 200 speakers will be more apparent using a tool that can measure them.
Art
Exactly more noted with Impulse Response.
Most errors in Transient will be at lower frequencies.
With a heavy cone for low distortion woofer it takes a very powerful magnet to keep Transients accurate.
Not only to start movement with the signal, but also stop movement when the signal ends.
If the motor has very good control, the impedance rise in typical reflex boxes will shoot to the moon.
Indicates a powerful motor with control.
And not much different carried up to midrange and treble.
Powerful motors and low distortion cone design/suspension in a nutshell
Problem with most " Simulation" and wild guessing. They are looking at response accuracy or transfer function of the compared filter relation.
Most people are obsessed with" more bass" or " Flat response" and is not the purpose of alignments to filter responses
The fake alignments they invent to make a " pretty picture" is killing real world transient.
.4 to .5 Qts drivers are starting to move towards a weak motor, actually both already relatively weak. But is typical magnet for many drivers.
People get confused in Sim. They make the box too small and tune too high for a " Flat" response.
They think the pretty picture is " good" but in reality making a weak motor slower with poor transient.
The so called " Soundstage" and " Imaging" magic myths is usually in the recording. It is stereo.
Reverb and Delay is rather common in recordings. And numerous widening effects are used.
This is usually the so called " imaging". The " no box" nonsense theory's of open baffle makes the sound more reverberant.
Or live in the room. Imaging is in the recording. Not skinny junk baffles, or dipole peaks and out of phase room reflections.
The producers add the fake delays/ reverb and panning to desired liking. For" open baffle" find a reverb plugin to your liking.
For forward or rearward " image" reverse left or right, for " wide" image spread apart the speakers till the vocals comb more.
Since they are typical centered and the delays are panned. AKA reverb effects
Most errors in Transient will be at lower frequencies.
With a heavy cone for low distortion woofer it takes a very powerful magnet to keep Transients accurate.
Not only to start movement with the signal, but also stop movement when the signal ends.
If the motor has very good control, the impedance rise in typical reflex boxes will shoot to the moon.
Indicates a powerful motor with control.
And not much different carried up to midrange and treble.
Powerful motors and low distortion cone design/suspension in a nutshell
Problem with most " Simulation" and wild guessing. They are looking at response accuracy or transfer function of the compared filter relation.
Most people are obsessed with" more bass" or " Flat response" and is not the purpose of alignments to filter responses
The fake alignments they invent to make a " pretty picture" is killing real world transient.
.4 to .5 Qts drivers are starting to move towards a weak motor, actually both already relatively weak. But is typical magnet for many drivers.
People get confused in Sim. They make the box too small and tune too high for a " Flat" response.
They think the pretty picture is " good" but in reality making a weak motor slower with poor transient.
The so called " Soundstage" and " Imaging" magic myths is usually in the recording. It is stereo.
Reverb and Delay is rather common in recordings. And numerous widening effects are used.
This is usually the so called " imaging". The " no box" nonsense theory's of open baffle makes the sound more reverberant.
Or live in the room. Imaging is in the recording. Not skinny junk baffles, or dipole peaks and out of phase room reflections.
The producers add the fake delays/ reverb and panning to desired liking. For" open baffle" find a reverb plugin to your liking.
For forward or rearward " image" reverse left or right, for " wide" image spread apart the speakers till the vocals comb more.
Since they are typical centered and the delays are panned. AKA reverb effects
Last edited:
Art,
I highly appreciate your input, but what I mean is a decay phase that actually IS in the input signal to the speaker from the source and amp. Not the decay phase after impulse that is - theoretically - infinitely short and then nothing, null. What I am trying to resolve is an interpretation by the speaker of a decay phase of, say, hi-hat or vocal sibilants reverberation that is present in the source mix. Not a decay time of the excited speaker driver. I have REW and made measurements of both speakers, Revox and Wega, but the problem in my opinion regards sensitivity characteristic in an input electrical power domain. Impulse plot of the Revox loudspeakers built with those tiny 12 cm mini but heavy, alloy gasket and big magnet woofers is exceptionally good compared to one big woofer, the bass is very fast and precise. I found a test from the 70's of these speakers and at the moment am trying to find it and attach to the thread. Just need to convert pdf to jpeg. REW SPL plot in my listening room mirrors the one what German testers measured several decades ago - a big dropdown about 500 Hz - but that's how these speakers interact with a room. The Germans did some oscilloscope tests at certain frequency impulses that REW - as far as I know - can not provide.
The point is that Revox BX-350 are two-way so frequencies up to 3200 Hz (and above to some extent) come from the mid-woofers and Wega has a big dome midrange with no support from any ari volume from the back (the back of the driver is a sealed magnet). That 39 cm dome with 39 cm coil should be significantly lighter than Revox 12 cm cone. Interesting detail, Revox mid-woofers have FS of 400 Hz (!) and go down to 45 Hz without hesitating. Impedance plot on the test copy below shows two impedance apex points - and that is a sealed box. Should not work but it does 😉 The crossover schematics also are not as usual for the vintage Revox and other German speakers - it has a LC resonance trap instead of a low-pass capacitor parallel to the woofer and another LC bypass of the woofer's inductor. Someone tried to "help" these four woofers go at some point and to eliminate another issue, but when recapping I missed measuring the inductors and can not tell what frequencies are addressed in these circuits.
All the best,
Mike
The gated tone bursts are an interesting test. Pretty harmless I'd bet, even at full power from the amplifier - as long as you limit the number of cycles to that which you need to see on a scope, to show the decay part after the signal ends. Since that's the part of interest, you dont even have to trigger the scope as nicely as shown above. The 60Hz one looks like it took a cycle to "get going".
At work we had these fancy Chroma AC sources. I put 1 cycle of 120VAC into a little computer monitor woofer; no smoke, survived. Could do it a second time.
At work we had these fancy Chroma AC sources. I put 1 cycle of 120VAC into a little computer monitor woofer; no smoke, survived. Could do it a second time.
I think it is most probable the result of frequency response. The decaying sound of the instrument is more clear in a speaker that is relatively stronger in this region. This has noting to do with impulse response - well FR is tied to FR so - maybe one should say, it is not a short term problem like 1 ms aspects. It is a frequency response problem. Try to EQ them equal and listen again.
This is my take on this.
//
This is my take on this.
//
Did anyone consider the room is in the equation? A simple fact: the broad dispersion of e.g. a dome midrange will produce more sound power output than the narrow dispersion of an array of four midrange cones. Decay in a room is all about the decay of the sound power in it, experienced as typical SPL on a listening spot. Redo the exercise in anechoic conditions.
Hi, my bet is the revox has at least one resonance, likely multiple, which make the signature sound you mention and mask any "small detail". This is from observation that strong room mode peaks on bass really make the whole system sound suck, bass feels either too loud or too wimpy if one tries to balance it with a bass tone control. But the second the resonance is targeted and notched out with high Q PEQ suddenly the whole system seems to be alive, sound "improves way up to midrange" when the very attention grabbing resonance is taken out.
So, i bet driver parameters have nothing to do with the difference in sound, but the whole thing, the construct. The worse one has issues (resonances) and your own auditory system poops the perception.
Loudspeaker design basics - combine all filters (drivers, crossover, enclosure, active/digital) in a way to achieve an as even frequency and phase response as possible. A system cleanly designed will have a close to optimal decay behavior across most of the frequency range (only the bass area will always suffer from some decay, because of the inherently necessary virtual high pass of the system). Doesn't require expensive components to achieve a great result either - just knowledge.
This is how EQ changed FR in my listening room. This particular plot comes from the listening position in a typical home room treated with four absorbers and five resonant bass traps. The answer of the room remains bad - that's no doubt about it - but EQ does a lot of job here. Here are results for Revox BX-350 pre and after EQ. Wega after EQ in REW looks pretty much the same but subjective listening experience differs a lot, it really does. These are different worlds. I even tried to compare white and pink noise passed through a band pass filter and measure SPL.
A pair of Revox BX-350 raw in a listening room, listening position:
...same pair after parametric EQ:
Speakers with similar FR plots in REW in the same room and positions can bring a very different listening experience. That's why I'm looking for that one missing parameter - if one exists - determining decay/reverb phase reproduction by the speaker. From my experience and experimenting it is not matter of frequency response characteristic. Of course, dark sounding speakers can kill HF reverb but that's extreme. With Revox and Wega examples I consider two pairs of speakers, each pair with its own EW as far to the neutral as possible without filter ringing or messing with phase. The EQ is an APO.
I enjoy the discussion going into more and more nuances 😉
Best,
Mike
It does exist but it is called interaction of room acoustics with directivity pattern of loudspeakers. Try to get hold of cardioid loudspeakers if you want the short decays, or get monopole speakers if you want the reverb of the room.
And now some photos, people like photos 😉
The protagonist and suspected: Revox BX-350:
...its tiny 16 Ohm and 400 Hz Fs drivers in detail:
...tweeter - fixing unglued wire with a superglue (cyano acrylate) and this is a bit different story, once I will run another thread with video showing exciting the tweeter to a resonance with a brush:
...sealing the tweeter back to the cabinet:
...x-over recapped with all-MKP from Monacor, back plate reinforcements and new banana-terminals:
...and the crossover schematic. It is evident that the designer tried to fix something and incorporated an LC resonant trap with 4,7 uF polarized elcap plus a bandpass boosting LC circuit with 47 uF cap but as I do not have inductor values it is hard to tell where exactly it is supposed to act:
...I did some experimenting with moving mic and the 500 Hz dip disappears about half meter from the baffle front of the Revox speakers. This is not lobing and from my point of view has something to do with a speaker-room system. German test from the 70's shows the same 500 Hz problem. As I have no access to the anechoic conditions, that's what I can provide to further discussion.
Best,
Miki
The protagonist and suspected: Revox BX-350:
...its tiny 16 Ohm and 400 Hz Fs drivers in detail:
...tweeter - fixing unglued wire with a superglue (cyano acrylate) and this is a bit different story, once I will run another thread with video showing exciting the tweeter to a resonance with a brush:
...sealing the tweeter back to the cabinet:
...x-over recapped with all-MKP from Monacor, back plate reinforcements and new banana-terminals:
...and the crossover schematic. It is evident that the designer tried to fix something and incorporated an LC resonant trap with 4,7 uF polarized elcap plus a bandpass boosting LC circuit with 47 uF cap but as I do not have inductor values it is hard to tell where exactly it is supposed to act:
...I did some experimenting with moving mic and the 500 Hz dip disappears about half meter from the baffle front of the Revox speakers. This is not lobing and from my point of view has something to do with a speaker-room system. German test from the 70's shows the same 500 Hz problem. As I have no access to the anechoic conditions, that's what I can provide to further discussion.
Best,
Miki
Provide a file with the results of REW measurements of both speakers with correction and necessarily with phase, THD and noise, obtained on a sweep tone. If you do not know how to measure the frequency response with phase in REW, then I can tell you how to get them in a single-channel measurement mode.
P.S. If you are looking for reverberation, then you need to look for it on other graphs. The frequency response does not show reverberation.
Maybe I don't understand the question correctly?
It seems to me that if we have two speakers, one speaker sounds good, and the second speaker sounds bad, then having received the measurements of these two speakers, we can compare them and find the differences between the two speakers. Then, looking at the differences, we can put forward a theory about the influence of this or that difference on the sound.
That's not exactly the point that one speaker sounds "bad" and another "good", they just sound completely different when it comes to reproducing recorded reverberation from a source material. Here you are with requested measurements. Listening position, single mic, REW, with/without EQ, Revox and Wega. Measurements level may differ because of different amp vol. So here we go 😉
1R) Revox BX-350 SPL and Phase, no EQ:
2R) Revox BX-350 SPL and Phase, EQ:
3R) Revox BX-350 Distortion, EQ off:
4R) Revox Distorion EQ on:
And now for Wega:
1W) Wega Direct 200, SPL and Phase, EQ off:
2W) Wega Direct 200, SPL and Phase, EQ on:
3W) Wega, Distortion, EQ off:
4W) Wega, Distortion, EQ on:
Best,
Mike
What is the problem with giving the source file in REW? You can give two files, I will combine them and analyze the measurements in a more convenient form than you provide.
I don't understand why it should be different. If we record two different speakers, they will sound different in the recording. Otherwise, if two speakers sound the same in the recording, then they are the same. The fact that you enter correction in the frequency response at the listening point does not mean that you made the same frequency response for two different speakers in the sound radiation zone. This can be easily seen by measuring the speakers in the near field after correcting the frequency response at the listening point. And this is only one of the reasons leading to different perception of the recorded material, there is also THD in speakers, parasitic resonances of speakers, there are crossover filter settings and different radiation patterns of speakers. There is not only one criterion for a speaker that can show which speaker does not steal reverberation ....