Hope this doesn't ignite a flood of debate..........
The term "baffle step loss" has always confused me in it's use.
The term loss in my mind would equate to some acoustic reduction in output of a loudspeaker at the listening point.
I understand that a woofer (fo example) in a given box has a much lower level of bass output compared to midrange (6db).
If you take a lose look at any frequency response graph showing output on a baffle, the bass though lower than the midrange is still at the output of the driver's unbaffled response.
Take for example the Peerless HDS 6.5inch woofer.
Let's say it's characteristic sensitivity at 2.83volts/rms at 1 metre is 88db. The woofer is placed in a ported box which is tuned correctly and from say 400hz to 800hz it's output when 2.83 volts is applied is 88db. At the point after baffle step happens the response is 94db at the same voltage.
It seems rather unlogical to refer to this phenomenon as a "loss".
The listener (ie normal people who site in front of a speaker to listen to it not beside it), actually receives more acoustic output than if the woofer where unbaffled!!!
I know that there is a loss of acoustic output off-axis to the speaker i.e waves become directional, but since we have to assume that a listener in the near or far-field will be relatively on-axis to the speaker system then this would still not be percieved as a loss.
I propose to change the term to something like "baffle step directivity change" or something along those lines since it can be argued that there is actually no net increase or decrease of acoustic energy fro the whole system.
Just my two bobs worth.
MADINOZ
The term "baffle step loss" has always confused me in it's use.
The term loss in my mind would equate to some acoustic reduction in output of a loudspeaker at the listening point.
I understand that a woofer (fo example) in a given box has a much lower level of bass output compared to midrange (6db).
If you take a lose look at any frequency response graph showing output on a baffle, the bass though lower than the midrange is still at the output of the driver's unbaffled response.
Take for example the Peerless HDS 6.5inch woofer.
Let's say it's characteristic sensitivity at 2.83volts/rms at 1 metre is 88db. The woofer is placed in a ported box which is tuned correctly and from say 400hz to 800hz it's output when 2.83 volts is applied is 88db. At the point after baffle step happens the response is 94db at the same voltage.
It seems rather unlogical to refer to this phenomenon as a "loss".
The listener (ie normal people who site in front of a speaker to listen to it not beside it), actually receives more acoustic output than if the woofer where unbaffled!!!
I know that there is a loss of acoustic output off-axis to the speaker i.e waves become directional, but since we have to assume that a listener in the near or far-field will be relatively on-axis to the speaker system then this would still not be percieved as a loss.
I propose to change the term to something like "baffle step directivity change" or something along those lines since it can be argued that there is actually no net increase or decrease of acoustic energy fro the whole system.
Just my two bobs worth.
MADINOZ
Checkout the Seas drivers measured in boxes, e.g. :
http://www.seas.no/seas_line/woofers/H571.PDF
the baffle step is obvious, the point is the measured bass
response is lower than it would be mounted on a large
"infinite" baffle , hence the term baffle step loss.
The bass level is 6dB below published sensitivity specs,
which is measured half field, not free field.
Your arguement regarding the Peerless woofer is incorrect.
If its sensisitivity is 88dB/2.83V this is above baffle step.
(As the driver is measured on a large baffle).
In the bass effective levels will be 82dB/2.83V.
sreten.
http://www.seas.no/seas_line/woofers/H571.PDF
the baffle step is obvious, the point is the measured bass
response is lower than it would be mounted on a large
"infinite" baffle , hence the term baffle step loss.
The bass level is 6dB below published sensitivity specs,
which is measured half field, not free field.
Your arguement regarding the Peerless woofer is incorrect.
If its sensisitivity is 88dB/2.83V this is above baffle step.
(As the driver is measured on a large baffle).
In the bass effective levels will be 82dB/2.83V.
sreten.The graph for the SEAS woofer has no absolute dB shown on the
dB scale therefore it is impossible to work out the absolute levels of output at any given frequency for a given voltage.
Please find attached a graph (copyright BESL) which shows a Vifa PL18 woofer in a Madisound box (from DIY2000).
You can note that the driver was placed into the speaker box and was when driven with a 2.83volt signal it exhibited a sensitivity of around 87/db/w/m which is basically Vifa's published sensitivity for this driver.
It's after the baffle step point is reached that the output rises to the +6db level.
You can view the full set of graphs at
http://www.t-linespeakers.org/tech/filters/Jeff/.
MADINOZ
dB scale therefore it is impossible to work out the absolute levels of output at any given frequency for a given voltage.
Please find attached a graph (copyright BESL) which shows a Vifa PL18 woofer in a Madisound box (from DIY2000).
You can note that the driver was placed into the speaker box and was when driven with a 2.83volt signal it exhibited a sensitivity of around 87/db/w/m which is basically Vifa's published sensitivity for this driver.
It's after the baffle step point is reached that the output rises to the +6db level.
You can view the full set of graphs at
http://www.t-linespeakers.org/tech/filters/Jeff/.
MADINOZ
Attachments
madinoz said:
The thing is... The sensitivity figures that are published are valid for an *infinite* baffle, or half space. At least those of Peerless, but also of most others. Decreasing the baffle size to that of an ordinary box, causes a *loss* of level in the bass. So, it's all a matter of selecting the viewpoint.
While we are at it; the effect is hardly a step function either. A step function goes *instantaneously* from one value to the other, that is how the matematicians define it anyway. The baffle "step" is very smooth.
So, Id suggest "finite baffle bass loss" or something, if I had to. But really, everyone knows what the baffle step is, so why confuse it?
BTW, what does EMF mean?
madinoz said:
If you read the Seas data sheet completely you will see it it
measured at 0.5m but driven by 2V to be equivalent to 1m.
0dB is 60dB as specified on the sheet.
It is the graph you mention that has no calibration.
I'll also note all the other graphs in the article show results
which plainly agree with the real situation as outlined in my
previous post.
sreten.
???? 0.5 woofer at back BSC
I'm trying to come to terms with baffle step compension using a 0.5 woofer at the back of the cabinet.
If we have a baffle of 200mm wide, a common size, then the baffle step frequency would be 575Hz. Is this the centre of the baffle step or where it starts to fall off? What is the effect on frequencies either side? That is, does the step go out to the woofer bottom end roll off and then rise again at the midrange at some frequency above 575Hz, a multiple of 575Hz or 575Hz + xHz???
If you used a 0.5 woofer, does it have to be the same as the front woofer with the same F3 or can it be a smaller woofer that has a higher F3 which rolls off at 575Hz + xHz at 6dB?
If you use a 0.5 woofer in parallel with the front woofer, I take it the sensitivity of the speaker remains the same as you are only using the 0.5 to cover the baffle step frequencies. Whereas if you used 2 woofers that cover the same range then the sensitivity goes up. So I assume the L-pad for the tweeter would not be effected.
Questions, questions...... have read a lot on the subject using passive filter compensation as well as using a 0.5 woofer but am having trouble puting the puzzle pieces together.
I'm lost in my own post........ I hope you understand what I mean.
I'm trying to come to terms with baffle step compension using a 0.5 woofer at the back of the cabinet.
If we have a baffle of 200mm wide, a common size, then the baffle step frequency would be 575Hz. Is this the centre of the baffle step or where it starts to fall off? What is the effect on frequencies either side? That is, does the step go out to the woofer bottom end roll off and then rise again at the midrange at some frequency above 575Hz, a multiple of 575Hz or 575Hz + xHz???
If you used a 0.5 woofer, does it have to be the same as the front woofer with the same F3 or can it be a smaller woofer that has a higher F3 which rolls off at 575Hz + xHz at 6dB?
If you use a 0.5 woofer in parallel with the front woofer, I take it the sensitivity of the speaker remains the same as you are only using the 0.5 to cover the baffle step frequencies. Whereas if you used 2 woofers that cover the same range then the sensitivity goes up. So I assume the L-pad for the tweeter would not be effected.
Questions, questions...... have read a lot on the subject using passive filter compensation as well as using a 0.5 woofer but am having trouble puting the puzzle pieces together.
I'm lost in my own post........ I hope you understand what I mean.
Hi Rabbitz,
An MTM 2 way with full baffle step compensation will have the
same tweeter level as a 2.5 way.
2.5 way forces you to have 6dB BSC, whilst a MTM 2 way can
have less BSC, thus it can have higher tweeter sensistivity.
THE baffle step is approximated to be a slope either side of the
centre frequency, how to estimate the centre frequency ?, I've
seen various ratio used.
The F3's of the two woofers don't have to be the same, see here :
http://home.hetnet.nl/~geenius/Auriga.html
The graphs should help somewhat,
sreten.
(P.S. if you place the extra bass unit on the back of the box
- do you mean like here ?
http://www.t-linespeakers.org/projects/tlB/intro.html
This gets confusing as the monopole and bipole cases are different.)
An MTM 2 way with full baffle step compensation will have the
same tweeter level as a 2.5 way.
2.5 way forces you to have 6dB BSC, whilst a MTM 2 way can
have less BSC, thus it can have higher tweeter sensistivity.
THE baffle step is approximated to be a slope either side of the
centre frequency, how to estimate the centre frequency ?, I've
seen various ratio used.
The F3's of the two woofers don't have to be the same, see here :
http://home.hetnet.nl/~geenius/Auriga.html
The graphs should help somewhat,
sreten.(P.S. if you place the extra bass unit on the back of the box
- do you mean like here ?
http://www.t-linespeakers.org/projects/tlB/intro.html
This gets confusing as the monopole and bipole cases are different.)
sreten said:
Hmm... You *could make it "2.7 way", I mean just *lower* the level of one driver towards higher frequencies.
The first order filter, that would be just an inductor in the 2.5 case would now be an inductor in parallel with a resistor. But then, we are almost at the "MTM-with-BSC" design... Oh well... Svante said:
Hmm... You *could make it "2.7 way", I mean just *lower* the level of one driver towards higher frequencies.
Hi Svante, ( I never get away with my generalisations...)
True, but your asking for trouble with the usual TMM
2.5 way set up allowing midrange thorough to the
lower driver. In a MTM 2.5 & a bit way this is less of
an issue but it does complicate the mid/treble c/o.
For the MTM 2.5+way :
If you are not careful the 0.5+ way driver will have more
output above the c/o frequency than the main unit, if you
only add a parallel resistor.
So you need to c/o this driver as well if you want it to work well.
Getting very complicated compared to the 2-way MTM case.
sreten.
A small baffle is like an ANTI-HORN
There are a couple of things people are forgetting here:
1)An ordinary lounge is NOT an anechoic chamber, and in fact most of the sound you end up hearing comes from reflections and not directly from the speaker whether you like it or not. That -6dB approximation of baffle-step loss only applies in outdoor or anechoic situations. In a room a big factor is how far you are from the speakers, which influences the ratio between direct sound and reverberant sound. The reverberant sound might easily make up for bass losses in the on-axis sound.
2)The REAL baffle-step bass "loss" is caused by something that I haven't heard anyone ever mention, and that is the air load on the speaker's cone. Note this: horn speakers are very efficient because they dramatically increase the air-load on the cone versus other losses such as heat, by forcing the cone to move a larger mass of air longitudinally. A large baffle has a similar effect because the air near the cone has less room to move and hence sound pressure is higher. At low frequencies the sound pressure is lower due to diffraction and the total real acoustical power drops. This is different to the usual excuse about the shape of the baffle influencing the directivity of different frequencies and an arbitrary "6dB" drop.
Anyone who can understand why a horn speaker design is more efficient compared with a flat baffle should understand that low frequencies diffracting around the box is like an "anti-horn". I only recently figured out why horn speakers are more efficient than other designs, and I think it's relevant.
A simple electrical equivalent to a speaker with a flat baffle would be a closed circuit containing a current source, a high resistance load (low air load) and low internal resistance in series (internal speaker losses). By increasing the external conductivity (attaching a horn), the output power can be dramatically increased even if the total power dissipated is the same. When the 2 resistances are equal the efficiency can't be increased anymore. (Please correct my example someone, I think I might be wrong somewhere!)
One interesting thing I noticed about the specs of Seas drivers, is that they quote the mass of the air load, eg: 1 gram. Obviously this will be different at different frequencies, and like I said will be influenced by diffraction. Everyone else has mentioned that SPL is lower at low frequencies, merely because the same acoustical power is spread over a wider area.
CM
There are a couple of things people are forgetting here:
1)An ordinary lounge is NOT an anechoic chamber, and in fact most of the sound you end up hearing comes from reflections and not directly from the speaker whether you like it or not. That -6dB approximation of baffle-step loss only applies in outdoor or anechoic situations. In a room a big factor is how far you are from the speakers, which influences the ratio between direct sound and reverberant sound. The reverberant sound might easily make up for bass losses in the on-axis sound.
2)The REAL baffle-step bass "loss" is caused by something that I haven't heard anyone ever mention, and that is the air load on the speaker's cone. Note this: horn speakers are very efficient because they dramatically increase the air-load on the cone versus other losses such as heat, by forcing the cone to move a larger mass of air longitudinally. A large baffle has a similar effect because the air near the cone has less room to move and hence sound pressure is higher. At low frequencies the sound pressure is lower due to diffraction and the total real acoustical power drops. This is different to the usual excuse about the shape of the baffle influencing the directivity of different frequencies and an arbitrary "6dB" drop.
Anyone who can understand why a horn speaker design is more efficient compared with a flat baffle should understand that low frequencies diffracting around the box is like an "anti-horn". I only recently figured out why horn speakers are more efficient than other designs, and I think it's relevant.
A simple electrical equivalent to a speaker with a flat baffle would be a closed circuit containing a current source, a high resistance load (low air load) and low internal resistance in series (internal speaker losses). By increasing the external conductivity (attaching a horn), the output power can be dramatically increased even if the total power dissipated is the same. When the 2 resistances are equal the efficiency can't be increased anymore. (Please correct my example someone, I think I might be wrong somewhere!)
One interesting thing I noticed about the specs of Seas drivers, is that they quote the mass of the air load, eg: 1 gram. Obviously this will be different at different frequencies, and like I said will be influenced by diffraction. Everyone else has mentioned that SPL is lower at low frequencies, merely because the same acoustical power is spread over a wider area.
CM
Hi CM,
Your point 2 is an accurate way of putting it.
But I'd use your own arguments in point 2 and later to
suggest that room reverberation cannot make up for the
BSL, as you suggest in point 1.
At low frequencies less acoustic energy is being coupled
into the room. As you say it is not simply the case of the
same amount of acoustic energy "folding" around the
loudspeaker.
The concencus of speaker builders is that for free
space placement clear of walls full BSC is required.
(For a compact monitor type box loudspeaker).
Reducing the amount of BSC requires the speakers
to be placed near, up to next to, the rear wall.
sreten.
Your point 2 is an accurate way of putting it.
But I'd use your own arguments in point 2 and later to
suggest that room reverberation cannot make up for the
BSL, as you suggest in point 1.
At low frequencies less acoustic energy is being coupled
into the room. As you say it is not simply the case of the
same amount of acoustic energy "folding" around the
loudspeaker
.The concencus of speaker builders is that for free
space placement clear of walls full BSC is required.
(For a compact monitor type box loudspeaker).
Reducing the amount of BSC requires the speakers
to be placed near, up to next to, the rear wall.
sreten.
OK, the radiation resistance, or the load, is another way of looking at it. The radiation resistance doubles in the presence of a mirroring plane and so does the radiated power, so here is a 3dB gain. The other 3 dB, which makes the baffle step totally 6 dB, comes form the increased directivity, ie only half space is "lit" by the driver.
Note that this directivity is *not* caused the driver (this is completely neglected in this model) but the baffle and the edge reflections.
So, the loading business only accounts for half of the baffle step.
One could say that a small baffle is like a no-baffle for low frequencies, and as a big-baffle for high frequencies. To call it a anti-horn is clearly seen from the perspective that the horn is the normal way of connecting a driver to the air, something that many of us simply don't do. But still, it is one way of seeing it, yes.
As for the anechoic environment assumed for the baffle step compensation, yes you are right about this, to some extent. A speaker place near a wall does exhibit a bass lift of 6dB, which would compensate for the baffle step (if the speaker was placed *in* the wall). A complete room complicates the picture even more. The excuse that can be made to ignore the room reflections is that we have a capability of separating early reflections from late reflections. If a person speaks in a room you have no problem hearing *where* he speaks from, still the majority of the acoustic energy that hits you has been reflected by the walls. This is psycho-acoustics and is called the precedence effect. So, there is a difference compensating for the baffle step (which is formed by early reflections) and the wall reflections (which are late-ish).
And third, the load resistance of the horn. Actually, if the load resistance increases *beyond* the source resistance the *efficiency* can in theory reach 100%, but the *power* in RHorn starts to decrease at Rsource=RHorn. This is very simplified, and a lot of things has to be taken into consideration to understand horn design completely, though.
Note that this directivity is *not* caused the driver (this is completely neglected in this model) but the baffle and the edge reflections.
So, the loading business only accounts for half of the baffle step.
One could say that a small baffle is like a no-baffle for low frequencies, and as a big-baffle for high frequencies. To call it a anti-horn is clearly seen from the perspective that the horn is the normal way of connecting a driver to the air, something that many of us simply don't do. But still, it is one way of seeing it, yes.
As for the anechoic environment assumed for the baffle step compensation, yes you are right about this, to some extent. A speaker place near a wall does exhibit a bass lift of 6dB, which would compensate for the baffle step (if the speaker was placed *in* the wall). A complete room complicates the picture even more. The excuse that can be made to ignore the room reflections is that we have a capability of separating early reflections from late reflections. If a person speaks in a room you have no problem hearing *where* he speaks from, still the majority of the acoustic energy that hits you has been reflected by the walls. This is psycho-acoustics and is called the precedence effect. So, there is a difference compensating for the baffle step (which is formed by early reflections) and the wall reflections (which are late-ish).
And third, the load resistance of the horn. Actually, if the load resistance increases *beyond* the source resistance the *efficiency* can in theory reach 100%, but the *power* in RHorn starts to decrease at Rsource=RHorn. This is very simplified, and a lot of things has to be taken into consideration to understand horn design completely, though.
Yeah, I've kind of tripped over on my own argument! However, I suspect that the variation in ultimate sensitivity directly caused by baffle-loading is around 1dB if not less. Perhaps it really is 3dB, considering the example where there can be up to 6dB gain when using twice as many drivers on a common baffle. This would be consistent in that when you use 2 drivers instead of 1, above a certain frequency the gain would be only 3dB due to summation of sounds with different phases.
Hmmm, I'm finding that I've reversed my own sceptical opinion on the subject.
CM
Hmmm, I'm finding that I've reversed my own sceptical opinion on the subject.
CM
Great points with the early and late reflections Svante! Come to think of it, I hate the delayed bass from some speakers, especially the room-loading of bass ports
I'm not a horn buff though, I've merely developed a curiosity. Perhaps a bass-horn may be forthcoming at some point.
An alternative is a very small baffle so that directivity and gain is consistent over a wide range of frequencies. Or an an extremely large baffle? But not medium sized.
With multi-way speakers (I don't buy that full-range madness!), the baffles should be somehow optimized, just like driver sizes are optimized. A bass speaker could have a really small baffle so it's consistently omnidirectional throughout its range, or it could have a wall-to-wall baffle. The midrange should have either a large-ish or small-ish baffle so that there isn't a step in sensitivity half-way through its frequency range, and the tweeter almost always has to have a relatively large baffle, as the small option is too small. An exception is a plasma tweeter which does have the option of being almost completely omnidirectional, but that's another story.
CM
I'm not a horn buff though, I've merely developed a curiosity. Perhaps a bass-horn may be forthcoming at some point.
An alternative is a very small baffle so that directivity and gain is consistent over a wide range of frequencies. Or an an extremely large baffle? But not medium sized.
With multi-way speakers (I don't buy that full-range madness!), the baffles should be somehow optimized, just like driver sizes are optimized. A bass speaker could have a really small baffle so it's consistently omnidirectional throughout its range, or it could have a wall-to-wall baffle. The midrange should have either a large-ish or small-ish baffle so that there isn't a step in sensitivity half-way through its frequency range, and the tweeter almost always has to have a relatively large baffle, as the small option is too small. An exception is a plasma tweeter which does have the option of being almost completely omnidirectional, but that's another story.
CM
CeramicMan said:
Yes, the baffle size should be larger than 17 metres or smaller than 17 millimetres if the speaker should be free from the baffle step and cover the entire range (20-20000Hz).
So, splitting the drivers to different baffles is in that sense a good idea, and indeed that is what we see done with subwoofers. On the other hand, the different drivers should be placed close to each other in order to avoid interference caused by delays between drivers at the crossover frequency. So in the end; mid-high drivers should be placed on separate baffles one large and one small and these baffles should be placed within centimetres from each other... Hmm... I wonder if that is why people compensate for the baffle step instead...
sreten said:
Hi sreten
Thanks for the reply. What I have in mind is more like the TLb arrangement with the 0.5 woofer (5") in the back in line with the existing woofer (6½").
Easy enough to experiment with as I'll just build a sealed box to shove at the back of the speaker and bi-wire it off the amp. Let the 0.5 woofer roll off on the bottom and play with inductors for the upper end starting off at 575Hz and going ½ an octave each side to see what happens.
Ta
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.