Sorry Sane, it's the best REW can do now. I certainly wasn't trying to be misleading in the smoothing depo as it's written on the graph so that anyone who can see, can see. The old dipole graph I just don't have the data saved for, just a graph. I could have done this:
That is misleading and actually is the no smoothing applied option in REW5, the new version. It still looks 1/48 to me however so I thought it would be misleading to post it that way. My point is/was that polar response can be seen in the reflections and the midbass of the dipole actually looks better. The modal excitation looks smoother as well in the graph.
Wayne gets 10 points for the right height.
Oh, I just remember I still have the old version of REW on the hard drive so I can make this more equal as to avoid upsetting anyone I'll make the graphs more similar if I can. This will take a minute..............
Keep in mind that these graphs aren't exact, just similarly placed but radically different graphing speakers under anechoic conditions. Man the polar response is prevalent in those graphs. The on axis is fairly flat on both. The off axis only looks great on the top graph, but there's the bubble on the bottom's. The comb is also very prevalent on the dipole, but I don't know that I can say I ever really heard it with music. The forward sound is what bothered me with the dipole and I'm not sure that visible in the graph, but I know a certain female someone said it had a hollow sound. 😀 The old audiophile cliche. The graph could make you think the sound was dull. In any case, I attributed this to different terminology and the Orion-like polar response, but maybe it was the comb. I don't know, but I always thought that we had too high of time resolution in the treble for the visible comb to audibly matter. The Orion shelf sure seems like a good idea from my experience--not that it matters.
Dan
That is misleading and actually is the no smoothing applied option in REW5, the new version. It still looks 1/48 to me however so I thought it would be misleading to post it that way. My point is/was that polar response can be seen in the reflections and the midbass of the dipole actually looks better. The modal excitation looks smoother as well in the graph.
Wayne gets 10 points for the right height.
Oh, I just remember I still have the old version of REW on the hard drive so I can make this more equal as to avoid upsetting anyone I'll make the graphs more similar if I can. This will take a minute..............
Keep in mind that these graphs aren't exact, just similarly placed but radically different graphing speakers under anechoic conditions. Man the polar response is prevalent in those graphs. The on axis is fairly flat on both. The off axis only looks great on the top graph, but there's the bubble on the bottom's. The comb is also very prevalent on the dipole, but I don't know that I can say I ever really heard it with music. The forward sound is what bothered me with the dipole and I'm not sure that visible in the graph, but I know a certain female someone said it had a hollow sound. 😀 The old audiophile cliche. The graph could make you think the sound was dull. In any case, I attributed this to different terminology and the Orion-like polar response, but maybe it was the comb. I don't know, but I always thought that we had too high of time resolution in the treble for the visible comb to audibly matter. The Orion shelf sure seems like a good idea from my experience--not that it matters.
Dan
Your probably right but I don't have the simulator and I'm more interested in the measurements. The last comparison wasn't fair at all and it only took a few minutes to take these measurements so as to produce a valid comparison/contrast to the previous presentation.
I've been meaning to measured my commercial speakers anyways... I have always been curious how my $1k+ commercial dynamic monopoles would compare to my DIY dipole setup and this discussion pushed me to do it.
Ah! That's where the passion is coming from. You missed my point blinded by your belief of what mine was. Reread my post for evidence. If you'd thought about where the reflections are coming from in that graph, you could have made the argument for the 'good' horizontal reflections and that our hearing is too acute in the time domain to hear the comb filter in the MR/treble.
next time perhaps eh?😉
Dan
I should have used the word "brand name" speaker instead of "commercial".. the monopole isn't my design and I'm not likely to try to commercialize my current dipole.
My point was that readers here can't really make any conclusion from the graphs and I'm not sure how you could either. (you can really see the polar response in that?)
I didn't think it was useful to compare graphs that were not comparable and perhaps could be construed as misleading. At least the graphs above are using what appears to be a similar resolution/smoothing and so they are *somewhat* comparable but they are still using a completely different scale so to make any sort of conclusion based on them is still questionable IMO.
semantics.... no matter to me.
I mentioned the conclusions as well, but horses for courses. Just giving some interesting data points that were actually intended to help your case. I'll never do that again. 😉 I don't desire to offend those whose positions I intend to support even if only for fun.
They are not on a completely different scale--look again and count. The graph shape is the difference between the 2 host sites as far as I know. Not that it matters one bit. They'll say the same thing regardless. I tried to help you even more, but it's certainly not that important to me. Dipole believers seem more passionate than sane and rational.
Yes I can see the polar response in mine--but I have also seen the polar responses and that gives me a great advantage. A couple reflection areas disagree with the polar response. I can't see the polar graph in yours, but I did correctly guess the dipole. It has the same advantages as the dipole in my graph. Interesting that the comb is not pronounced. Can you hear the comb on a sine sweep? With mine it's clearly audible. Can't say the same with music, but its presentation would be different.
Dan
I mentioned the conclusions as well, but horses for courses. Just giving some interesting data points that were actually intended to help your case. I'll never do that again. 😉 I don't desire to offend those whose positions I intend to support even if only for fun.
They are not on a completely different scale--look again and count. The graph shape is the difference between the 2 host sites as far as I know. Not that it matters one bit. They'll say the same thing regardless. I tried to help you even more, but it's certainly not that important to me. Dipole believers seem more passionate than sane and rational.
Yes I can see the polar response in mine--but I have also seen the polar responses and that gives me a great advantage. A couple reflection areas disagree with the polar response. I can't see the polar graph in yours, but I did correctly guess the dipole. It has the same advantages as the dipole in my graph. Interesting that the comb is not pronounced. Can you hear the comb on a sine sweep? With mine it's clearly audible. Can't say the same with music, but its presentation would be different.
Dan
sorry... I can't seem to add smilies in my posts any more to give perception of less negativity. No offense intended nor taken... just a rigorous debate.
Beside my computer is not positioned by microphone, I mostly try not to listen to the sine sweeps ... but maybe could hear it if I did slower sweep and positioned speaker with more direct reflection and less diffusion (I do not have entirely a flat wall behind).
Last edited:
I listened to the sims. They demonstrate nicely a source of extra "zippiness" or liveliness I hear in OB sound. I think folk find it attractive because it's somewhat reminiscent of a live performance space which has a large amount and variety of reflections and thus gives an impression of "naturalness". The downside, in my view, is that it tends to mess with timbre.
Even coherent sounds far apart in angle and time are not perceived as separated. There's still one sound sensation. If lumping together direct sound and reflections would represent our hearing then we would hear a flanging sound whenever our ears move but we do not. This particular behavior of our hearing first needs to be fully understood before we can draw conclusions.
Nevertheless I do believe that front wall reflections are detrimental, not because they lead to colored sound but because they affect localization of phantom sources.
How far apart in time ?
I think you're jumping to false conclusions to say we never hear flanging effects when moving our ears - in the specific case of equidistant stereo listening flanging is actually a significant problem, especially if the angular separation of the speakers at the listener is significantly less than +/- 25 degrees or so.
Have you ever listened to mono pink noise over a stereo speaker pair at the "ideal" listening position and tried moving your head sideways ? You'll hear a flanging effect. With narrow speaker separation angles the effect will be obvious, with a wider angle (+/- 30 or more) it will be subtle but still present.
It's worse with narrower speaker separation angle because direct path inter-aural crosstalk gets worse at higher frequencies the closer both speakers move towards the centre line.
However even with wide speaker separation if you turn your head at 90 degrees and perform the same test (this time moving your head forward and back - lateral relative to the speakers) you'll get obvious flanging.
We rely on the reduction of high frequency inter-aural crosstalk provided by our binaural hearing when listening to a typical widely separated stereo pair to minimize this problem, as well as the fact that flanging is much more difficult to detect on constantly varying sounds like music than it is on a steady signal like pink noise, but it's still there, and IMHO is one source of image instability.
For flanging to be audible the time delay needs to be very small though, on the order of 0.3ms or less, or a differential path length of under 10cm. The only thing which approximates this in the listening room is the differential delay between a stereo speaker pair listened to not quite at identical distances.
Reflections from any of the walls represent a much longer delay and don't cause a noticeable flanging effect with moderate movements of the head because the density of the peaks and dips from the comb filtering is far greater than the "critical bandwidth" of the ear in the higher frequencies and so get averaged out on typical broadband spectra, even though a microphone with a narrow band analyser will show them when it's moved.
You will still notice a difference at lower frequencies though - low midrange at typical side wall distances, but this won't sound like flanging because it takes a large movement of the head to cause a change in response, so it's perceived more as a shift in tonal balance with listening position in the room.
Flanging is caused by comb filtering, but not all comb filtering sounds like flanging depending on the time delay, and comb filtering is not even audible on most programme material once the peaks and dips are dense enough to be averaged by the critical bandwidth of the ear. (EG at a high enough frequency relative to the time delay)
I find the opposite actually - colouration of the lower midrange is the biggest problem I notice from having a speaker too close to the front wall, and as long as a small EQ change is made to restore balance, I don't usually notice any loss of localization of phantom sources. But then I tend to use speakers with very wide baffles, putting most if not all of the midrange above the baffle step frequency - which makes the speaker far less fussy about how close it is to the wall behind it..
Last edited:
At least not as severe as David's samples imply. Sound coloration caused by reflections is not the real issue. The real issue is ILD, ITD and pinna localization cues.
Within the time precedence is in effect.
Repetition pitch (see Bilsen).
Exactly my point.
The region between modal and statistical region is somewhat difficult to tackle. Simple steady-state equalization doesn't help anymore. I'd agree that speaker and room treatment needs to solve this.
Not sure I see the significance of this. "Repetition pitch" is just another way of describing the effects of flanging - peaks and dips in the response that are related to delay time. A peak that forms between two dips in a broad band noise will appear to have a sense of pitch to it. So ?
Except that it doesn't help your argument, because inter-aural crosstalk reduction reduces the amount of sound coming from the left hemisphere of the room entering your right ear at high frequencies - in Dave's example the reflection from the front wall will be coming from the same hemisphere as the direct signal, so no reduction due to this effect.
Well with a 50cm wide baffle you put the baffle step frequency at ~230Hz - right near the border of modal and statistical region in most rooms, so moving a speaker with that size baffle back towards the front wall primarily effects frequency at or below 230Hz, which makes compensation a whole lot easier.
Also, when I said applying some EQ, I didn't suggest that I would measure the steady state response, as that's very unlikely to yield a meaningful measurement that corresponds with what we hear in that frequency range. Instead I'd adjust it by ear on a speaker that was already well balanced before being pushed closer to the wall.
If you're already familiar with the tonal balance of the speaker out from the wall it's not hard with a little bit of trial and error to get the EQ right when it's pushed a bit closer to the wall.
The amount of correction needed isn't large, and the baffle step frequency of the speaker is a good guide as to how high in frequency you need to go with the correction.
Is it really that simple - direct sound and reflections coming from the same side will result in sound coloration whereas reflections coming from the other side are fine?
I didn't say that.
All I said is inter-aural crosstalk reduction won't help you when the reflection is coming from the same hemisphere as the direct signal, which is the case with a front wall or side wall reflection.
I only mentioned it in passing as a reason why flanging effects aren't usually noticed in playback despite their potential existing with any stereo speaker pairing, but you seized on it as a possible reason why rear wall reflections from dipoles could be discriminated against and ignored by our perception process - but since they're from the same hemisphere, and often almost exactly the same direction this can't be the case.
Besides, the cross talk reduction only happens at higher frequencies, (upper mid ~2Khz to treble) while the main frequencies of concern with wall distance comb filtering are lower midrange frequencies. Comb filtering at treble frequencies from wall distance reflections is not a concern (at least as far as flanging etc goes) because the peaks and dips are densely packed within the critical bands and get smoothed out, as I mentioned.
Personally I think all reflections (in playback, not recording) are colourations, but some are more detrimental than others, as I mentioned a few posts back. If you have to trade off speaker design and or room placement and acoustics to reduce one type of reflection at the expense of another, it's useful to know which are the more detrimental ones, which is what Dave was getting at earlier with the importance of floor bounce and front wall bounce reflections vs side wall reflections.
Last edited:
I'm pretty sure this can be the case with music. Listening to pink noise is admittedly different 🙂
I'd agree but the problem isn't coloration, the problem is that we don't have a good understanding how spatial hearing works. Coloration is just one effect next to many.
All I know is that I do. Maybe not optimal setup, but it works well enough for me.
If you set up a dipole with maybe a 30 deg. angle, wide placement, the rear energy is going to be reasonably coming from the side since part of the energy will still go directly to the side wall while the rear axis energy will go first to the far side of the front wall, reflect outward towards the side wall that will then reflect towards the listening space. To me there seems to be some over-simplification of where all of the energy passes.
The energy from the side of the dipole closer to the front wall will also be somewhat reduced due to the dipole pattern at that angle, so the front wall treatment stands a better chance of sufficiently reducing the ratio of direct to reflected energy at the listening position. In practice I just haven't found the magnitude to be the problem suggested.
I also always position the speakers along the long wall as well and prefer a roughly equilateral triangle positioning if I can do that. This to me always seems to be the better position in a room. I started doing that right out of college in my first home before I knew anything about acoustics. That part has never changed for me.
Dave
My wall is acoustically flat so that might explain why the comb is so visible. The backside was firing into a corner as well.
I'm not sure a debate takes place when people are saying the same thing. Well, I guess it still can as evidenced by.... Rigor is great when it can be accomplished. If there was much rigor here, there would be no debate just a body of useful evidence. Accepting facts is hard for any human to do and more so if it contradicts their beliefs. Everyone seems to have a point to prove with a mix of belief and supporting evidence. Anyway, I think it was John K who also posted the little smoother low end of a dipole before. Again not radically so, but there is a trend.
Dan
And me, and thousands (many thousands, if you count Maggie owners) of others whose installations do not suffer from the "sim'd" effect. From which the obvious conclusion is either that they're all deaf (speakerdave's position), or that the sim has little-to-nothing to do with reality. I don't for a moment doubt that a single discrete "echo" can be heard (with varying effect, depending on the amplitude and the delay) . . . but that single discrete echo "simulated" at -5dB simply does not occur in my listening room, or any other that I have heard . . . so I don't regard the "sim" as having any bearing on reality. It's kind of like shutting off one rail of an amplifier's power supply (the "sim"), noting that it sounds terrible (if it works at all), and concluding that if the power supply rails don't track within .001% the amplifier will sound bad. Just not useful . . .
While talking about perception vs. measurements, some interesting examples:
Top 10 Incredible Sound Illusions
"MgGurk Effect": DON'T listen first with your eyes closed. First watch and then listen again with your eyes closed.
Top 10 Incredible Sound Illusions
"MgGurk Effect": DON'T listen first with your eyes closed. First watch and then listen again with your eyes closed.
- Status
- Not open for further replies.