5 ms·

Dave Rat, of Rat Sound/RHCP/Bassnectar/etc fame, has some very interesting takes on the wall of sound idea using modern equipment. The core of it is that speak

Dave Rat, of Rat Sound/RHCP/Bassnectar/etc fame, has some very interesting takes on the wall of sound idea using modern equipment. The core of it is that speakers are bad at polyphony, so if you can avoid it, you can produce something that sounds more natural to human ears, and do so in a larger area. The way to avoid it is more speakers, more stacks/arrays/etc. You don't necessarily need an array per instrument, because modern loudspeaker arrays are indeed much better than they used to be, and modern loudspeaker processing fixes a multitude of problems.

mrob 2y ago
Another advantage of individual speakers for each instrument is that it allows positioning them without the phase cancellation artifacts you get with a stereo setup. It's obvious with some sounds, e.g. try playing some mono pink noise on stereo speakers and move your head. Then try again after hard panning the audio to only one of the speakers. This is why adding a real physical center channel makes dialogue clearer in movies.
- hunter2_ 2y ago
  The center channel also helps by allowing the listener to localize the dialogue differently than the music/FX. When the dialogue comes from a different point in space, it doesn't need to be louder than the music/FX for sufficient intelligibility. When downmixed to stereo or mono or any other non-center-having configuration, the dialog must be mixed well above the music/FX to avoid being masked by it. Sometimes a downmix occurs without this step being given enough consideration, resulting in complaints about the mix, when the original 5.1 (or whatever) mix was perfect but an engineer wasn't involved for downmixing. This is also what Dave Rat is on about, in addition to IM and combing concerns, when explaining why a speaker per instrument sounds better than a mix: not just that it's a single point source, not just that it's doing one single job, but that all the sounds come from discrete points nowhere near each other. Just like an acoustic band using no reinforcement.
amlib 2y ago
I wonder how do they deal with the phasing over so many speakers? Wouldn't an array of speakers require you to do something about that? Specially so if you are stuck with 70s tech.
- JohnBooty 2y ago
  deal with the phasing over so many speakers? Each performer had their own vertical stack of speakers. So, no phase issues / comb filtering on the horizontal axis. https://www.soundonsound.com/techniques/line-arrays-explained https://www.soundonsound.com/techniques/line-arrays-explaine...
  - amlib 2y ago
    I was thinking more in terms of the Wall of Sound as shown in the article. Were they also doing vertical stacks per channel/instrument or were they doing arrays of speakers arranged in a grid per channel/instrument? From what I take it was the later as line arrays weren't a thing at that point, right?
    - ssl-3 2y ago
      With the Wall of Sound, it worked more-or-less like this: The bass guitar gets its own pile of speakers. The lead guitar gets its own pile of speakers. The vocals get their own pile of speakers. (And so on, and so forth, until everything on the stage was covered -- maybe with some mixing on a stack to fit something else in, but probably not for things like Jerry's guitar and Phil's bass). So, yes: There were aspects of the Wall of Sound that absolutely behaved like modern line arrays do (even though "line arrays" as we know and use them today did not begin to gain wide popularity until somewhere around the middle of the 1990s). There's no magic necessary to implement a line array: It can just be an array of speakers that are arranged in a line. (Of course, it can also be much, much more complex than that -- but that complexity isn't an inherent part of what a "line array" be. It can involve things like phase-steering and per-element EQ and 3D predictions [and 3D measurements!], but it does not have to be that way in order to be a line array.)
JohnBooty 2y ago
The core of it is that speakers are bad at polyphony You can look at intermodulation measurements for some popular and affordable home speakers here. The TL;DR is yeah, you're going to go from something like -60dB distortion to -40dB when doing synthetic multitone tests, but I would not remotely characterize this as being "bad" at polyphony nor the primary reason for the Wall of Sounds primordial "one vertical speaker array per instrument" design. https://www.erinsaudiocorner.com/loudspeakers/kef_r5_meta/ https://www.erinsaudiocorner.com/loudspeakers/kef_r5_meta/ https://www.erinsaudiocorner.com/loudspeakers/sony_sscs3_tower/ https://www.erinsaudiocorner.com/loudspeakers/sony_sscs3_tow... more: https://www.erinsaudiocorner.com/loudspeakers/ https://www.erinsaudiocorner.com/loudspeakers/ The way to avoid it is more speakers, more stacks/arrays/etc. If we're talking about his remarks here I'd characterize his take as "more vertical arrays" and not just "more speakers." I realize it's a bit pedantic of me, but some people think that simply adding more speakers equals more betterer sound and it's quite far from the case. In general, multiple loudspeakers arrayed horizontally (this includes MTM center channels in home theater setups) lead to comb filtering. I'm sure you know that, just clarifying for others. If he has written about this elsewhere I'd love to read more!
- jtriangle 2y ago
  I'm simplifying things significantly because I don't expect most HN'ers to know the ins and outs of pro sound. Dave Rat has a good youtube channel, he also has some sort of insider subscription thing that I've never bothered with. I wouldn't say his ideas necessarily translate to every situation, but, he really presents these ideas as tools to use in a toolbox, not as gospel. One interesting thing is that, we generally view comb filtering as universally 'bad', when, in reality, our ears do an excellent job of sorting out comb filtering when it comes to natural sounds. In fact, comb filtering is how we can locate a sound in 3d space with only two reference points (and, if you try, only one reference point moved around a little). That's remarkable, and points back to how speakers comb filter instead of mere comb filtering itself. In practice, say you have a rock band, and your sound system has two arrays with subs spaced 40ft apart. Now, you're going to get a less than ideal pattern from that in the ranges where the bass guitar and kick drum live. How do you fix it? The answer is fairly simple, you simply run bass/kick in stereo, then, you delay the bass on one side by a little, kick on the opposite side just a little, then add some kind of EQ difference to the delayed side of each, then play with the delays until it sounds right. Why does that work? Or does it really work? It's odd, because the math says "no no, it'll sound bad", the reality is, it can tighten up that comb to the point you don't really hear it and you can get good bass coverage out of a fairly poor system design.

mrob 2y ago

Another advantage of individual speakers for each instrument is that it allows positioning them without the phase cancellation artifacts you get with a stereo setup. It's obvious with some sounds, e.g. try playing some mono pink noise on stereo speakers and move your head. Then try again after hard panning the audio to only one of the speakers. This is why adding a real physical center channel makes dialogue clearer in movies.

hunter2_ 2y ago
The center channel also helps by allowing the listener to localize the dialogue differently than the music/FX. When the dialogue comes from a different point in space, it doesn't need to be louder than the music/FX for sufficient intelligibility. When downmixed to stereo or mono or any other non-center-having configuration, the dialog must be mixed well above the music/FX to avoid being masked by it. Sometimes a downmix occurs without this step being given enough consideration, resulting in complaints about the mix, when the original 5.1 (or whatever) mix was perfect but an engineer wasn't involved for downmixing. This is also what Dave Rat is on about, in addition to IM and combing concerns, when explaining why a speaker per instrument sounds better than a mix: not just that it's a single point source, not just that it's doing one single job, but that all the sounds come from discrete points nowhere near each other. Just like an acoustic band using no reinforcement.

amlib 2y ago

I wonder how do they deal with the phasing over so many speakers? Wouldn't an array of speakers require you to do something about that? Specially so if you are stuck with 70s tech.

JohnBooty 2y ago
deal with the phasing over so many speakers? Each performer had their own vertical stack of speakers. So, no phase issues / comb filtering on the horizontal axis. https://www.soundonsound.com/techniques/line-arrays-explained https://www.soundonsound.com/techniques/line-arrays-explaine...
- amlib 2y ago
  I was thinking more in terms of the Wall of Sound as shown in the article. Were they also doing vertical stacks per channel/instrument or were they doing arrays of speakers arranged in a grid per channel/instrument? From what I take it was the later as line arrays weren't a thing at that point, right?
  - ssl-3 2y ago
    With the Wall of Sound, it worked more-or-less like this: The bass guitar gets its own pile of speakers. The lead guitar gets its own pile of speakers. The vocals get their own pile of speakers. (And so on, and so forth, until everything on the stage was covered -- maybe with some mixing on a stack to fit something else in, but probably not for things like Jerry's guitar and Phil's bass). So, yes: There were aspects of the Wall of Sound that absolutely behaved like modern line arrays do (even though "line arrays" as we know and use them today did not begin to gain wide popularity until somewhere around the middle of the 1990s). There's no magic necessary to implement a line array: It can just be an array of speakers that are arranged in a line. (Of course, it can also be much, much more complex than that -- but that complexity isn't an inherent part of what a "line array" be. It can involve things like phase-steering and per-element EQ and 3D predictions [and 3D measurements!], but it does not have to be that way in order to be a line array.)

JohnBooty 2y ago

The core of it is that speakers are bad at polyphony You can look at intermodulation measurements for some popular and affordable home speakers here. The TL;DR is yeah, you're going to go from something like -60dB distortion to -40dB when doing synthetic multitone tests, but I would not remotely characterize this as being "bad" at polyphony nor the primary reason for the Wall of Sounds primordial "one vertical speaker array per instrument" design. https://www.erinsaudiocorner.com/loudspeakers/kef_r5_meta/ https://www.erinsaudiocorner.com/loudspeakers/kef_r5_meta/ https://www.erinsaudiocorner.com/loudspeakers/sony_sscs3_tower/ https://www.erinsaudiocorner.com/loudspeakers/sony_sscs3_tow... more: https://www.erinsaudiocorner.com/loudspeakers/ https://www.erinsaudiocorner.com/loudspeakers/ The way to avoid it is more speakers, more stacks/arrays/etc. If we're talking about his remarks here I'd characterize his take as "more vertical arrays" and not just "more speakers." I realize it's a bit pedantic of me, but some people think that simply adding more speakers equals more betterer sound and it's quite far from the case. In general, multiple loudspeakers arrayed horizontally (this includes MTM center channels in home theater setups) lead to comb filtering. I'm sure you know that, just clarifying for others. If he has written about this elsewhere I'd love to read more!

jtriangle 2y ago
I'm simplifying things significantly because I don't expect most HN'ers to know the ins and outs of pro sound. Dave Rat has a good youtube channel, he also has some sort of insider subscription thing that I've never bothered with. I wouldn't say his ideas necessarily translate to every situation, but, he really presents these ideas as tools to use in a toolbox, not as gospel. One interesting thing is that, we generally view comb filtering as universally 'bad', when, in reality, our ears do an excellent job of sorting out comb filtering when it comes to natural sounds. In fact, comb filtering is how we can locate a sound in 3d space with only two reference points (and, if you try, only one reference point moved around a little). That's remarkable, and points back to how speakers comb filter instead of mere comb filtering itself. In practice, say you have a rock band, and your sound system has two arrays with subs spaced 40ft apart. Now, you're going to get a less than ideal pattern from that in the ranges where the bass guitar and kick drum live. How do you fix it? The answer is fairly simple, you simply run bass/kick in stereo, then, you delay the bass on one side by a little, kick on the opposite side just a little, then add some kind of EQ difference to the delayed side of each, then play with the delays until it sounds right. Why does that work? Or does it really work? It's odd, because the math says "no no, it'll sound bad", the reality is, it can tighten up that comb to the point you don't really hear it and you can get good bass coverage out of a fairly poor system design.