Headliner speakers in R1S only work with 3D audio

[R1Sezejay]

Former audio engineer here who posted some of the EQ recommendations and analysis on this forum. R1S Elevation and R1T Meridian.

The “headliner” speakers are toggled on and off by the 3D modes. That's by design. We definitely want the ability to turn them off. Here's why.

99%+ of music is recorded and played back in stereo, which means it contains two separate channels of information: Left and Right. When a system has only two speakers, mapping which channel of information goes to which speaker is straightforward: Left to left, Right to right.

However, an audio system with more than two speakers (e.g., a typical car system) has to make decisions about how to split/copy the Left and Right channels of information to other speakers. This process is called "upmixing."

The two biggest factors that influence upmixing decisions are:

1. The bandwidth capability of the other speakers, or the frequency range they can play at what volume before distorting.
2. What the listener should hear when the other speakers play audio back.

Let's break these two factors down with the most basic example of an additional speaker beyond the left and right: the subwoofer.

First, a system with a subwoofer has to decide what information to send to the subwoofer to play back. Generally, it splits off content below 80 Hz from both the Left and Right channels via a crossover and sends the sum to the subwoofer because the subwoofer is much more efficient at playing back low frequencies at higher volumes than your typical left and right speakers. Efficiency has to do with how much air a speaker can move (loudness) at what frequency before it breaks up. Subwoofers move air at lower frequencies more efficiently than tweeters, for example.

Second, we only ask the subwoofer to play frequencies below 80 Hz (even if it can efficiently play back frequencies up to 150-200 Hz) because those frequencies are hard to localize. Our ears and brains don't/can't use differences in volume and phase to determine where low frequencies are coming from in space. If we asked a subwoofer to play back the frequency ranges of a violin or voice (assuming it had the bandwidth capability to do that), we'd notice immediately they were coming from the trunk or under the seat. This is probably not what people want to hear.

All of the above applies to the headliner speakers (or the center speaker, for that matter), but the problem is actually more complicated because all of the frequencies we would send upward are both highly localizable and we can't use a simple crossover to decide what goes into those speakers and what doesn't.

If, for example, we use a crossover to send any frequency above 10 kHz to the headliners, a single cymbal hit that spans 400 Hz to 12 kHz would suddenly sound like multiple cymbals coming from different directions. Worse, it would sound confusing because part of the information—mainly the split "hits" occurring at the same time—tells our brains that they're the same cymbal hit but another part of the information—frequency range—tells our brains they have to be two different cymbals. Our brains get confused and spend energy trying to make sense of what is going on. It's actually quite amazing how much of our world we understand only through our ears...

So if we can't use a simple crossover to decide what content to send upward, then we use Digital Signal Processing, i..e, "3D Mode," to make "smarter" decisions. Engineers design algorithms to selectively choose what parts of the content to cut/copy and send upward based principally on two things:

1. Frequency range.
2. Volume similarity and phase coherence between the Left and Right channels.

Like we said earlier, it doesn't make sense to send low frequencies upward because those speakers can't handle them (they're small), so the DSP sends mid and high frequencies.

Here's an oversimplification of how it decides which "occurrences" of those frequencies to send upward: It uses Fast Fourier Transforms to break down content with complex waves (pretty much all music unless you like listening to pure sine waves) into constituent sine waves. Then it determines how similar the Left and the Right waves are to each other based on volume and phase. The most basic example is a voice recorded on a mono channel microphone and mixed equally between the Left and the Right channels so that the voice sounds like it's coming from dead center even if no speaker exists in the center. When our brain hears sound in the left and right ears that are the same volume and same phase (the constituent frequencies arrive at both ears at the same time), it decides the sound came from either dead in front or behind us (how it determines which is another topic entirely).

When the DSP compares the Left and Right channels to each other and sees that certain content has the same volume in both and the same phase in both (all the waves line up in time), it assumes that it's probably something like a voice or snare or kick drum that the mixing engineer probably doesn't want you to hear from anywhere but dead ahead and that you probably don't want to hear from dead ahead.

If the content differs in volume and/or phase—like a high-hat that's been panned right or room reverb—the DSP might decide to send that upward because it assumes those sounds contribute to us hearing "space" and we don't want to localize it anyway.

It does all this in real time.

The challenge any upmixer like 3D Enhance has is that it can only guess. It has no idea how to differentiate between a snare drum and a voice and a plucked nylon guitar or what any of those things should sound like if they came from overhead or if they were reflecting off a high ceiling and providing a larger sense of spaciousness. It's a lot like trying to convert a 2D photo into 3D so that you can move your head around and see what's behind the person in the foreground: software can do it, but it has to guess and probably won't look convincingly real.

Now to the key question: Should you turn 3D modes on or leave them off?

Answer: If you like them, turn them on.

Hopefully, all this explanation helps you understand what is going on behind the scenes so you can make a more "informed" decision about what you like and dislike. =)

Thank you for sharing your talent! Really appreciate it!

Sponsored

 

[superfluid]

First, thank you for the detailed explanation and the helpful information regarding the equalizer settings.

Since you’ve probably forgotten more about audio systems than I’ve ever known, I have a question which I don’t recall reading about. If I missed this somewhere else please accept my apology and feel free to redirect me.

My question is: Does the front center speaker truly improve the stereo experience or does it level out the left/right volume at the cost of true stereo separation? My (possibly wrong) understanding is that it is supposed to give each front seat occupant a more level left/right balance by filling in the middle with a mix so you can set the balance at the center and not be overwhelmed by the volume of the side dash speaker which is closer to you. In my simplistic, and likely completely incorrect, view, this is essentially a mix of the left+right.

While I am not a true audiophile, when I listen to stereo music on a surround-capable 5.1 receiver at home I prefer the sound when it is set to simply stereo: left/right/subwoofer on with center & rears off. There doesn’t seem to be a way to do this on the car audio systems I’ve had which include a front center speaker.

Thank you again.

This is a deep question that definitely pushes up against my knowledge limits because car audio is quite different from home audio (which differs still from headphone, cinema, and live audio). The main difference that applies to your question about stereo separation is proximity and relative position of the listener to the speakers.

In car audio, everyone is really close to some speakers, but no one is equidistant from any of them. That makes creating a sound field everyone hears similarly very difficult because so much of how we determine the position of a sound source is by hearing time similarities and differences between speakers. Optimizing the sound field for one person overcorrects it for everyone else. I could be wrong, but I think that's why we never really talk about "imaging" or "sound stage" when it comes to cars; giving one person an excellent image would probably make the music hard to listen to for others.

With that, let's jump into stereo imaging, the center channel, and your question.

In home audio, the center channel is designed to anchor sound to the screen. This is particularly important for dialog. Dialog needs to sound like it's coming from where the speaking character is—which is almost always on the screen. If we hear the dialog come from somewhere other than where the actor is, we get confused and lose our suspension of disbelief. With a center channel, dialog mixed to the center channel will sound like it's coming from the person talking on screen regardless of where you sit. Without a center channel, the dialog sounds like it comes from whichever speaker you're sitting closest to. This is one of the reasons home theater enthusiasts often insist on using a projection screen instead of a light-emitting display like an OLED: projection lets you put the center channel behind an acoustically transparent screen and make dialog sound like it's literally coming out of the actor's mouth. This adds a lot to realism.

The center channel's importance also applies to sounds that pan from one side to the other. If a train speeds from left to right on the screen, you want to hear the train move smoothly from the left to the right speaker as the train moves across the screen. The further the left and right speakers are from each other, the harder that is to do without a channel in the middle, especially if people are sitting off-center.

The reason you tend to prefer turning off the center channel when listening to music is a bit complex but mostly comes down to this:

As mentioned earlier, 99% of music is mixed in stereo. The mixing engineer made all their choices on a stereo set of speakers, including how loud each instrument is relative to the others, where to place the instruments in the stereo sound field, and what kind of space to present those sounds in (i.e., a large reverberant space a la Enya or small dry space like a recording studio a la Billie Eilish). That music is published with two channels of information. When we upmix it to include a center channel (or more channels), the upmixer has to guess what to put in the center in a very similar way as I described earlier because the music has no center channel information. Some do a decent-enough job (I prefer Auro's upmixers), but none do a perfect one.

It won't sound perfectly coherent because it isn't perfectly coherent. Anything mixed down the middle, like a vocal, is easy to decode and put in the center channel in a passably convincing way. But Dave Matthew's guitar in his live recording of Crush at Radio City, which is panned about 10 degrees left of center, is much harder to figure out. In the stereo recording, some of the information is in the left channel and some (less) in the right. How much of that guitar do you take from the left and right and put through the center so that it still sounds like it's panned 10 degrees to the left? Not easy to figure out in real time and render convincingly.

We are actually very good at hearing when something is not correct, just like we're good at seeing when a special effect in a movie doesn't look right. The difference between you and most other people is that you care and they don't =)

And finally to your question.

I think you're right: I don't believe the center channel in a car is supposed to improve the "stereo" experience because stereo imaging isn't a car audio engineer's aim. Instead, it's supposed to "fill in the hole" like you said, but in a somewhat more specific way.

A stereo image is created between two speakers at a specific point. The further outside of that point a listener is, the less of a coherent image they will hear. The problem in a car is that every listener is really close to one speaker and really far from every other. The driver is effectively listening to the left speaker by itself and the passenger is listening to the right speaker by itself. Neither can experience a stereo image.

I think the center is designed to provide the driver and passenger a closer speaker to immerse them more than they would otherwise be. Effectively, the driver is listening to the left and center speakers while the passenger is listening to the center and right speakers. They both get sound hitting both of their ears at similar levels because that center speaker is "filling in the hole." (Another clue that this is true is that center speakers in a car are often pointing up at the windshield, which probably bounces the sound into the cabin more widely and diffusely like an Atmos speaker does of a home's ceiling.) But they don't get a stereo image because that center channel can't act as a true right channel for the driver or a true left channel for the passenger at the same time because it would have to play two different channels of audio at the same time, which would be hella confusing.

Instead of a stereo image, I think car audio engineers aim to create a sound field where the music sounds like it's coming from everywhere and nowhere at once and none of the speakers grab individual attention. The aim of car audio seems to be making things sound clear but not making their positions clear. A center channel in a car can help with that.

Hope that answers your question.

 

[SoCal Rob]

This is a deep question that definitely pushes up against my knowledge limits because car audio is quite different from home audio (which differs still from headphone, cinema, and live audio). The main difference that applies to your question about stereo separation is proximity and relative position of the listener to the speakers.

In car audio, everyone is really close to some speakers, but no one is equidistant from any of them. That makes creating a sound field everyone hears similarly very difficult because so much of how we determine the position of a sound source is by hearing time similarities and differences between speakers. Optimizing the sound field for one person overcorrects it for everyone else. I could be wrong, but I think that's why we never really talk about "imaging" or "sound stage" when it comes to cars; giving one person an excellent image would probably make the music hard to listen to for others.

With that, let's jump into stereo imaging, the center channel, and your question.

In home audio, the center channel is designed to anchor sound to the screen. This is particularly important for dialog. Dialog needs to sound like it's coming from where the speaking character is—which is almost always on the screen. If we hear the dialog come from somewhere other than where the actor is, we get confused and lose our suspension of disbelief. With a center channel, dialog mixed to the center channel will sound like it's coming from the person talking on screen regardless of where you sit. Without a center channel, the dialog sounds like it comes from whichever speaker you're sitting closest to. This is one of the reasons home theater enthusiasts often insist on using a projection screen instead of a light-emitting display like an OLED: projection lets you put the center channel behind an acoustically transparent screen and make dialog sound like it's literally coming out of the actor's mouth. This adds a lot to realism.

The center channel's importance also applies to sounds that pan from one side to the other. If a train speeds from left to right on the screen, you want to hear the train move smoothly from the left to the right speaker as the train moves across the screen. The further the left and right speakers are from each other, the harder that is to do without a channel in the middle, especially if people are sitting off-center.

The reason you tend to prefer turning off the center channel when listening to music is a bit complex but mostly comes down to this:

As mentioned earlier, 99% of music is mixed in stereo. The mixing engineer made all their choices on a stereo set of speakers, including how loud each instrument is relative to the others, where to place the instruments in the stereo sound field, and what kind of space to present those sounds in (i.e., a large reverberant space a la Enya or small dry space like a recording studio a la Billie Eilish). That music is published with two channels of information. When we upmix it to include a center channel (or more channels), the upmixer has to guess what to put in the center in a very similar way as I described earlier because the music has no center channel information. Some do a decent-enough job (I prefer Auro's upmixers), but none do a perfect one.

It won't sound perfectly coherent because it isn't perfectly coherent. Anything mixed down the middle, like a vocal, is easy to decode and put in the center channel in a passably convincing way. But Dave Matthew's guitar in his live recording of Crush at Radio City, which is panned about 10 degrees left of center, is much harder to figure out. In the stereo recording, some of the information is in the left channel and some (less) in the right. How much of that guitar do you take from the left and right and put through the center so that it still sounds like it's panned 10 degrees to the left? Not easy to figure out in real time and render convincingly.

We are actually very good at hearing when something is not correct, just like we're good at seeing when a special effect in a movie doesn't look right. The difference between you and most other people is that you care and they don't =)

And finally to your question.

I think you're right: I don't believe the center channel in a car is supposed to improve the "stereo" experience because stereo imaging isn't a car audio engineer's aim. Instead, it's supposed to "fill in the hole" like you said, but in a somewhat more specific way.

A stereo image is created between two speakers at a specific point. The further outside of that point a listener is, the less of a coherent image they will hear. The problem in a car is that every listener is really close to one speaker and really far from every other. The driver is effectively listening to the left speaker by itself and the passenger is listening to the right speaker by itself. Neither can experience a stereo image.

I think the center is designed to provide the driver and passenger a closer speaker to immerse them more than they would otherwise be. Effectively, the driver is listening to the left and center speakers while the passenger is listening to the center and right speakers. They both get sound hitting both of their ears at similar levels because that center speaker is "filling in the hole." (Another clue that this is true is that center speakers in a car are often pointing up at the windshield, which probably bounces the sound into the cabin more widely and diffusely like an Atmos speaker does of a home's ceiling.) But they don't get a stereo image because that center channel can't act as a true right channel for the driver or a true left channel for the passenger at the same time because it would have to play two different channels of audio at the same time, which would be hella confusing.

Instead of a stereo image, I think car audio engineers aim to create a sound field where the music sounds like it's coming from everywhere and nowhere at once and none of the speakers grab individual attention. The aim of car audio seems to be making things sound clear but not making their positions clear. A center channel in a car can help with that.

Hope that answers your question.

Thank you very much for helping me, and possibly others, with this information!

I was wondering if the center speaker was worth the vibrating dashboard on rough roads given the speaker’s weight attached to the poorly-supported speaker grill assembly. It seems like it is actually improving the overall experience.

You saved me from going down the path of disassembling the dash, disconnecting & removing the speaker, and checking to see if disconnecting it caused any problems/errors with the audio system. If it did, then I’d be trying to find out what could replace the speaker electrically: Inductor? Resistor?

Thanks for letting me leave things as they are without feeling that I’m just being lazy. It’s for better audio.

 

[Rizzian]

Another former audio engineer here... I'm only on day 6 with my 2024 R1S and haven't spent enough time with it to put together as thoughtful of a post as some have done here, but I will agree that the sound system is somewhat disappointing.
Prior to purchase, I had read something indicating that the new sound system had 18 speakers and 1100 watts of power, which had me excited coming from a stock Toyota 4Runner that was about as average as car audio gets. My first impression was that the Rivian seemed capable of hitting all the requisite frequencies, but that the system lacked "tuning" as a whole. The speakers and power seem adequate, but perhaps the acoustics of the truck just stink. Subs are a bit lacking, especially considering how much more power this system has compared to what I had been listening to. The first thing I did was turn off the 3-D/ fake surround nonsense and try to fiddle with the EQ setting. At the end of the day, I liked the curves posted by the other audio engineer who used measurement equipment to try and flatten things out as best as he could. There's definitely something wonky about the mid-range, mono audio information that just isn't translating correctly in the R1S. Snares are bright, but missing punch and lead vocals (which typically are panned right up the middle) are subdued and even obscured by the music. The first thing my wife said about the sound system was that "I can't hear the singer...all the vocals sound quiet." The EQ did help with this significantly. If people hadn't said that the original Meridian system sounded so good, I'd probably be blaming it all on the glass roof. Since it's a lease, I probably won't be installing any upgrades. Overall, it sounds better than my stock 2016 4Runner, but I was expecting much more.

 

[superfluid]

Thank you @superfluid I went back and read your whole super thread from Feb. one question: has anyone posted recommended EQ settings for the R1S Meridian system? Of course we have the super thread for R1S Elevation system but I haven’t seen anyone do a scientific type measurement on the R1S Meridian. Thanks!

Not to my knowledge. I plan to measure an R1S Meridian at some point but haven’t made time for it yet. But it will happen because I want to test the hypothesis that the R1S Elevation system is flawed at an engineering level and try to get Rivian to fix it.

 

[UnsungZero_OldTimeAdMan]

Adding to discussion about L/R and F/R bias and listening positions. If you were to tune the system only to you, the driver, and no one else, you want the fader to be slightly right side biased... since you're closer to the left speaker. But in doing so, your front passenger is not going to get balanced sound. As for F/R bias, I've always been taught to set up rear speakers as fill—meaning you bias rear just enough to hear them. With my Rivian audio, I have the fader set centered, but slightly ear of front row.

 

[Glembi2]

There is much talk about the crossovers in the R1S being off. Are crossovers adjustable via over the air updates? Or just fixed? Or only adjustable on the hardware.

An older sub in my home theater had an adjustable crossover but that was on the speaker, not on the amp or pre-amp. I think that means it had an adjustable input filter. There was no separate control of which frequencies went where

 

[superfluid]

Another former audio engineer here... I'm only on day 6 with my 2024 R1S and haven't spent enough time with it to put together as thoughtful of a post as some have done here, but I will agree that the sound system is somewhat disappointing.

The first thing I did was turn off the 3-D/ fake surround nonsense and try to fiddle with the EQ setting. At the end of the day, I liked the curves posted by the other audio engineer who used measurement equipment to try and flatten things out as best as he could. There's definitely something wonky about the mid-range, mono audio information that just isn't translating correctly in the R1S. Snares are bright, but missing punch and lead vocals (which typically are panned right up the middle) are subdued and even obscured by the music. The first thing my wife said about the sound system was that "I can't hear the singer...all the vocals sound quiet." The EQ did help with this significantly.

Nice to hear from a fellow engineer. Glad the EQ settings helped out. I hope to measure an R1S Meridian s00n to show Rivian they messed up the crossover at 1 kHz on the Elevation system. There's just no other good explanation for why there's such a massive peak and dip on each side of that band.

 

[superfluid]

There is much talk about the crossovers in the R1S being off. Are crossovers adjustable via over the air updates? Or just fixed? Or only adjustable on the hardware.

An older sub in my home theater had an adjustable crossover but that was on the speaker, not on the amp or pre-amp. I think that means it had an adjustable input filter. There was no separate control of which frequencies went where

Not sure the Rivian crossover is adjustable OTA, but I doubt it.

Home theater subwoofers have adjustable crossovers, but most people don't/shouldn't use them because the receiver already crosses over the main L/R speakers to the sub for you. If you use the sub's crossover on top of the receiver's crossover, you stack crossovers and cause major frequency response issues. You would only use a subwoofer's built-in crossover if you're running the sub and mains in series—but no one should do that in a home theater situation because then you can't effectively use the receiver's bass management system (e.g., Dirac Live, Audyssey, ARC, etc); the system will get majorly confused about what it hears.

For the Rivian's crossover to be adjustable OTA, it has to be a digital crossover, which means the signal is split and crossed over in the computer before it gets amplified. That's what typical receivers/stereo processors do when they cross the sub to the mains at, say, 80 Hz. The receiver splits the signal to the mains and sub, then sends an amplified signal to each main speaker and a line-level signal to the sub (subs generally have their own amplifiers). In a 5.1 setup, the receiver sends 5 amplified signals so it only needs 5 channels of amplification.

The problem I measured at 1 kHz occurs within the frequency range of the woofer/tweeter—and happens to be right where typical woofer/tweeter crossovers tends to operate. To have an adjustable crossover at 1 kHz, the woofer/tweeter can't be crossed over with a passive network but a digital one. If it has a digital crossover, then each woofer and tweeter would need its own separate channel of amplification.

If the Rivian's 18 speakers each comprise a woofer and a tweeter (they might not; the overheads might just be tweeters), the system would need 36 amplifiers, double the digital crossover processing, and a lot more engineering.

That kind of system is called a fully active system. They're very expensive and usually reserved for the truly high-end.

The cheapest home theater processor I know of that works with active speakers is the Trinnov Altitude 16, starting at $18,500.

Sponsored