Audiophiles: Touring factory EQ curve doesn't change with volume!

DANANDME said:

ti have a 1.5 T CIVIC HATCHBACK made in englend - SPORT model - do you think it will measure the same ?

what kind of equipment you used to measure it ?

Click to expand...

I think it will. Another member here measured his EX's headunit, which is one that has the amplifier built into the headunit rather than external like mine, and he found the same results. So it seems all Civics sold in the US probably do this, and I can't imagine why they'd do differently in other markets if they have the same speakers.

The equipment I used is no longer made, but it's basically just an accurate soundcard. Soundcards have a much lower maximum voltage they can accept as input than the stereo sends to the speakers, but you can get around this by either creating a voltage divider circuit with some resistors, or creating pink noise test tracks that are at a very low volume level.

For just checking if the bass level changes relative to the rest at different headunit volume settings, you can also just use an accurate microphone with an RTA program to measure the speaker's output. That's what the guy who checked his EX's headunit did.

josby said:

I think it will. Another member here measured his EX's headunit, which is one that has the amplifier built into the headunit rather than external like mine, and he found the same results. So it seems all Civics sold in the US probably do this, and I can't imagine why they'd do differently in other markets if they have the same speakers.

The equipment I used is no longer made, but it's basically just an accurate soundcard. Soundcards have a much lower maximum voltage they can accept as input than the stereo sends to the speakers, but you can get around this by either creating a voltage divider circuit with some resistors, or creating pink noise test tracks that are at a very low volume level.

For just checking if the bass level changes relative to the rest at different headunit volume settings, you can also just use an accurate microphone with an RTA program to measure the speaker's output. That's what the guy who checked his EX's headunit did.

Click to expand...

I see an JL audio device ...
What is it?

DANANDME said:

I see an JL audio device ...
What is it?

Click to expand...

Oh lol no actually that is a M-Audio MobilePre. That's the USB soundcard I used.

Hey Josby,

I see you stopped your build log over at DIYMA but you're still active here. I was curious to see if you found a summing solution (or other) for the center channel information on our cars.

I keep getting lost following the signal path that leads to the center.
The log says you made the adapter to send the 12 signals (front mids and tweets, center, and sub) but that it didn't work for your ms-8. You continued to use a center and found that the voice/nav sound was missing, so I am assuming that the ms-8 "created" the center info you sent to your center speaker from the front (?) info. This must have been musically better, but was it that much better that it was worth losing the voice/nav?

You did not go into detail on why adding the factory center info didn't work for you, so I'm curious as to why you decided not to add it as it only created more obstacles for you to plan around. I'm assuming, again, the factory center info was a mix of info of the FL and FR channels along with the voice/nav sounds that the ms-8 didn't "like"?
Do you think the factory center signal sent right back to the center (not to a dsp) would interfere with the other DSP'd speakers? I have the 12 channel system in the hatch, so same question relating to the rear fill speakers..

And it looks like you fed the rear speakers power from the ms-8, so did you loose your factory fader controls? I know you had the controller mounted up front, but I would like to retain as much factory control as possible. In order to do this however, it seems like you have to have two processors with summing, or one with a TON of inputs, to send clean signals to the amps.


I'm trying to plan my build and using yours as a guide and I like how you added your own plugs and kept it as stock/reversible as possible. I also am wondering if using the 18/9 wire will support around 50-65 watts to send back to the factory connection to then send to the doors.
But I think I would send it back as a 4 full summed signals back to the doors midbass signals to go to component crossovers and then split from there to the door and pods. It seems like a hassle to sum it only to split it back up, but you wind up with less required amplified channels, right?
yes, i'm losing my marbles trying to come up with the best solution, but it reminds me of the old days of home theater when component cables came out and there were adapters and cables going everywhere!

Anyway, I can't find a single processor that can handle all this split up channel crap that these new vehicles have. They either need to come out with an upgraded DSP or allow us to keep the factory connectivity with an aftermarket

Sorry for the ramble of questions, but i'm working through this one step at a time.
I just upgraded from a 2009 Honda Fit Sport which was much simpler to upgrade as it had no bells and whistles.

Wow, lots of questions here, but I will try my best to answer them.

MrStix said:

I was curious to see if you found a summing solution (or other) for the center channel information on our cars.

Click to expand...

I have since found a solution that I think would work, but haven't had a desire to buy the gear I'd need to try it.

MrStix said:

The log says you made the adapter to send the 12 signals (front mids and tweets, center, and sub) but that it didn't work for your ms-8. You continued to use a center and found that the voice/nav sound was missing, so I am assuming that the ms-8 "created" the center info you sent to your center speaker from the front (?) info.

Click to expand...

Yes, to my knowledge there is no DSP that will integrate a factory center channel's signal in with the rest, because they can't know what algorithm the car used to create it. So they don't know how to downmix it back to stereo. And you're correct, the MS-8 upmixes its own center channel output based on what it finds in the L and R inputs.

MrStix said:

This must have been musically better, but was it that much better that it was worth losing the voice/nav?

Click to expand...

Yes, I find I don't really miss them. I always found the nav voice annoying because it interrupts too often, so I kept it turned off anyway. And the only handsfree command I use is music search, and I already know what each voice prompt is asking me from experience, so I can still use that even though I can no longer hear its voice prompts. So I'm fine with it.

MrStix said:

You did not go into detail on why adding the factory center info didn't work for you, so I'm curious as to why you decided not to add it as it only created more obstacles for you to plan around. I'm assuming, again, the factory center info was a mix of info of the FL and FR channels along with the voice/nav sounds that the ms-8 didn't "like"?

Click to expand...

Yes, the MS-8 expects to get L and R signals on its inputs. The only exception is a subwoofer signal, which it expects to be mono. So if it sees a mono input, it thinks it's a sub signal and applies a 200Hz low-pass filter to it. So it did see my center channel signal coming in, but thought it was a subwoofer due to it being mono, and therefor only "used" the bass from it. So I'd just hear a faint sound of the lowest frequencies of the nav/HFL voices.

MrStix said:

Do you think the factory center signal sent right back to the center (not to a dsp) would interfere with the other DSP'd speakers? I have the 12 channel system in the hatch, so same question relating to the rear fill speakers..

Click to expand...

Yes, because it has music on it. So you'll have the rest of your speakers playing music that sounds good, while the center speaker is still playing music that sounds like crap. Plus there's some delay as the signal is processed in a DSP, so the rest of the speakers would be playing slightly behind in time from the center speaker.

If you wanted to do this, I'd look for a DSP that will accept a center channel and let you apply DSP to it as a discrete channel. That way you could tailor the sound playing from the center to match the new speaker you'd put there. I'm not sure such a DSP exists, though. You'd want to have DTS Neural turned on in the headunit if you did this.

MrStix said:

And it looks like you fed the rear speakers power from the ms-8, so did you loose your factory fader controls? I know you had the controller mounted up front, but I would like to retain as much factory control as possible. In order to do this however, it seems like you have to have two processors with summing, or one with a TON of inputs, to send clean signals to the amps.

Click to expand...

Yes, the factory fader control is lost. You could use something like a JL Audio FiX-86 to retain the factory fader, though. You still lose the center, though.

MrStix said:

I'm trying to plan my build and using yours as a guide and I like how you added your own plugs and kept it as stock/reversible as possible. I also am wondering if using the 18/9 wire will support around 50-65 watts to send back to the factory connection to then send to the doors.

Click to expand...

Yes, I'm now running 100W to all my speakers using two JBL MS-A1004 amps, and it all runs through the 18 gauge 9-wire cables and then into the factory wiring. The wires don't get warm.

The only issue I have is "cross-talk", where the input signal is inducing a signal on adjacent output wires inside one of the 9-wire bundles. What happens is, even if I turn my amps off completely, if I turn the headunit up to 36 or higher, I can hear music coming from my speakers VERY quietly. Like, so quiet that I can't hear it if the engine is running. But still, if I was doing this over, I'd make sure the wires carrying the signal back to the DSP were not in the same bundle as ones running up to the speakers.

MrStix said:

Anyway, I can't find a single processor that can handle all this split up channel crap that these new vehicles have. They either need to come out with an upgraded DSP or allow us to keep the factory connectivity with an aftermarket

Click to expand...

I think the DSP manufacturers don't want to tackle factory center channels because there are so many variables.

Now, for that potential solution I mentioned.

There are two pins in the connector that runs into the factory amp, C5 and C12 on the 16-pin connector. The factory service manual describes them as "Inputs sound signal for voice guidance and Voice Recognition (VR) prompts". I've tested and confirmed that these contain an analog signal of the just the voice prompts - no music.

So, if you don't care about retaining music on the center speaker, and you want to retain factory fading, then my recommendation would be to get a FiX-86 and connect the signals for the front door mids and the rear door mids to it. Inputs from the tweeter and sub aren't needed. Then run the FiX's outputs to whatever amps and speakers you put in. Meanwhile, buy a small cheap low-power amp and run the signals from C5 and C12 into one of that amp's inputs. Then hook that amp's output to the center channel speaker. You'll get voice prompts on the center and music from the rest of the speakers.

josby said:

Wow, lots of questions here, but I will try my best to answer them.


Now, for that potential solution I mentioned.

There are two pins in the connector that runs into the factory amp, C5 and C12 on the 16-pin connector. The factory service manual describes them as "Inputs sound signal for voice guidance and Voice Recognition (VR) prompts". I've tested and confirmed that these contain an analog signal of the just the voice prompts - no music.

So, if you don't care about retaining music on the center speaker, and you want to retain factory fading, then my recommendation would be to get a FiX-86 and connect the signals for the front door mids and the rear door mids to it. Inputs from the tweeter and sub aren't needed. Then run the FiX's outputs to whatever amps and speakers you put in. Meanwhile, buy a small cheap low-power amp and run the signals from C5 and C12 into one of that amp's inputs. Then hook that amp's output to the center channel speaker. You'll get voice prompts on the center and music from the rest of the speakers.

Click to expand...

@josby
First of all, thanks for replying.
Yes, I apologize for all the questions at once, but figured I'd ask as much as I could in case I grabbed your attention. I appreciate you taking the time to help me understand it better. And I'm sure it took time to dissect my post.

I remember you mentioned the analog voice signals in your build log, as well as a potential amp for them.

So if I had a 6 channel amp, I could use 4 of the channels for the 4 component sets (1 per door), 1 for the center, and then split the last one to the two ambient speakers somehow? Or just leave those disconnected completely so they no longer get any signal. I know you don't have them in your vehicle, but I doubt they do much of anything useful. I have not tested them yet though.

Also, in your log you mentioned that you got a better signal when sending the mids, tweets, and sub signal to the ms-8. Does it sum all 3?!?
I think I remember you saying that just sending a mid signal was your initial goal due to recommendations on the ms-8, but it cut off the lower and upper spectrum when you tried to do that.

I never understood what summing was til I started look at DSP's for this vehicle. I see recommendations for the bitone, ms-8, helix, and others, but the only ones I will likely be pursuing are the AudioControl DM-810 (or similar) or D-6.1200 or the Alpine PXE-0850S; either in combination with some Kicker KX amps I have. Those are high on my priority list simply because of the discount I get on them. I am only supplying that info for reference as I am taking the recommendations you are giving me and trying to see how I can apply them to those products.

Also, I apologize for jacking this thread, but I'm hoping the info is still relevant enough. I could also DM you, but this may be good info for others like me somewhere down the line.

Thanks again for letting me spitball this stuff.

josby said:

I also checked the center channel (purple) out of curiosity just to see how much bass and treble they allow to go to that little speaker:

Click to expand...

Hi Josby,

I have the 2019 Civic Type R and am working with that infotainment setup.
Thank you so much for going through the trouble of taking these measurements.
My topic is the Center Speaker. Right now my main problem is in the factory setup there is little to no "Center" or mono image audible in the soundstage. I hear either the right or the left, there's no anchor for the vocals, etc. So I took out the factory 3.5" and put in the a Kicker 3.5" and ended up cutting out the kicker tweeter, due to it's being shrill. Running the system in the "Neural" DSP setting with the Kicker does increase the anchoring of the soundstage in the center speaker, however, the sensitivity quoted by the manufacturer of the Kicker 3.5 is 88 dB. The Kicker is more audible than the factory center, but still can't keep up with the pillar tweeters and door speakers. I've just ordered a Infinity REF-3032CFX 3.5" with a sensitivity of 91 dB. Hopefully this is legit at 1 Watt 1 Meter. If so, then the output of the Infinity should be noticeably louder than the Kicker was, and give me some headroom. The Infinity driver is specced at 3 Ohms so hopefully it will be loud enough for me to have to put a dropping resistor in series to lower the level. I'm trying to create a believable listenable presentation of the music in the dash. Right now I can get the frequencies approximately correct but sound stage is shit. I also have replaced the shrill aluminum factory tweets with Polk Dubs (they fit well and are silk dome, which aren't as 2-3.5K one note as the shrieking factory aluminums).
Way to go Honda.

MrStix said:

So if I had a 6 channel amp, I could use 4 of the channels for the 4 component sets (1 per door), 1 for the center, and then split the last one to the two ambient speakers somehow? Or just leave those disconnected completely so they no longer get any signal. I know you don't have them in your vehicle, but I doubt they do much of anything useful. I have not tested them yet though.

Click to expand...

I'm not sure about those rear ambient speakers that the hatchbacks have. I think you're probably right that they don't add a whole lot. Maybe try a preliminary test of just disconnecting those two while leaving everything else the same and see if you can tell a difference?

But yes, your plan with the 6 channel amp sounds good to me as far as powering the center speaker using one of the channels and using the signal from those voice-only pins in the OEM amp connector. I would guess that it's a low-level signal, so I think you'd want to connect the wires you run from it to an RCA connector and plug that into your amp (i.e. rather than using the amp's line-level inputs for it).

MrStix said:

Also, in your log you mentioned that you got a better signal when sending the mids, tweets, and sub signal to the ms-8. Does it sum all 3?!?
I think I remember you saying that just sending a mid signal was your initial goal due to recommendations on the ms-8, but it cut off the lower and upper spectrum when you tried to do that.

Click to expand...

Yeah, I honestly can't remember it all now, but I did think that at the time. I'm not sure if I was calibrating the MS-8's input at a different headunit volume or what, but I'm currently using just the mids for source signal. I can't remember what prompted me to try it again with just the mids, but I know I measured the output response after input calibration again recently and was satisfied with the flatness of it. And I don't feel like it lacks bass or treble.

MrStix said:

the only ones I will likely be pursuing are the AudioControl DM-810 (or similar) or D-6.1200 or the Alpine PXE-0850S;

Click to expand...

I've seen a video on the DM-810 and so I know it has a visualizer that will let you see the input signal it's getting, so you could check to make sure you're getting a full-range signal from the mids with that. I just glanced through the manual for that Alpine unit, though, and it seems to not have any input correction features at all, surprisingly.

MrStix said:

Also, I apologize for jacking this thread, but I'm hoping the info is still relevant enough. I could also DM you, but this may be good info for others like me somewhere down the line.

Click to expand...

No problem, and I agree, you never know when someone else will run across posts later when they're looking for the same info.

Fountainhead said:

My topic is the Center Speaker. Right now my main problem is in the factory setup there is little to no "Center" or mono image audible in the soundstage. I hear either the right or the left, there's no anchor for the vocals, etc. So I took out the factory 3.5" and put in the a Kicker 3.5" and ended up cutting out the kicker tweeter, due to it's being shrill. Running the system in the "Neural" DSP setting with the Kicker does increase the anchoring of the soundstage in the center speaker, however, the sensitivity quoted by the manufacturer of the Kicker 3.5 is 88 dB. The Kicker is more audible than the factory center, but still can't keep up with the pillar tweeters and door speakers.

Click to expand...

I did some further measurements and analysis of the center signal (not posted here) to try to figure out exactly what they're doing with it.

With DTS Neural turned off, the center channel is a simple sum of L+R. So, as I'm sure you experienced, the center image is no good with DTS Neural off.

With DTS Neural on, a true upmixing is done, where the center speaker only plays sounds that are present in both the left and right signal. Furthermore, it reduces those sounds in the L and R channels by 10dB.

So, for example take a singer that is supposed to be at the center of the soundstage. In the original stereo recording, his voice will be at equal levels in both the left and right channels. With DTS off, it will play at equal level from the left and right speaker, but also play from the center speaker at +6dB (from summing two identical sounds). With DTS turned on, his voice will play from the left and right speaker at -10dB from what it originally was, and play from the center speaker at the volume it originally was.

Other upmixers, like Harman's Logic 7, do the center similarly to DTS, so this is the generally accepted way to do things. So, it should produce a good center image, and I'm not sure why it doesn't.

I think you're right that the center needs to play louder, so I think your idea of putting a more sensitive speaker into the center has merit.

But I think another part of the issue may be the range of frequencies that are sent to the center. I think the most noticeable thing about hearing a proper center channel is the way the voice of the singer sounds like it's in the center. According to Wikipedia, the fundamental frequency of the human voice is about 85Hz to 180Hz for men and 165Hz to 255Hz for women. And I've read elsewhere that one's center channel should ideally be able to play down to about 100Hz. But as you saw in the graph you quoted, the Honda center signal rolls off around 250Hz. So those fundamental frequencies of vocals won't be coming out of the center speaker no matter what speaker it is. I think that limits the effectiveness of the center channel in the Honda setup.

josby said:

I did some further measurements and analysis of the center signal (not posted here) to try to figure out exactly what they're doing with it.

With DTS Neural turned off, the center channel is a simple sum of L+R. So, as I'm sure you experienced, the center image is no good with DTS Neural off.

With DTS Neural on, a true upmixing is done, where the center speaker only plays sounds that are present in both the left and right signal. Furthermore, it reduces those sounds in the L and R channels by 10dB.

So, for example take a singer that is supposed to be at the center of the soundstage. In the original stereo recording, his voice will be at equal levels in both the left and right channels. With DTS off, it will play at equal level from the left and right speaker, but also play from the center speaker at +6dB (from summing two identical sounds). With DTS turned on, his voice will play from the left and right speaker at -10dB from what it originally was, and play from the center speaker at the volume it originally was.

Other upmixers, like Harman's Logic 7, do the center similarly to DTS, so this is the generally accepted way to do things. So, it should produce a good center image, and I'm not sure why it doesn't.

I think you're right that the center needs to play louder, so I think your idea of putting a more sensitive speaker into the center has merit.

But I think another part of the issue may be the range of frequencies that are sent to the center. I think the most noticeable thing about hearing a proper center channel is the way the voice of the singer sounds like it's in the center. According to Wikipedia, the fundamental frequency of the human voice is about 85Hz to 180Hz for men and 165Hz to 255Hz for women. And I've read elsewhere that one's center channel should ideally be able to play down to about 100Hz. But as you saw in the graph you quoted, the Honda center signal rolls off around 250Hz. So those fundamental frequencies of vocals won't be coming out of the center speaker no matter what speaker it is. I think that limits the effectiveness of the center channel in the Honda setup.

Click to expand...

i agree, the center channel needs more low end, but anything would be an improvement over what I’ve got now. I work with the Analog Devices Adau1452, (Audio Company) I’ve got a demo kit and the SigmaDSP SPI interface and I'm using the SPDIF optical input on the ADau1452 for another audio type thing-work related, theoretically I could create my own DSP setup but decoding the digital audio signal from the H/U isn’t worth the effort. I have a Salae logic analyzer I could hook up and try to analyze the data stream but again, too much trouble lol.

thans again!

Well I installed the Infinity driver in the center position. The imaging is slightly better. It's now apparent to me more than ever before that the pillar tweeters (awful awful arrangement) are drowning out the center signal. I'm torn between putting in dropping resistors on the tweeters or just 3M gluing felt over the tweeter grills, it's that bad. The Infinity doesn't quite fit the circumference of the original plastic ring on the center dash plastic speaker mount, it's slightly larger, so I had to add washers to get the surround above the ring and seal it with this insulating tape we use here for rattle control and sealing.
Also noticed that as before with the Kicker Center driver (88 dB) the tweeter in the Infinity driver is slightly noticeable and annoying, so same scenario, either mute with felt or cut tweeter leads.
Now I'm paranoid about the center dash cutout getting wonky and worn due to my messing with the center speaker grill. You know, if you mess with anything long enough you'll break it .

Fountainhead said:

i agree, the center channel needs more low end, but anything would be an improvement over what I’ve got now.

Click to expand...

Yeah, at first I thought one could just put use a DSP to undo the high-pass filter on the center, but then you still have a problem because the upmixing algorithm has routed those lower frequencies to left and right since it couldn't play them through the center, so if you're bringing them up in the center, you need to bring them down correspondingly in left and right. But you can't put a low-pass filter on left and right since that would effect signals that are left-only or right-only too.

I found the center image to at least be better when I upgraded my center speaker, but still nowhere near as good as it is now that I'm using a separate Logic7 upmixer, which is really a shame since DTS Neural *should* be a better upmixer than Logic7.

Fountainhead said:

I work with the Analog Devices Adau1452, (Audio Company) I’ve got a demo kit and the SigmaDSP SPI interface and I'm using the SPDIF optical input on the ADau1452 for another audio type thing-work related, theoretically I could create my own DSP setup but decoding the digital audio signal from the H/U isn’t worth the effort. I have a Salae logic analyzer I could hook up and try to analyze the data stream but again, too much trouble lol.

Click to expand...

Yeah, I have been curious about that idea too. I found some info on an Acura forum from a few years ago where a guy had done some work decoding the RS-485 packets sent from his headunit's touchscreen. That could probably be used to help with decoding the RS-485 packets sent from our headunit to control the amp.

Further update:
Well the drivers side tweeter is at the same spl level as the passenger side tweeter but about 5 times closer to my left ear than the other tweeter in my right, so right there is a huge sonic imbalance. The inverse/square rule is not working in my favor here.
I mean really? Who at Honda thought putting a tweeter in your ear was a good idea? If I had the time I'd mount them in the dash, where they should be. in an effort to tame the tweeter output levels (before I eventually insert an in-line dropping resistor with the tweeters) I'm putting 1mm felt in two layers on the tweeters, it made a big difference in revealing the output of my new center speaker.
So I'm forging forward with further tuning in an effort to create a good soundstage.

Before I traded my Sport (with tweeters) in for my LX (no tweeters); I did the same. Even with the 'highs' turned down to zero; the sound from those things was brutal.

Back in the 1980s; the go-to (scientific) solution for overly bright tweeters was to cover them with Kimwipes. A magazine called 'Recording Engineer/Producer' did a high tech analysis similar to yours and concluded that Kimwipes were the most accurate comb filter with a relatively flat curve.

shoegazer said:

Before I traded my Sport (with tweeters) in for my LX (no tweeters); I did the same. Even with the 'highs' turned down to zero; the sound from those things was brutal.

Back in the 1980s; the go-to (scientific) solution for overly bright tweeters was to cover them with Kimwipes. A magazine called 'Recording Engineer/Producer' did a high tech analysis similar to yours and concluded that Kimwipes were the most accurate comb filter with a relatively flat curve.

Click to expand...

Ahh the good old days of nsm 10s