[lucksin]

I have been struggling for a few days now trying to tune my DSP to get a relant better sound and it all failed miserably. My initial steps were to cut and boost the bands to see what works and what not.

I was shooting in dark thinking at least something would work however everything failed. The best I could do was time aligned and able to get somewhat of a front stage, yet no center stage.

I ordered an RTA mic to understand what is that I am dealing with and below are the EQ curves that I am dealing with:

If you check the Front, there is a boost at 40hz, then it rolls down till 100, there there is another boost at 200hz the it gradually rolls down till 1.5khz, there is another boost at 5khz mark.

compared to the rear which has no boost until 100hz and then gradually increases until 10k.

Also, My speakers can very well play until 25khz, however, infotainment is restricting anything beyond 10k. however, it doesn't matter much but still. The above samples were taken with flat DSP setting with pass through. Also, there is a difference of 10db of sound level between front and rear. Rear being lower.

It is for FYI for everyone trying to upgrade speakers on stock infotainment.

[Rehaan]

Interesting data lucksin, thanks for sharing.

Re-uploading the charts with X and Y axis labels shown would really be helpful....

Originally Posted by lucksin

Also, there is a difference of 10db of sound level between front and rear. Rear being lower.

Absolutely hate this cost-cutting trend!! Happens not only in budget hatchbacks, but even 20+ lakh SUVs.

[lucksin]

Originally Posted by Rehaan

Interesting data lucksin, thanks for sharing.

Re-uploading the charts with X and Y axis labels shown would really be helpful....



Absolutely hate this cost-cutting trend!! Happens not only in budget hatchbacks, but even 20+ lakh SUVs.

My bad, I will re-upload with the axis.

i am not sure how to tackle the low output on rear, plus boosting lower frequencies is generally not good. I will be bridging the front 2 channels to rear. I am near completion on front 2 channels though.

Thinking of going full active with polks

[alpha1]

Originally Posted by lucksin

I ordered an RTA mic to understand what is that I am dealing with and below are the EQ curves that I am dealing with:

Which mic? Each mic will have its own EQ calibration curve (although even cheap microphones are lot more straight and linear compared to speakers/drivers)

If you check the Front, there is a boost at 40hz, then it rolls down till 100, there there is another boost at 200hz the it gradually rolls down till 1.5khz, there is another boost at 5khz mark.

compared to the rear which has no boost until 100hz and then gradually increases until 10k.

How did you measure the spectrum? Did you factor in the room modes (standing waves)?
Did you keep the mic stationary or did you move it around in space?
Did you use pink/white noise or was it since sweep or perhaps maximum length sequence?

[lucksin]

Originally Posted by alpha1

Which mic? Each mic will have its own EQ calibration curve (although even cheap microphones are lot more straight and linear compared to speakers/drivers)

I ordered a Dayton iMM6 calibrated mic with a USB soundcard(Practically making it a UMM mic)

Originally Posted by alpha1

How did you measure the spectrum? Did you factor in the room modes (standing waves)?
Did you keep the mic stationary or did you move it around in space?
Did you use pink/white noise or was it since sweep or perhaps maximum length sequence?

sweeps were taken with the moving microphone method(MMM) with a full range pink periodic noise at -5db. I use ear-to-ear movement with >100 averages.

[Rehaan]

Originally Posted by lucksin

i am not sure how to tackle the low output on rear, plus boosting lower frequencies is generally not good.

On the Kia's with the Arkamys systems, one hack that I read on the forum that works surpsisingly well is selecting the "Lounge" advanced preset. The rears instantly get louder.

Maybe see if anything like that is possible with the Baleno.

[lucksin]

Originally Posted by Rehaan

On the Kia's with the Arkamys systems, one hack that I read on the forum that works surpsisingly well is selecting the "Lounge" advanced preset. The rears instantly get louder.

Maybe see if anything like that is possible with the Baleno.

Nothing like that in Baleno. It's more vanilla than vanilla.

Anyway, here's my new tune for the front that I will be importing tomorrow morning and playing with. There are still very minor positive and negative peaks that I am not able to flatten, however, I will still give it a try tomorrow.

I still need to do the Phase alignment of all speakers, however, I think I will do that once I go fully active on the front this weekend.

[Rocketscience]

The rears have known to be much less loud in many cars but curiously in many of them when I got the speakers changed the drivers along with the rear magnets and markings on them seem identical, no idea why?

A curious case along with an accidental solution was discovered by me in my first gen i20, so I was disappointed by its audio quality and decided to get all the speakers changed, I narrowed down to JBL Power Series and they sounded fantastic in the store but once installed in my car they sounded garbage and the volume level was too low even at max, the car accessories guy suggested either a infotainment system change but that would have ruined the then great looking interiors of the car or otherwise adding an amplifier but I had anyways spent the entire 25-30K or so of budget allocated by dad on speakers and good quality damping. (I was 17 then and although it was my first car, I was not yet driving it, dad was before I turned 18 a few months later)

The solution anyways found was using AUX, I tried my then phone Nokia 5800 and the output was instantly better, I then tried iPod Touch and the output was better + louder. Then I tried my iPod Nano and that was the ultimate solution, the volume level was dramatically louder, the sound quality was night and day better and the system came to life!

It was like switching from a poor quality cassette system to a Pioneer CD system, although I didn't have mic or the knowledge of audio spectrum then, the HU for some reason had poor output when playing CDs or USB drives directly but using AUX when the DAC part was being handled by external device, the best of which was iPod Nano, the characteristics changed.


It should be worth persuing with different input modes by you as well, no matter how redundant it would feel.

Another example is, in my Honda City 4th Gen the audio quality between Bluetooth mode and USB using Pen Drive or iPod connected was indistinguishable and great in both cases but in Creta (1st Gen SX), the USB quality is great but using bluetooth, the quality is way worse, not only is the bottom end punch missing but the top end too is non existent and there are audible cracks and pops in louder sections of songs making it completely unusable.

Bottomline, do check all the modes one by one and in detail, you might be surprised to find the variances.

[alpha1]

Originally Posted by lucksin

Nothing like that in Baleno. It's more vanilla than vanilla.

Anyway, here's my new tune for the front that I will be importing tomorrow morning and playing with. There are still very minor positive and negative peaks that I am not able to flatten, however, I will still give it a try tomorrow.

I still need to do the Phase alignment of all speakers, however, I think I will do that once I go fully active on the front this weekend.

Which peaks are you referring to? The one between 50-55 Hz?
I think it would actually sound better rather than worse compared to a ruler flat frequency response.

What is the green smooth line?

[Bass&Trouble]

Originally Posted by lucksin

I ordered a Dayton iMM6 calibrated mic with a USB soundcard(Practically making it a UMM mic)

Instead of measuring with a mic why dont you tap into the speaker line, perhaps reduce the level to line level using a line output converter and feed that to a RTA? That way you eliminate the acoustics completely and get the exact EQ applied to the signal.

Then use the input EQ section of a DSP to flatten the signal and finally apply DSP as required in the output section.

The only things an RTA wont be able to catch are things like delays etc applied to any of the channels in high end OEM systems, but I dont think Maruti infotainment must be using any.

[lucksin]

Originally Posted by Rocketscience

...

The solution anyways found was using AUX, I tried my then phone Nokia 5800 and the output was instantly better, I then tried iPod Touch and the output was better + louder. Then I tried my iPod Nano and that was the ultimate solution, the volume level was dramatically louder, the sound quality was night and day better and the system came to life!...

The difference in sound between Bluetooth/CD vs AUX/iPOD is due to processing happening within infotainment where as in case of Aux its a pass-through (no processing happens, 2 analog channels in and 4 amplified analog channels out), and in the case of iPOD - it's via apple inbuild DAC which is much better compared to the infotainment processessing.

I know this cause many years back I was working with ETON in designing their home audio speakers(majorly their Mobile apps and firmware integration) and AUX was always a passthrough on the board with no processing happening.

Unfortunately, Baleno has only 3 input source

- CarPlay/AA (I will 99.9% of the time will use this)
- USB
- Bluetooth

Originally Posted by alpha1

Which peaks are you referring to? The one between 50-55 Hz?
I think it would actually sound better rather than worse compared to a ruler flat frequency response.

What is the green smooth line?

If you look at my first post, there are 5 images, the first 2 are from front Left and right. If you look at it, the peaks are at 60hz, 200hz, 1Khz, 5Khz etc, The one you are referring to is my eq'd curve which is much smoother. And indeed it sounds much nicer everything just blends better, The boost at 60hz is something I have to re-eq since it is a little bass-heavy sound that I am getting at the moment. I actually wanted to try this way.

The green line is called the house curve a kind of target response that you want to align for a signal response from a speaker while correcting the response

There are many house curves :

- JBL
- Whiteledge
- Flat
- Jazzi's and so on.

JBL is the most common house curve used

Originally Posted by Bass&Trouble

Instead of measuring with a mic why dont you tap into the speaker line, perhaps reduce the level to line level using a line output converter and feed that to a RTA? That way you eliminate the acoustics completely and get the exact EQ applied to the signal.

Then use the input EQ section of a DSP to flatten the signal and finally apply DSP as required in the output section.

The only things an RTA wont be able to catch are things like delays etc applied to any of the channels in high end OEM systems, but I dont think Maruti infotainment must be using any.

Awesome questions, however here is my thought which may differ from others,

1. I need an analog RTA to tap into one of the speaker line which is quite expensive compared to an RTA mic($40)
2. I believe EQ is always done based on the environment. Each speaker works differently and we have to tackle the curve according to the speakers and environment, for ex, with the same set of EQ, 2 different brands of speakers will sound completely different. In fact, my left is 1.5db louder then right with same EQ and this is normal. no 2 speakers are manufactured identical.
3. with RTA Mic, you can see the dips(sound cancellation) and the peaks happening because of sound reflections.

remember most of what we hear is from sound reflection in a closed environment, and hence It is always better to eq the system within the environment, The line level RTA can only tell me how the signal is coming from the source not how it performs or sounds from a speaker

With RTA using Mic, you can not only eq, but can do time alignment, phase alignment, independent speaker leveling etc

[alpha1]

Originally Posted by lucksin

If you look at my first post, there are 5 images, the first 2 are from front Left and right. If you look at it, the peaks are at 60hz, 200hz, 1Khz, 5Khz etc, The one you are referring to is my eq'd curve which is much smoother. And indeed it sounds much nicer everything just blends better, The boost at 60hz is something I have to re-eq since it is a little bass-heavy sound that I am getting at the moment. I actually wanted to try this way.

You've actually done pretty good I don't think it is going to become more straight - which DSP/ parametric EQ correction you are using?

The green line is called the house curve a kind of target response that you want to align for a signal response from a speaker while correcting the response

There are many house curves :

- JBL
- Whiteledge
- Flat
- Jazzi's and so on.

JBL is the most common house curve used

But my understand was the house curves are meant for "house".
What that means is that a flat anechoic chamber frequency response translates to a "house curve" when subject to real life rooms and halls.

e.g. a flat frequency response speaker when measured in a large hall will show a reduction in high frequencies. Hence, house curve will matter when measuring speaker response in a hall or room; but here you are measuring in a car, which is practically near field application.

Ofcourse deliberately EQing it to sound like something else is perfectly understandable (for some reason most people prefer more bass, less midrange)

remember most of what we hear is from sound reflection in a closed environment, and hence It is always better to eq the system within the environment, The line level RTA can only tell me how the signal is coming from the source not how it performs or sounds from a speaker

With RTA using Mic, you can not only eq, but can do time alignment, phase alignment, independent speaker leveling etc

Additionally: it is the speakers and their placement that is the most non-linear (and most troublesome) of all components in the entire signal chain of sound reproduction. Taking a line-out will give perfect ruler straight measurements mostly.

[Bass&Trouble]

Originally Posted by lucksin

Awesome questions, however here is my thought which may differ from others,

1. I need an analog RTA to tap into one of the speaker line which is quite expensive compared to an RTA mic($40)
2. I believe EQ is always done based on the environment. Each speaker works differently and we have to tackle the curve according to the speakers and environment, for ex, with the same set of EQ, 2 different brands of speakers will sound completely different. In fact, my left is 1.5db louder then right with same EQ and this is normal. no 2 speakers are manufactured identical.
3. with RTA Mic, you can see the dips(sound cancellation) and the peaks happening because of sound reflections.

remember most of what we hear is from sound reflection in a closed environment, and hence It is always better to eq the system within the environment, The line level RTA can only tell me how the signal is coming from the source not how it performs or sounds from a speaker

With RTA using Mic, you can not only eq, but can do time alignment, phase alignment, independent speaker leveling etc

The first job should be to isolate the source signal and clean it out, correct it and then take the subsequent variables in the chain 1 by 1. You could analyze the head unit's output by merely using a PC spectrum analysing software, no need for any expensive gear. You would get far better and relevant data about what the head unit is doing to the signal than a $40 mic. Armed with that info, all you need to do is EQ it to get a flat signal and then do the crossover and all other processing. Finally you can use the mic and analyze the sound as produced in car using the RTA, impulse response etc and make the corrections for the deficiencies in the speakers and the car's acoustics.

[lucksin]

Originally Posted by Bass&Trouble

The first job should be to isolate the source signal and clean it out, correct it and then take the subsequent variables in the chain 1 by 1. You could analyze the head unit's output by merely using a PC spectrum analysing software, no need for any expensive gear. You would get far better and relevant data about what the head unit is doing to the signal than a $40 mic. Armed with that info, all you need to do is EQ it to get a flat signal and then do the crossover and all other processing. Finally you can use the mic and analyze the sound as produced in car using the RTA, impulse response etc and make the corrections for the deficiencies in the speakers and the car's acoustics.

Sorry for long pause from my end, had health issues due to which i was on bed rest.

Could you help me with how can I take a reading on my PC? I do have REW but how do I connect the output signal to my computer to see what the infotainment is outputting?

Whatever you mention makes lot of sense. Thanks

[greenhorn]

Imm6 low end accuracy is supposedly all over the place. Was told this when i made some measurements with the imm6. Got a umik from olx and much happier!

Regarding getting the Hu output, i would take the input from one of the speaker wires, pass it through a resistor divider (10k:1k) should do fine
https://en.m.wikipedia.org/wiki/Voltage_divider

Take the output and plug it to the usb soundcard line in using the appropriate connector (you can get solder less rca connectors if you don't want to solder)
The rest should be the same. If you want a bit more safety then use a hilo convertor or add a 47uf capacitor in series with the input - but both of these will impact the frequency response a bit.

Remember to use the laptop and usb soundcard on battery power so that there is only one common ground (signal ground)