My Focal based audio build

[TOOSTUBBORN2FAIL]

I would try to keep the door speakers crossed over at 80hz or preferably 100hz. Modeling them in a door shows them reaching xmax at 80hz with just 25 watts. Or at 100hz with just 50 watts. This will also prevent the doors from rattling as much as well.

[P3tras]

Thank you! Yes, I was planning to cross them at ~80hz (HPF on amplifier) when subwoofer is installed.

[P3tras]

Getting there... I will leave on road trip in 2 days, so I hope to finish everything until then.

Mounted tweeters like this. It fixed dip at 6-8kHz. Now there is new dip at 15kHz when mic is pointed straight at ear level (no dip when mic is pointed slightly down). But it isn't bad and 15kHz is not as important.


Figured out the noise problem. It is coming from RCA cables. Cables pick up noise easily from factory wiring harness that goes near passenger door (when engine is started). Added external metal shield to the cable, it helped.

Modified source code for Viper4Android, so it would allow all settings for phone speaker (in this case car audio). Haven't tested it, though . Now I should be able to access DDC (basically sound processor, I can cut/add any frequency >~200hz) and set time delay. For time delay there is no selection which channel to delay. Might look into modifying source code for that as well. But that comes later (no time right now and I have 0 Android programming experience).

Last frequency response test (only right side, not both sides). Measurements at various spots to get general idea of the system. I will probably cut 6kHz and add 8kHz (if my modified Viper software will work).


And here's test of rear speakers (Rockford Fosgate). At high frequencies they really perform good.



Also found one strange thing. Head-unit has mixed up left/right channels. Left channel on Android system is right channel on output. Android sends stereo audio signal to second processor (MCU) which then divides signal to 4 channels according to panning set on unit. It seems engineers mixed up channels between Android system and second MCU Not a big issue, swapped RCA connectors and swapped pins for rear speakers. Just panning window is inverted now.

[TOOSTUBBORN2FAIL]

Be careful adding at 8khz. If that's a null, adding there won't do much to the frequency response, but will increase the amount of power the tweeter is seeing and add distortion. You'll know if you add 3db and it barely moves. If so, back it back down and just leave it alone.

[P3tras]

The more I dig into this build, the deeper I get.

Decided that after all I will get proper DSP. Main reason is to get subwoofer in phase with rest of the system (I want bass, but don't want subwoofer to ruin overall sound).

Been looking at different solutions from mini-dsp:

• miniDSP 2x4 - cheap, but only 4 outputs. So 2 for front, 2 for rear speakers and 1 subwoofer = 5. Unless I run subwoofer without DSP and other speakers through DSP (natural time delay ). But not sure if it is going to end well.
• miniDSP 2x4 HD - better processor than standard version, has FIR filters. I could try USB audio, but then issue is with volume controls.
• C-DSP 6x8 - designed for car, has enough outputs. But worse processor than 2X4HD, doesn't have FIR filters.
• Custom solution: miniSHARC + miniDAC8 (8 outputs) + VolFP (remote volume control). Similar processor as 2X4HD (has more computing power, FIR filters + it is floating point) The most complicated and expensive, but probably the best solution. It has I2S input (audio signal that is used inside electrical devices). I could dig into head-unit, find DAC and its I2S lines. Such way I would get true digital signal, without unnecessary digital to analog conversions. But I2S signal is designed to be transferred only few cm, will need some kind of circuit to "boost" signal and convert it back to normal near DSP. For volume controls I could wire stock volume encoder in head-unit to Vol-FP (it also has encoder for volume). In theory, this way will get highest possible quality, while still maintaining stealth look. Complicated, though, complicated...

[sly]

You'll just have to focus on what you want to get out of this system. If all you really want is time alignment then you could buy 2 miniDSP 2x4 units. Just use one for the left and one for the right. That will give you the potential for a stereo 4-way system (8 channels). You could run the 2x4 advanced plugin https://www.minidsp.com/products/plu...dvanced-detail which can give you matrix mixing to create your subwoofer channel.

If you want a simple solution to go digital, you could look at several options for converting your RCA signals to S/PDIF. Here's one I found. https://www.amazon.com/Digital-Optic.../dp/B0055EBH5S Don't know if it's any good but it's cheap and available. Then you could use the carDSP or any of the miniSHARC based products without any complicated wiring. Doing it this way you will still retain your factory volume controls without having to fool with i2s.

With that said, i2s is finicky. You are not going to get it to transfer very far. Most devices that use i2s are designed to fit inside the same enclosure. Even then you have to be shielded to prevent signal degradation. It would be far better to convert i2s to S/PDIF and then make your transfer. Looking at the datasheet, it looks like the miniDIGI can support i2s inputs while giving S/PDIF outputs. https://www.minidsp.com/images/docum...f-MiniDIGI.pdf I would also check out twisted pear audio's website: http://www.twistedpearaudio.com/digital/wm8804.aspx This is an S/PDIF and i2s tranceiver that can give you the S/PDIF output you need.

I have a custom made miniSHARC unit in my car. If you want to eventually play around with FIR filtering, either the 2 2x4HD units or a openDRC-DA8 would do fine. https://www.minidsp.com/products/ope...es/opendrc-da8

You just need to convert your RCA signals to S/PDIF or tap into your head unit's i2s and convert that to S/PDIF. If you use the i2s signals, you are going to have to create a volume control. The openDRC has one on the front of the unit which would be simpler. I believe you can also use an external volume control for the 2x4HD units. Tapping into your head unit's volume control information is beyond the scope of what I can advise. Your best bet would be to check out miniDSP's forums for support on that.

[P3tras]

Thanks for the tips!

You got some good ideas. The main reason I do not want to convert analog to Spdif is because of unnecessary conversions (there are always losses each time signal is converted). Talking about twisted pear devices, I was looking at "Teleporter": http://www.twistedpearaudio.com/digital/teleporter.aspx

It is basically device to transmit I2S signal over long distances. I am pretty sure it converts signal to LVDS and then back to I2S. Such way you would avoid quality losses related to spdif (as I2S has separate lines for clock and data). Bit too expensive for my taste, though.

I have disassembled my head-unit and removed CD drive (I have never used it). There is more than enough space to fit miniSHARC inside. Such way I could easily tap into I2S signal, as distance is very short. But few problems related to that:

• Didn't find DAC IC itself. Probably it is in "Android" PCB which is under metal shield. Going to remove metal shield today and check.
• Such way I will only get sound from Android (at least I think so). As I understood on device there are separate audio streams. From android, from the radio, from AUX and from bluetooth. They all go to sound processor (which I found) and MCU controls which input to use. If I use I2S, I will only get sound from android.

But I also came up with few ideas how to avoid issues.

• I was thinking about having 2 audio input signals to miniSHARC. First one would be high quality I2S straight from android and second is head-unit's analog OUT converted to Spdif. So first input would give highest possible quality, while second input allows to use radio, AUX and etc.
• Volume and signal can be controlled through VOL-FP. It has encoder for controls, so I think I could hook stock encode on head-unit.

However, everything is still in "thinking process". First I need to find I2S lines in head-unit, then make sure that miniSHARC can switch between Spdif and I2S inputs (it has both, just I am not sure if it can switch between them). Why I have to make everything so complicated

[sly]

You have 2 options for switching between SPDIF and i2s with the miniSharc... First you could just leave both inputs open. You can use either input or both at the same time. Whenever you switch sources, the other input will just not receive a digital signal.

Another option is to create a profile that uses a particular input. The miniSharc can store 4 user selectable profiles. These can include settings for every parameter from time alignment to EQ to source. You could have one profile set to listen to i2s and another profile to listen to SPDIF. It would be a manual process to switch between sources, though.

[sly]

Originally Posted by P3tras

I have disassembled my head-unit and removed CD drive (I have never used it). There is more than enough space to fit miniSHARC inside. Such way I could easily tap into I2S signal, as distance is very short. But few problems related to that:

• Didn't find DAC IC itself. Probably it is in "Android" PCB which is under metal shield. Going to remove metal shield today and check.

That's a good idea. Do you have any pictures of the inside of the head unit? I've been wanting to add RCA outputs to mine and have been looking at this site: https://www.ifixit.com/Guide/Subaru+...+Out+Mod/17985


It's not the same head unit but it uses a similar processor to ours. There's got to be a DAC in there. We just need to find it. I would much rather skip the factory DAC all together, tap into i2s and use my own DAC.

[P3tras]

Oh, I am not using stock head-unit. But I found the "DAC". In fact it is also ADC for microphone and video encoder. It is Rockchip RK-1000S circuit (on right from the big chip)



Not sure if it runs in master or slave mode (according to datasheet it supports both). I suspect that it is slave, but no way to be sure without oscilloscope.

Also, as I understand you cannot connect 2 devices (inputs) to one I2S source? Such way I would need to cut PCB tracks and it would be difficult going back if something does not work

Good advice regarding 4 pre-sets. I knew about them, but didn't think that you can also use them to switch between inputs. VOL-FP also has "source" selection, but I am not sure if it is implemented, yet (as it is not described in data sheets)

P.S. If you are looking for low level signals, why not just google amplifier's data sheet and check which pins are low level inputs?

[sly]

Originally Posted by P3tras

Also, as I understand you cannot connect 2 devices (inputs) to one I2S source? Such way I would need to cut PCB tracks and it would be difficult going back if something does not work

Good advice regarding 4 pre-sets. I knew about them, but didn't think that you can also use them to switch between inputs. VOL-FP also has "source" selection, but I am not sure if it is implemented, yet (as it is not described in data sheets)

Correct, only one input can be used with an i2s source. If you want to send a signal to multiple devices, you will need an i2s matrix mixer such as a miniDSP 2x4HD. Feed it an i2s input and then have it send SPDIF out to multiple sources. I wouldn't be too concerned about converting i2s to SPDIF. They are both PCM formats. SPDIF just combines the clock signals rather than running them through separate channels. It has better shielding and is less prone to signal degradation. The tradeoff is that with SPDIF you can introduce a small amount of jitter.

The VOL-FP will select between sources. Just press the button and the display will let you rotate between presets 1-4. Press the button again and it will switch to that preset. It functions both as a volume control and as a preset selector. Like I said, each preset can store any number of variables. You can store individual EQ, level, FIR filters, compression settings, delays, matrix mixing and sources. Everything is settable via the presets. I use 4 presets on my miniSHARC unit. Presets 1,2 and 3 are for time alignment, gain and EQ settings for driver, passenger and middle seating positions. I use preset 4 to add dynamic range compression for when I play DVD's and want to compress the vocal tracks. You can literally set anything with the presets.

You can also use a universal remote control to do these wirelessly but I haven't played around with that.

[P3tras]

Thank you! I bought miniSHARC + miniDAC8 + VolFP + miniDC isolator

Stock system works like this. Android > DAC > Audio processor > amplifiers. There are also radio, bluetooth and AUX analog signals going to the audio processor and MCU selects which input to use (or which inputs to mix together).

My solution: Android > miniSHARC I2S > miniDAC8 > amplifiers (solution for best quality, when I do not need radio, AUX input and etc).

However, when I need radio, it gets bit more complicated. Android > MiniSHARC I2S > miniDAC outputs stereo signal (with no processing) > stock Audio processor (it mixes android sound with radio, AUX and etc) > analog to spdif adapter > MiniSHARC > miniDAC8 > amplifiers. In this case I basically use processor as stock DAC and then convert output analog signal from radio to Spdif and supply it to MiniSHARC again.

If I2S signal will not work, my backup plan is to use USB audio with I2S converter.

[sly]

How about something like this...


Android>i2s input to miniSHARC> DAC8
Stock audio processor>RCA to SPDIF converter>SPDIF input to miniSHARC>DAC8


Leave both miniSHARC inputs unmuted. Now you can play off the Android but still get system sounds when needed, including radio.

[P3tras]

Quite smart, quite smart! Sometimes the best ideas are the simplest ones

Tomorrow I hope to check I2S signals via oscilloscope to make sure everything is OK. Bought most of the supplies for subwoofer box build. I hope to build the box until DSP components arrive.