It seems that a few folks might want a sub out option on their Empegs. In the words of CHiP, it would be great if there was a

"User selectable, adjustable low pass filter and volume control accessable from the empeg menus."

for a sub out board.

Stu

This would be nice to add to Hijack, wouldn't it?

Agreed!

Stu

Easy enough to add, if/when there's a use for it.

Cheers

So are you saying I should put this circuit together and get back with you?

Stu

Basically, yes.

Cheers

I was really hoping this RioCar would have come shipped with a sub out. I mean, to put alot of energy into the 20 bands of fully parametric EQ without adding a parametric sub output was negligent of the car audio community!

Currently, I have a four amp system in my car (true 4-way). The probalem comes when I want to individually EQ the front and rear. EQing the front is easy using the 4 x 5 band mode, but leaves the remaining pair of bands to allocate for my rear and sub amp. That configuration sucks, so I just end up using the front EQ bands for both the front AND rear and use the rear EQ bands for the sub (overkill in a way).

If there is realy an easy way to make one RCA output and a program to control it, that would be sweet.

I don't know anyone with subs that would find the 4 x 5 mode that much more helpful. Let's make it 4 x 5, and one output with fully parametric and lowpassed (adj. too) control!

I'll PayPal you if you can figure it out. WHere would one tap into the board to get this extra output too?

For me, the hardware will be pretty easy. I will be designing it this week while I'm out of town. It will be very small and hopefully of good quality (there is an unavoidable tradeoff between size and quality). The problem for me is the software aspect. I have no programming experience and minimal exposure to C and Linux (esp. Linux), so developing the software would be a very daunting task. Shonky (Christian) has made and tested a breadboarded sub out hack for his Empeg, complete with software, however, I don't believe his software is of the parametric variety you want though. It has preset filter points only.

The sub out board would tap into the "IIS" header located under the 2nd drive bay.

Stu

I only have one drive in my unit and never plan on upgrading, so size doesn't matter too much and I prefer quality over bulkiness. I can personally build my own RCA jack with op-amp and filters, but the Hijack programming is not my forte either. So, if one was able to bring both the hardware and Hijack to the table, I'd gladly pay $100 USD.

Features I want:
24dB/ oct. and 12dB/oct lowpass crossover (selectable), gain, 180 degree phase switching via Hijack (more preferrably, fully variable or semi-variable phase adjust), and it would be very nice to have a coule bands of fully parametric EQ like the front and rear already have (I'd pay ya extra for that if necessary).

My stereo system is the most important thing for my car.

Jonathan

My stereo system is the most important thing for my car.

Kinda goes without saying for a large number of us, doesn't it?

Most of your requests are dictated by the DSP.

24dB/ oct. and 12dB/oct lowpass crossover (selectable)

It only has a 4th order filter (24dB/octave - I think I said decade somewhere else) in mono sub/centre channel mode. In stereo sub mode, the crossover is 2nd order (12dB/octave)

Crossover frequency is not fixed as such but requires intricate knowledge of the DSP to select the 20 odd coefficients required. The manual supplies coefficients for 40,60,80,100,120 Hz only. These even change with sampling frequency which is not a problem on the empeg since it only uses 44.1kHz

gain

Easy and done.

180 degree phase switching via Hijack (more preferrably, fully variable or semi-variable phase adjust)

Technically very easy but the DSP code isn't programmable. Could be done in hardware quite easily though - wouldn't be hijack switchable then. I'm not sure about variable phase though. How does that work? I understand how you can do 180 degree shift by inverting the signal but not variable really.

it would be very nice to have a coule bands of fully parametric EQ

Again not without programming the DSP and that ain't going to happen.

It uses 38kHz when in radio mode, so you need to deal with 2 sets of crossover frequencies.

I have a DSP test program from philips which, in theory, will give me coefficients for other frequencies - if you have a request I can see if it'll give me values.

Hugo

OK thanks Hugo, that should be no problem to do, although I think I might wait until a board actually exists before I do anymore. I kinda jumped the gun. There is no way I am going to have time to design anything properly in the next couple of weeks.

I assume this program you mention is not for public release? If so, perhaps you could tell me what it's called and I can go ask Philips for it. They didn't seem to mind giving out the doco so maybe they won't mind with this program given the DSP is getting on a bit now. I don't have any specific frequencies though. Perhaps 40,60,80,100 and 120 will be enough for everyone? I would perhaps like to go in units of 10Hz though. The kernel space used would be so minimal.

Also why do you say "in theory"? Sounds like it's a bit dodgy.

If I had the program, I'd run it for all of the frequencies from 40 - 120 Hz in 1 Hz steps (minus the ones already listed in the PDF), just to make it fully parametric, at least if there is room for that many settings in the kernel. I will have a hardware design out this week, roughly based on the active filter Crystal eval. board for the 4334.

Stu

To do that you would take up:

number of crossover freqs X number of entries per freq X 2 bytes per entry X number of sampling freqs (Need 38 and 44.1kHz)

So 80 X 20 X 2 X 2 = 6400 bytes which isn't too bad I guess. Some one has to type them in manually though... Just typing in the ones I have now was a bit of a chore. I assume also that on the fly calculation is out of the question.

Depends on this special program. It might do a range and print it to a file which could be massaged into nice C code.

Having this large number will also required more bits in the hijack space. I have already taken 15 more bits for the centre gain (6), sub gain (6) and the sub cutoff (3). What you are talking about is an extra 4 bits the for the cutoff. The gains are also a bit "chunky" at the moment.