Do you have access to an oscilloscope? If not, you can't really set them the "best" way. This is what I did, as a rookie (so don't take this as correct, but just my experience):

• Unplug speaker outputs.
• Play 1Khz test tone just under clipping volume on the headunit. Since I don't know the clipping point, I set my volume to 30/35 assuming that was sufficiently conservative but not far off the point where it clips.
• Look at input levels in the DSP (I have the miniDSP c-DSP 6x8) and adjust. I adjusted to -5dB, to leave headroom to EQ up.
• Using a DMM set to AC voltage, set gains on the amp. My JL amp came with the target voltage based on speaker impedance. What I noticed is my 6 channel amp has 3 gain adjustments which each control 2 channels. The output voltage varied a small amount between both channels. I set the gain to the lower voltage of the 2 channels was correct, and then I took 0.1dB adjustments down in the DSP software for the higher channel in order for the voltage to match.

Again, not sure how much of the above is correct, but it's what I put together reading guides here and not having access to an oscilloscope.

 

or, from Andy at Audiofrog:

Andy's tech for today:
Settings gains with a DSP.

After tech supporting someone yesterday, it occurred to me that this process may require a little explanation because it's a little more complicated than the standard process we're all used to using with analog gear.

With analog gear and a scope, we're used to finding the maximum undistorted output of a piece of gear and setting the input sensitivity of the next piece of gear to provide maximum undistorted output for it and then ultimately, setting the "gain" of the amplifier with some "overlap" so that the system sounds loud and isn't noisy. A little clipping isn't a big deal in ANALOG devices. This process was originally super important because we had a long chain of gain stages, each with its own noise floor and head unit output was really low-like 100mV. Systems were noisy.

In most cases, products don't have input sensitivity AND output gain controls, with the exception of Audio Control preamp processors. Most only have an input sensitivity control.

With DSP, there is an additional consideration. For DSPs, we have to be concerned with analog input gain when we are using analog input. We also have to be concerned with digital gain. They aren't the same.

When an analog signal goes into a DSP, it's converted to a digital signal. ADCs (analog to digital convertors) are designed to convert analog signals of up to some maximum voltage into digital signals and some number of volume control steps are available to describe the voltage. The bit depth (16 bit, 24 bit, 32 bit, etc) determines the number of possible volume levels that can be used to describe the input voltage when it's converted into digital.

This ADC has a maximum input voltage and it isn't adjustable. If the maximum signal sent from the radio to the ADC is only a small part of that range, then the resolution of those volume steps is compromised. For a 16 bit signal, there are 65,536 steps (2^16). For a 24 bit signal, there are 16,777,216 steps (2^24).

If our ADC has a max input voltage of 4V and our input signal is 2V max, we waste half of those volume steps--we reduce the resolution with which the ADC can describe the signal. If our input signal is higher than the maximum input, we "clip" the input and that produces distortion.

Digital distortion isn't like analog clipping. It isn't smooth and it becomes nasty sounding instantly.

So, we don't want to overdrive the input to the ADC. In some DSPs, this is set with standard jumpers or pots to match the input voltage to the ADC's maximum. This is the case in the Helix Pro and it's well documented in the manual. For many other DSPs, this isn't so obvious. If the bit depth of the ADC is 16 bit, matching these is pretty important. If it's a 24 bit ADC, it's less important because there's plenty of resolution to throw away. It's ALWAYS important to be sure that we don't OVERDRIVE the input.

Once the signal has been converted and sent to the processor, we often have input and output level controls. The input level control in the digital chain is NOT the same as the input sensitivity control for the analog signal before the ADC and the distinction is important.

In digital audio, there's no gain. There are no values above zero. Maximum signal is defines as "all high bits". There are no numbers to describe a signal greater than that and if it's exceeded, the numbers are sort of undefined. That's digital clipping and it sounds horrible--nasty--mechanical--kind of like a square wave, but it happens instantly.

So, in the DSP UI, if you have your input level control at 0dB (or 100% in some GUIs), you have no room for any boost. In that case, a signal that comes from the radio that reaches maximum in the ADC and is then sent to the DSP as all high bits is as loud as it can be. If you boost the EQ by 1dB...BAM...clipping. If you use a crossover filter with a Q higher than .707...BAM clipping.

It matters where in the digital chain these level controls are placed. Attenuating the digital input is important if you're going to do any boosting of the signal in the EQ or crossover or if you're going to boost the output of the sub. If the input reaches all high bits and you boost the level of the sub channel in the DSP...BAM...clipping--nasty clipping.

So, when you are tuning with a DSP and you hear nasty distortion, it's important to try to figure out where it's coming from. If you attenuate the DIGITAL input and it goes away, then it's in your EQ/Crossover/Output level settings. If you attenuate the digital input and it doesn't go away, then you're overdriving the ADC and you should adjust the analog input sensitivity. If there's no analog input sensitivity adjustment on your DSP, then you have to use the radio's volume control judiciously or you need to build a voltage divider to attenuate the signal.

My suggestion for this is to set the analog sensitivity to give you the maximum resolution if possible. Then, attenuate the digital input level in the UI. Then, tune the DSP. Then, with the amp gains turned way down, adjust the output levels of the DSP so that you hear NO distortion on any channel.

Then, set adjust your amp gains all by the same amount to set the system level. A scope and a sine wave track aren't going to be all that helpful unless you know precisely the frequency in your signal that's at the highest level and you have a sine wave track for that particular frequency.

Just get close doing this by ear so they system plays loud and so that noise is at a minimum. Or, if you MUST use a scope or a DD1 and s sine wave disc, set the gains with the input and output levels in the DSP at -3dB or so and with the EQ set to flat and all crossovers at a Q of .707 or lower. Then, turn the digital input and output gains down, do your tuning and then turn them up proportionately until the system is loud and there's still no distortion. When you do that, you are ONLY optimizing the resolution of the digital signal.

 

Thank you all. Sorry in the delay. I have been busy with a few stuff.

No I have no oscope. The installers tune it by ear, but I get some hiss and I believe this is due to the gain.

For the head unit, I went ahead and pulled the trigger for the new joying head units, have read a few positive reviews, hope they turn out to be good. Plan to use the coax output of Joying to DSP. The reason why I went for it is the flexibility of screen positioning, my truck has no option for an aftermarket. So planning ot get a 3d printed screen holder above the current screen with some sliding assembly. Let me see how it all turns out.

 

ugh, i'm still confused about setting gains from my HU to the DSP despite re-reading Andy's guide since he doesn't really address how to detect the HU's clipping point.

I have a Kenwood DMX906S that will connect via RCA input to a Helix Mini DSP.

I don't know how best to determine the HU's max volume output before clipping. Should I use my oscilloscope to detect flattened peaks with 1kHz and 40kHz 0dB test tracks?

The Helix manuals says to set the HU volume to 90% of the volume range and then adjust the input sensitivity pot on the Helix until the Clip light illuminates, then back off the sensitivity slightly until it goes out.

What if my HU's output at 90% volume is already clipped? How would I know it's not until after connecting the rest of the down stream gear?

 

I don't have a link to the testing, but that head unit should be completely free of clipping until the last couple of clicks on the volume knob, maybe even at 100%. Using about 75-80% of the volume for setting gains is a good idea because it allows for a bit of extra range on the head unit to get your quietest recordings to play loud enough. The ideal gain structure would have you using about 80%, or more, of the head unit's volume for loud listening. If you find that you're only using 50% of the volume range for enthusiastic listening, your gains are too high, and you're not going to be able to fully take advantage of some of the head unit's features, like loudness, to compensate for low volume listening.

 

found this that should help me with setting the HU volume output:

Then I'll adjust the Helix Sensitivity until it clips, then back off to no clip light.

Also found these instructions in my archives:

Tools needed :
Oscilloscope or DD-1 (use 40hz and 1khz )
Sine wave tracks

Start by turning off your amps. Play a 1kHz sine wave recorded at 0 db (full volume). You can download these and burn them to a disc or put them on a thumb drive if your head unit will play them. Connect the oscilloscope to an RCA output on your head unit. Bypass all eq, tone controls, and any other effects the head unit may have. Turn everything off... Now play the sound file and look at the oscilloscope. Turn the receiver up all the way (amps off). If you see the wave start to develop flat lines on the top and bottom, you are clipping and need to turn the head unit down. The point where you get no clipping is the max volume of the head unit. My Kenwood will go to 40 but I got clipping around 38. 37 is the maximum volume I can use and avoid clipping. Repeat this process for both channels.

Do this with a 40hz tone for sub channels. Use the lowest volume of clipping between high sine wave and low and that is now max volume for the head unit.

So if low wave clips at 34 and high wave clips at 37, 33 is now your max volume from the headunit.
Next you need to connect the dayton 408 and do the same thing. Mute all channels (amps still off) and then play only one. Turn your head unit to the maximum volume you determined previously and connect the oscilloscope to one output on the dayton dsp . Ensure all EQ is bypassed on the dayton dsp. But you have some basic crossovers set to protect the speakers. Turn the output gain down about half or so to ensure the dsp doesn't clip and put the master volume at 0db. Now raise the input gain on the DSP for that one channel until you just start to see clipping. Turn it down until it just goes away. This is your max input setting. Do this for the other channel.
***set the above with 1khz tone for mid and tweeter channels, and 40hz tone for sub channels

Next you want to set the outputs of the dayton dsp 408. Pick one channel and connect the oscilloscope to it's RCA output. You've already set the gain on the head unit and the input section of the D408. Now set the output gains the same way. Raise the output slider within the D408 software until you see clipping on that RCA. Now back it down until it goes away. Do this for all your output channels (amps still off).
***set sub output channels with a 40hz tone. And all others outputs with a 1khz tone.

Now you have unity gain from the head unit all the way through the D408. Now it's time to set the gains on the amps. You essentially do the same thing. Provide a 1kHz signal to each speaker, one at a time with the speakers hooked up to give the amp a proper load.
(40hz tone for sub channels)

Connect the oscilloscope to a speaker output on the amp . Turn the amp gain all the way down. Turn the head unit to max volume without clipping. Play the signal and then turn the amp gain up slowly until you start to see clipping. Turn it down until it's clean and then mute the channel and go on to the next one.
When you do this for tweeters, make sure you have a crossover setting enabled within the D408. Put in a high pass at 6kHz or so to provide some speaker protection. The other speakers won't need any crossover. 20kHz is a safe signal to play. Just don't take too long tuning the tweeters. They can overheat if played too long.

Now when you set your levels, you will only turn down channels, not up. Use your SPL meter to find the quietest speaker. Now turn down all the other speakers until you match the level of the weakest. Never boost signal when setting gains this way or you will clip. The system is already at max level and so any tuning should only involve cutting, not boosting. You can boost some individual levels within the EQ as long as the overall signal is not boosted. Tune EQ by cutting the high spots first. Anything that is too loud in the response needs to be cut. Then provide a minimum boost to any weak areas. Ensure your overall signal level doesn't increase to avoid clipping.

 

Hello,

Thanks for the information.

I have a kicker front row and two amps that powers front components and one sub.
I do not have an oscilloscope. Is it possible to set the gains using a digital multimeter..?
Could you please share it.?
Thanks in advance

 

You know, you'd think that todays equipment would be smart enough to make all of this "automatic"... You play a track on the head-unit and then each device will just adjust input/output levels to be optimal. Or at the very least, if a piece of equipment detects clipping, just automatically limit things to avoid it. It just seems like gain setting is always just a big deal - and it really shouldn't be that hard, IMO.

There has to be a better, easier, more fool-proof way...

Maybe I'll add that to the amp that I invent that just has one big pool of power and you get to configure how much power is allocated to each channel. ;-)

 

@jtrosky - i could see this being a feature of an integrated DSP/amp where it comes with a test tone loaded USB that you use with your head unit. The DSP/amp has two gain setting phases. One for you to play the test tone and increase the HU volume until the DSP/amp beeps or a clip light activates, then back off. Then the DSP/amp auto-adjusts the DSP-to-amp gains to just before clipping while your speakers are connected for the loading.

 

Isn’t this the idea behind the kicker key? I haven’t used one yet but it seems to auto configure most of itself.

 

Well, not really (I have a Kicker KEY in my other car). You still set gains manually on the KEY - at least on the one that I have (the original one). It handles EQ and time-alignment automatically - and it may even level the channels automatically for you (can't remember), but even it doesn't set the actual gains for you (it has regular gain dials just like every other amp).

 

Did you use the Kicker pink noise for setting gains on your Key?

 

The easiest way to set amp gain is to use an amplifier that has an automatic clipping detector to prevent turning its gain too high.