TOPIC:

Can someone explain how the target relates to the all before? And how I should use the all before to set my levels.

It says the all before is normalized, I assume that's because we don't make measurements anywhere near 0db otherwise we would blow our speakers. So I understand that isn't just so we can see it and don't have a scroll down option on the screen.

I also understand it's using inverse filters so there's some boost in comparison to the signals relative output.

If I set my target or set the output lower than The highest peak and the lowest null db difference divided by 2 that should get me the amount of cut to use on output correct?

When I do that I finally get no clipping. However the output is so low how do I compensate?

Is it at that point okay to use the same amount of db that I used to cut Dirac and add it back in boost in my dac/dsp (for me 2x4HDs) and my dsp allows 12db boost at output. Would it be safe to say I could add that boost in dsp and succeed in restoring output with no clipping?

Thanks ,
Andrew

"All (before)" shows the single measurements you took.
"Target" lets you control how the average response should look like after equalization.

Please see manual:

Avg. spectrum (before)
The average of the measured magnitude responses. These plots are shown in light blue.

Avg. spectrum (after)
The predicted average magnitude response after correction. These plots are shown in green, and can be viewed only after filters have been generated with the Optimize button.

Target
The target curve – that is, the desired in-room magnitude response. This curve is user-adjustable so you can fine-tune it to best suit your speakers, room, and preferences

All (before)
All of the measured magnitude responses. These plots are shown dark blue.

All (after)
All of the predicted magnitude responses after correction. These plots are shown in dark green,
and can be viewed only after filters have been generated with the Optimize button.

Okay so how do we get the maximum amount of output signal out of Dirac without clipping?

Yeah we could say -10db should be enough but is it? What if we have a 22db difference in the "all before" or the measurement. Divide that by 2 and that's 11db. I figure it's going to make a inverse filter so 22db mirrored would be two 11db halves.

I'm just saying. How do we know by use of graphs that is correct?

Because the "all before" is normalized so it's db scale is already wrong. I mean I take my measurements at -35db but it shows on screen as part of it going above 0db..... so
I just want to be sure what we see in all before is actually the right magnitude. there's no crosshairs to click on parts of graph to make measurements that way just a little hand that shows the x,y data.

Anyway just want to maximize the signal as much as possible...and ask if the lowered output would allow me to safely boost the output on the dac respectively to restore lost signal.


Andrew

Hi, I think that for what you asking you need to connect the DDRC-xy output directly to your audio measurement device input.

Or, you could just listen for issues. Bear in mind that any form of EQ (including Dirac Live) tends to reduce available headroom. That's the way it is, and the reason the default gain is (IIRC) -10 dB.

denver8me4dinner wrote:

Okay so how do we get the maximum amount of output signal out of Dirac without clipping?

Depends on a) the max. input signal and b) how much DL has to boost to get to the intended target curve (boost is limited to 10dB though).

P.S. You seem to be confused what "All (before)" is showing. It is NOT a normalized view. It shows the acoustical response of each speaker at the mic positions. Please re-read my post and the manual.
P.P.S. Aligning speaker levels is part of DL's optimization process. In order to maximize SNR of the miniDSP device you should get the speaker levels in line in the "Output & level" tab of DLCT (use the input gain controls of your amp for that).

markus wrote:

denver8me4dinner wrote:

Okay so how do we get the maximum amount of output signal out of Dirac without clipping?

Depends on a) the max. input signal and b) how much DL has to boost to get to the intended target curve (boost is limited to 10dB though).

P.S. You seem to be confused what "All (before)" is showing. It is NOT a normalized view. It shows the acoustical response of each speaker at the mic positions. Please re-read my post and the manual.
P.P.S. Aligning speaker levels is part of DL's optimization process. In order to maximize SNR of the miniDSP device you should get the speaker levels in line in the "Output & level" tab of DLCT (use the input gain controls of your amp for that).

Oh gosh I think you may have solved it .

Boost is limited to 10db....well than that makes sence because I have as much as 22db differences on the horns.

On another forum I've been saying a good pre tune seems to be the only good way to get a good result but I haven't been practicing what I preach. I need to do some pre peq than run DL.

I have been getting clipping only in the 3k band exactly where horns have the worst response...

I've just been happy with what DL has done on it's own so I been lazy. Just the clipping I've recently noticed. So I wanted to get solid answers before I go and spend a day tuning .

So how does pre tuning with peq affect its phase correction, iirs seem to make a funky lookin phase when peq is applied will it know how to overcome that on its own. Does it compare phase to the test signal? If so I should be good right?

Much appreciated

Andrew

john.reekie wrote:

Hi, I think that for what you asking you need to connect the DDRC-xy output directly to your audio measurement device input.

Or, you could just listen for issues. Bear in mind that any form of EQ (including Dirac Live) tends to reduce available headroom. That's the way it is, and the reason the default gain is (IIRC) -10 dB.

John,
Not sure I'm following what you mean,
I have it hooked up as per the manual and use umik1

Are you suggesting I check for clipping with another program like rew and analysis that way

Andrew

denver8me4dinner wrote:

Oh gosh I think you may have solved it .

Boost is limited to 10db....well than that makes sence because I have as much as 22db differences on the horns.

On another forum I've been saying a good pre tune seems to be the only good way to get a good result but I haven't been practicing what I preach. I need to do some pre peq than run DL.

I have been getting clipping only in the 3k band exactly where horns have the worst response...

I've just been happy with what DL has done on it's own so I been lazy. Just the clipping I've recently noticed. So I wanted to get solid answers before I go and spend a day tuning .

So how does pre tuning with peq affect its phase correction, iirs seem to make a funky lookin phase when peq is applied will it know how to overcome that on its own. Does it compare phase to the test signal? If so I should be good right?

Much appreciated

Andrew

Speaker (crossover) development requires anechoic data. The approach is outlined here: www.grimmaudio.com/site/assets/files/1088/speakers.pdf

DL itself shouldn't produce any clipping. You're probably overdriving DL input or somewhere at the output stage. Use test signals and watch the meters.

markus wrote:

denver8me4dinner wrote:

Oh gosh I think you may have solved it .

Boost is limited to 10db....well than that makes sence because I have as much as 22db differences on the horns.

On another forum I've been saying a good pre tune seems to be the only good way to get a good result but I haven't been practicing what I preach. I need to do some pre peq than run DL.

I have been getting clipping only in the 3k band exactly where horns have the worst response...

I've just been happy with what DL has done on it's own so I been lazy. Just the clipping I've recently noticed. So I wanted to get solid answers before I go and spend a day tuning .

So how does pre tuning with peq affect its phase correction, iirs seem to make a funky lookin phase when peq is applied will it know how to overcome that on its own. Does it compare phase to the test signal? If so I should be good right?

Much appreciated

Andrew

Speaker (crossover) development requires anechoic data. The approach is outlined here: www.grimmaudio.com/site/assets/files/1088/speakers.pdf

DL itself shouldn't produce any clipping. You're probably overdriving DL input or somewhere at the output stage. Use test signals and watch the meters.

You were right. I have to get my measurements within 10db of 0db either +\- 10db. It's not clipping now. I super appreciate your help.

Yeah it's toslink in/out so it was the measurement. I used peq on my hds and got it sounding right...

So now I need to keep my hands off the vol knob when I want extra gain and find another way to get gain. I suppose I'll take the noise compromise and gain up amps

So..... I got the clipping fixed now. So now I need gain , and I really would like to know how much of the 12db on the HDs output I can boost safely. Why did minidsp put boost on the output of the hds ? Was it for these reasons and I'll be ok.

I guess another way I was trying to ask ,
If I'm now 10db quieter (on average) let's just say can I put 10 DB of boost in the 2 x 4 HD and not clip anything ? I don't own a scope so I don't know any other way to test for clipping .

Thank you

denver8me4dinner wrote:

I don't own a scope

...yet

Recommended read:
www.hometheatershack.com/forums/home-the...ent-your-system.html

markus wrote:

denver8me4dinner wrote:

I don't own a scope

...yet

Recommended read:
www.hometheatershack.com/forums/home-the...ent-your-system.html

Oh thank you good read indeed I will definitely try that today in fact

Andrew

So I got a much better gain setting. Had to crank up amps gains too scared to clip the outputs of 2x4HDs...

However I think I really want to understand exactly what Dirac is doing and the minidsp hardware. Together.

I noticed on the DDRC2x2 plugin by default it comes at -10db....I also noticed that that setting in the plug in is configuration specific. Meaning if I change it to -6db or -3db it only applies that change to the preset that is currently playing (1-4) .....

That article is very good about gains ....and I get it. However when I'm making filters it doesn't help me understand what's happening working DL and the ddrc box itself....

1. Is there separate input and output gains in DDRC22DI? One is vol knob and one is the attenuation?

2. How does drawing the target above/below 0db line in "filter making" affect the input or output? Meaning if one were to draw a line 10db under the 0db line would that eliminate the need for having the -10db setting? How does that 0db line relate to signal strength and boost/cut on the filter?

3. Is the "all before" even something to consider on where you draw your target? Or are they completely unrelated and you just simply draw the responce you want?

Thanks everyone,
Andrew

I think I finally found the actual source of my clipping.....okay I DID find the actual source

And this is a question maybe DevTeam or anyone well versed in electronics would be able to answer ...

So if my 4 amplifiers 3 are class D and one class AB , the one that is clipping is the AB
It's model is FOCAL FPS4160 ... it's gain range is 5v-250mv. This amp starts to distort if the the gain is turned anywhere over 3v setting. And it doesn't matter why that actual voltage going into the amp is. It has a distorted sounding output even at very low low listening levels.

This amplifier boasts a "direct signal" input
And a "soft clipping" output. It's a car amp and is designed to be used at 2-4ohms however I am using it 8ohms. I figured what difference does it make all speakers eventually go above 4ohms right....

So anyway I think it's a bad design or maybe I need a stronger signal going in. When I had my old source it put out a healthy 5v and I never heard this. Now that I'm down to 2v and -10db on processor I can't get any gain from just this amp. All my class D amps work just fine with gain up a bit.

So.... would anyone know how what soft clipping in a amps output means and how that would relate to my system? It's such a good sounding amp when gain is low and have more input signal going in. It's not a gain structure problem now it's a product compatibility issue.

Should I ditch this amp and go with another class D or is there a work around?

Help much appreciated
Andrew

I'm not sure how you adjust the sensitivity on the amp but you want to reduce the setting to under 2V and maybe lower. (i.e. reduce the voltage for full scale output == increase the sensitivity)

john.reekie wrote:

I'm not sure how you adjust the sensitivity on the amp but you want to reduce the setting to under 2V and maybe lower. (i.e. reduce the voltage for full scale output == increase the sensitivity)

Exactly right.

So I have been talking to different folks and apparently my amps are "high end" and only accept a high voltage line. Which is dumb they shouldn't have made the amp like this.

Anyway a nice gent informed me that adding a line amp will get this amp do what I need with my equipment. Which is opposite of what I read on home theatre shack thread. He was saying something about how a dac runs at lower volt and any boost in mini might not be quite enough and instead of stressing the mini to its peak and get distortions use a line amp. Makes sence only because this amp requires it. I have never had to do this in my life as a hobbyist. I've always found a gain structure that is good with whatever I've use home pro car don't matter. And my other amps work just fine with this setup of different brand....so it seems this amp issue is rear it big head now n needs addressed

So we'll try that and hopefully doesn't add a bunch of noise. It's 114db sens seems ok and they hand make it for my dsp and amp so.....we'll see

Andrew