Welcome, Guest
Username: Password: Remember me
All tutorials related to equalization of a system with miniDSP platforms.

TOPIC: Limitation of signal levels to avoid clipping

Limitation of signal levels to avoid clipping 1 year 3 months ago #20739

  • samoht
  • samoht's Avatar
  • Offline
  • Expert Boarder
  • Posts: 92
  • Thank you received: 5
  • Karma: 1
Hi all,

I'm using a miniDSP 4x10 for crossovers and PEQing my 3-way speakers. In this respect the frequency response of the subwoofer is extended by a Linkwitz Transform, which means (like the tutorial shows) that the signal below the original rolloff frequency is lifted significantly.
My question is: With the signal input level and output level set to 0 dBFS and the enhanced signal from the Linkwitz Transform being far above 0 dB at very low frequencies, does this signals create distortion and/or clipping of the miniDSP?
Or doI have to reduce the input signal by the same level which could be added by PEQs or others filters at a later stage in the DSP process?

Another practical problem: According to the datasheet my tweeter shows a heavy resonance peak at 26 kHz, which is out of range of the miniDSP processing frequency band. Is it possible to equalize this peak by using a biquad? The spreadsheet allows calculations of filters above 20 kHz.

Any advices are wellcome.

Best regards

Thomas
The administrator has disabled public write access.

Limitation of signal levels to avoid clipping 1 year 3 months ago #20747

  • john.reekie
  • john.reekie's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 2286
  • Thank you received: 859
  • Karma: 99
Hi Thomas, my understanding is that there's enough headroom in the DSP internally that clipping is unlikely. Clipping will occur if the signal exceeds 0dB FS at the output though. I suppose to be sure you could feed in a 0 dB sweep in and check the output.

[Edit] I've just noticed that you said "With the signal input level and output level set to 0 dBFS " - in that case yes, you will get clipping if the input signal exceeds -(LT boost) dB FS at very low frequencies.

On your second question, I've equalized an ultrasonic peak with the spreadsheet, so yes it works :) Of course people will tell you it's unnecessary, but... ;)

-> www.hifizine.com/2014/09/mini-convertible-crossover/
I am not miniDSP support.

"You must ask the right questions." - Dr. Alfred Lanning's hologram.
-> Have you read the User Manual??
-> Have you drawn and posted a diagram?
-> Have you posted a screenshot?
-> Have you posted your config file?
Last Edit: 1 year 3 months ago by john.reekie.
The administrator has disabled public write access.
The following user(s) said Thank You: samoht

Limitation of signal levels to avoid clipping 1 year 3 months ago #20750

  • samoht
  • samoht's Avatar
  • Offline
  • Expert Boarder
  • Posts: 92
  • Thank you received: 5
  • Karma: 1
Hi John,

Thanks for your quick answers.
I will do some testing with the miniDSP at different signal levels. In the meantime i had a look to the datasheet of the DSP and found a headroom of 4 bit on top of the 24-bit bus resulting in an internal 28-bit bus. However the 28-bit bus will be cut back to 24 bit just before the D/A converter by a clipping circuit. In my opinion this cutoff leads to distortion if the signal is broader than 24 bits.

Thank's also for the answer to my second question. Interestingly i'm using the same tweeter.
Could you tell me your parameters as ARTA does not measure above 20 kHz?

Best regards

Thomas
Last Edit: 1 year 3 months ago by samoht.
The administrator has disabled public write access.

Limitation of signal levels to avoid clipping 1 year 3 months ago #20753

  • john.reekie
  • john.reekie's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 2286
  • Thank you received: 859
  • Karma: 99
4 bits is about 24 dB headroom. But yes, the signal must not exceed 0 dB FS at the output stage. To put it another way, the (input signal level in dB FS) + (gain in all DSP processing) must be no more than 0 dB. If everything (signal, input gain, output gain) is at 0 dB and you have gain in the filters, then you will get clipping, no doubt. Therefore, one of those three must be reduced.

I used the PWR-ICE but the parameters would be the same for the 4x10 I assume. Here they are. This is for the tweeter I have, it is possible there is variation between units/production batches.

biquad1,
b0=1.08039810944833,
b1=-0.216085259387282,
b2=0.758040243457985,
a1=0.216085259387282,
a2=-0.838438352906311

biquad1,
b0=0.744849906488439,
b1=0.23827938867872,
b2=0.717924863817948,
a1=-0.23827938867872,
a2=-0.462774770306387
I am not miniDSP support.

"You must ask the right questions." - Dr. Alfred Lanning's hologram.
-> Have you read the User Manual??
-> Have you drawn and posted a diagram?
-> Have you posted a screenshot?
-> Have you posted your config file?
The administrator has disabled public write access.

Limitation of signal levels to avoid clipping 1 year 3 months ago #20754

  • john.reekie
  • john.reekie's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 2286
  • Thank you received: 859
  • Karma: 99
PS are you sure that ARTA does not measure above 20 kHz? I seem to recall setting it up one time to run at 96 kHz sample rate, so it should be able to sweep past 20k.
I am not miniDSP support.

"You must ask the right questions." - Dr. Alfred Lanning's hologram.
-> Have you read the User Manual??
-> Have you drawn and posted a diagram?
-> Have you posted a screenshot?
-> Have you posted your config file?
The administrator has disabled public write access.

Limitation of signal levels to avoid clipping 1 year 3 months ago #20755

  • samoht
  • samoht's Avatar
  • Offline
  • Expert Boarder
  • Posts: 92
  • Thank you received: 5
  • Karma: 1
Hi John,

Thanks for your answers.
I will check whether ARTA can go above 20 kHz.

Are your biquad parameters based on 96 kHz sammpling rate?

Best regards

Thomas
The administrator has disabled public write access.

Limitation of signal levels to avoid clipping 1 year 3 months ago #20756

  • john.reekie
  • john.reekie's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 2286
  • Thank you received: 859
  • Karma: 99
Yes, 96 kHz :)
I am not miniDSP support.

"You must ask the right questions." - Dr. Alfred Lanning's hologram.
-> Have you read the User Manual??
-> Have you drawn and posted a diagram?
-> Have you posted a screenshot?
-> Have you posted your config file?
The administrator has disabled public write access.

Limitation of signal levels to avoid clipping 1 year 3 months ago #20762

  • dreite
  • dreite's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 842
  • Thank you received: 408
  • Karma: 78
ARTA will indeed measure to above 20khz. This is effectively determined by your hardware capability. I routinely have ARTA set for a 96khz sampling rate and can measure to 48khz.......if I want to. :)

I'm not sure I understand the requirement to notch filter something at 26khz. You'd need to have content at that frequency in order to excite that resonance. :)

Dave.
The administrator has disabled public write access.
The following user(s) said Thank You: devteam

Limitation of signal levels to avoid clipping 1 year 3 months ago #20764

  • samoht
  • samoht's Avatar
  • Offline
  • Expert Boarder
  • Posts: 92
  • Thank you received: 5
  • Karma: 1
Hi Dave,

You may be right that ARTA itself can measure above 20 kHz. But the rest of my equipment is limited to 20 kHz.

Concerning the notch filter at 26 KHz you are in principle right, but i can image that 2nd or 3rd harmonics can stimulate the resonance frequency, too.

Best regards

Thomas
The administrator has disabled public write access.
Time to create page: 0.135 seconds