Forum Rule: Always post complete source code & details to reproduce any issue!
Results 1 to 10 of 10

Thread: Reduce noise on Guitar Input

  1. #1

    Reduce noise on Guitar Input

    Hi there!
    I'm trying to build a FAUST-based guitar amp profiler (Kemper-like) on the teensy 4.0, currently using only external hardware and the standard audioshield (no custom input circuits). The signal path giving me the best results so far is:
    Guitar - DI-BOX - Mixer with Preamp - Audioshield-LineIn
    I use a balanced signal between the mixer and the shields line-in (substraction is done in FAUST), that reduces the noise to acceptable levels for everything up to AC/DC-like distortion, but it gets really bad from there on. The same FAUST-program, as well as the teensy using USB-audio input, runs super-silent, so I assume the audioshield is the noisy part.
    Since I'm not into electronics so much, are there any standard prebuild boards, shields or ICs out there I could try? I only know of blackaddr's TGA Pro, but that doesn't ship to Germany.
    Or is it an option to use an external ADC/DAC and using the SPDIF ins and outs? Would that affect the latency?
    Looking forward to any replies

  2. #2
    Senior Member
    Join Date
    Jul 2020
    Posts
    1,222
    Could it just be clipping that you are experiencing? The Audio shield inputs have a limited voltage range they can tolerate
    (somewhat less than 3.3V peak-2-peak) before they clip, and once above 3.3V peak-2-peak there's a risk of damage to the
    shield if no current-limiting input resistors are employed, since the input signal goes straight into the chip (via DC-blocking
    capacitors).

  3. #3
    Thanks for you reply, but I'm pretty sure there's no clipping involved (I have a clipping indicator in the sketch that lights the LED when the input is saturated). It is actually more of a white noise when there should be silence. I'll share the sketch and record a sample tonight

  4. #4
    I think my old mixer is the noisy one. While recording the sample I wanted to create a reference recording through the same route (DI - Mixer) into my audio interface, which gave me the very same noise... Since I was partly successful reducing the noise on the teensy (by using the input path and balanced inputs) I didn't reference it again before posting. I'll try it again with a better mixer and post back the results!
    Thanks for the assistance and sorry for taking your time!

  5. #5
    Ok, I got a better mixer but the problem persists. I uploaded the sketch and some samples, using the same faust program through the teensy and as a jack-client (same input path):
    https://github.com/robinrosenberger/noisyteensy

  6. #6
    Senior Member
    Join Date
    May 2019
    Posts
    133
    Noise (hiss) or EMI (buzzes) ?

    Is it still there when you ground the input of the mixer?

  7. #7
    @tscharma: Sounds like noise to me, not buzz (samples on the github link). I'm sorry, but I dont really know what you mean by "ground the input of the mixer".
    The noise is actually still present when I completely disconnect the audioshields input, so I assume it is the ADC's noise?

  8. #8
    Senior Member
    Join Date
    May 2019
    Posts
    133
    yes, if the noise is still present with the ADC grounded (meaning, the input of the ADC shorted to GND, so no signal can reach the ADC, no noise and no EMI), all the noise that you still get is inherent from your data aquisition system, and unrelated to your sensor of signal conditioning. Meaning, the guitar, preamp , mixer etc etc is not the problem.

  9. #9
    Senior Member
    Join Date
    Jul 2020
    Posts
    1,222
    An audio device with nothing plugged in is going to see the thermal noise from its internal load resistance, which can be considerable if this is
    the standard value (50k). With a low-impedance device plugged in, or the input shorted, the device sees no/little thernal noise from this resistance.

    Thermal (voltage) noise power scales with the resistance value, so 50k has 30dB more noise than 50 ohms.

  10. #10
    Thank you both for helping with the analysis, I actually never tried connecting the inputs to ground until today. There is still an audible noise with the inputs grounded, and since there's a lot of compression happening in the tube emulation, even a little noise is a big problem for my application. So I think I'll have to look for alternatives to the audioshield. Do any of you know any besides TGA?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •