What is the reason for SNR and thereby dynamic range to degrade in the low frequency range ?
I was just testing a soundcard using RMAA and the charts show that the snr and dynamic range gradually degrade in the low frequency range. It starts off at -108db at 20Khz and by 100hz it is down to -93db, -90db @ 30hz and -87db @ 20hz.
What does this generally indicate ? Is this something to do with the "digital" part of the system (jitter, etc..) or "analog" (psu decoupling, filtering, bad caps, etc..) ?
I was just testing a soundcard using RMAA and the charts show that the snr and dynamic range gradually degrade in the low frequency range. It starts off at -108db at 20Khz and by 100hz it is down to -93db, -90db @ 30hz and -87db @ 20hz.
What does this generally indicate ? Is this something to do with the "digital" part of the system (jitter, etc..) or "analog" (psu decoupling, filtering, bad caps, etc..) ?
the cable is just a short (6") shielded stereo patch cord going from line-out to line-in, and its not too close to the mains cable.
Why do I faintly remember that I observed the same thing when I was testing an old sony discman which was running off batteries. There was a sudden drop in SNR/DR below ~1Khz.
Why do I faintly remember that I observed the same thing when I was testing an old sony discman which was running off batteries. There was a sudden drop in SNR/DR below ~1Khz.
Check freq response, high pass filters in dac and adc multiply and cut signal at low freq
Op amps can have 1/f noise corners higher than 1 KHz, fet and cmos inputs usually have higher 1/f than bipolar
bipolar op amp input current noise can have a higher 1/f - too high coupling/feedback impedance can make input current noise prominent
High freq noise can alias down to audio band, so noise at ~ n*sampling frequency translates to low freq noise, oversamping/delta-sigma adcs move this aliasing up to their oversampling freq
On board soundcards can be pretty ugly
Op amps can have 1/f noise corners higher than 1 KHz, fet and cmos inputs usually have higher 1/f than bipolar
bipolar op amp input current noise can have a higher 1/f - too high coupling/feedback impedance can make input current noise prominent
High freq noise can alias down to audio band, so noise at ~ n*sampling frequency translates to low freq noise, oversamping/delta-sigma adcs move this aliasing up to their oversampling freq
On board soundcards can be pretty ugly
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.