Can Raspberry Pi be optimized for sound quality?
JP said:
Click to expand…
The SMD ceramic caps on the RPi itself and probably the HiFiBerry DAC+ Pro too are said to be of the least expensive variety. If I’m wrong about that I’d love to be enlightened, but on various fora this has been said including one’s in which Daniel has posted at times, and I’ve never seen him refute it.
In other words, these MLCCs were chosen for cost, not for their superior resonant characteristics and how that might affect clock performance. This has been referred to as the singing capacitor effect, in which mechanical resonance can generate noise on the voltage rails, thereby degrading clock performance, the piezoelectric effect.
So my original question was regarding these timing errors that appear to be the basis of what you were listening to in those delta waveforms, do you think that’s purely a clock quality thing, or could other factors such as the above be at play, especially given the Pi’s tiny amount of available real estate that limits placement options of various parts on the board, and the restricted BOM necessary to hit the price point they are.
The RPi is intended for underprivileged school children in places where they have no hope of ever affording most other computers let alone the electricity needed to power them. That price point rules out audio grade capacitors.
Yes the Pi’s GND is noisy, possibly due to the switching regulator on the board, or perhaps other factors too, as previously stated I’m not an EE, but a HAT such as the HiFiBerry DAC+ Pro sounds better with that GND severed in my listening tests, using an isolator board such as those from Allo or IanCanada.
I tried both the Collybia MamboBerry LS DAC+, and the aforementioned HiFiBerry DAC+ Pro with the isolator and thought the larger improvement was with the latter, possibly due to it’s master clocking arrangement.
I make no proclamations here, just throwing it out there as it may or may not relate to the small timing errors you mentioned.
The SMD ceramic caps on the RPi itself and probably the HiFiBerry DAC+ Pro too are said to be of the least expensive variety. If I’m wrong about that I’d love to be enlightened, but on various fora this has been said including one’s in which Daniel has posted at times, and I’ve never seen him refute it.In other words, these MLCCs were chosen for cost, not for their superior resonant characteristics and how that might affect clock performance. This has been referred to as the singing capacitor effect, in which mechanical resonance can generate noise on the voltage rails, thereby degrading clock performance, the piezoelectric effect.So my original question was regarding these timing errors that appear to be the basis of what you were listening to in those delta waveforms, do you think that’s purely a clock quality thing, or could other factors such as the above be at play, especially given the Pi’s tiny amount of available real estate that limits placement options of various parts on the board, and the restricted BOM necessary to hit the price point they are.The RPi is intended for underprivileged school children in places where they have no hope of ever affording most other computers let alone the electricity needed to power them. That price point rules out audio grade capacitors.Yes the Pi’s GND is noisy, possibly due to the switching regulator on the board, or perhaps other factors too, as previously stated I’m not an EE, but a HAT such as the HiFiBerry DAC+ Pro sounds better with that GND severed in my listening tests, using an isolator board such as those from Allo or IanCanada.I tried both the Collybia MamboBerry LS DAC+, and the aforementioned HiFiBerry DAC+ Pro with the isolator and thought the larger improvement was with the latter, possibly due to it’s master clocking arrangement.I make no proclamations here, just throwing it out there as it may or may not relate to the small timing errors you mentioned.
