Hi,
I would like to upgrade the ADC on my ADAU1701, so I've been looking for the ADC to I2S that are available in Aliexpress, and found these:
THD+N (dB) SNR (dB) DR (dB)
PCM1808 -93 99 99 96 SE
ADAU1701 -83 100 SE
AD1938 -94 107 107 SE
Cirrus Logic CS5343 -92 98 98 96 SE
AKM 5720 -94 102 102 SE
PCM1802 -96 105 105 SE
The PCM1802 is one of the best (spec wise) and it's also very cheap...
My question is, should I capture @96kHZ of program the ADC to capture @48kHz.
The analog source it's AVR's preamplified outputs, working @48kHz but once the sound becomes "analog" again, shouldn't be better to convert it at a higher sample rate (96kHz vs 48kHz)?? If the capture is made @96kHz I would change the ADAU1701 schemes to work at the same sample rate before output to the DACs.
I am understanding these things correctly?
Thanks in advance.
I would like to upgrade the ADC on my ADAU1701, so I've been looking for the ADC to I2S that are available in Aliexpress, and found these:
THD+N (dB) SNR (dB) DR (dB)
PCM1808 -93 99 99 96 SE
ADAU1701 -83 100 SE
AD1938 -94 107 107 SE
Cirrus Logic CS5343 -92 98 98 96 SE
AKM 5720 -94 102 102 SE
PCM1802 -96 105 105 SE
The PCM1802 is one of the best (spec wise) and it's also very cheap...
My question is, should I capture @96kHZ of program the ADC to capture @48kHz.
The analog source it's AVR's preamplified outputs, working @48kHz but once the sound becomes "analog" again, shouldn't be better to convert it at a higher sample rate (96kHz vs 48kHz)?? If the capture is made @96kHz I would change the ADAU1701 schemes to work at the same sample rate before output to the DACs.
I am understanding these things correctly?
Thanks in advance.
I think your question isn't clear enough.
You start with a DAC working at 48 kHz sample rate, then you go into an ADC at either 48 or 96 kHz, then through a DSP and then into a DAC at 48 or 96 kHz? Why are the first two conversions needed?
You start with a DAC working at 48 kHz sample rate, then you go into an ADC at either 48 or 96 kHz, then through a DSP and then into a DAC at 48 or 96 kHz? Why are the first two conversions needed?
Sorry If I've not been clear...
I want to upgrade the ADC on my DSP.
My question is:
Is there any SQ advantage in using my ADC @96kHz vs @48kHz (both 24bits)?
Things to consider:
-. All the DSP programming /scheme in SigmaStudio will be done @96kHz
-. The "analog" signal I am feeding to my DSP is coming from my AVR's preamp.
Cheers.
I want to upgrade the ADC on my DSP.
My question is:
Is there any SQ advantage in using my ADC @96kHz vs @48kHz (both 24bits)?
Things to consider:
-. All the DSP programming /scheme in SigmaStudio will be done @96kHz
-. The "analog" signal I am feeding to my DSP is coming from my AVR's preamp.
Cheers.
It's unclear to me what you mean by: "The analog source it's AVR's preamplified outputs, working @48kHz". Does that mean it's actually a digital source converted to analogue that gets converted back to digital? If not, what does that 48 kHz refer to?
I don't know what AVR means in this context, maybe that's why I don't understand it.
I don't know what AVR means in this context, maybe that's why I don't understand it.
The "analog" source that feeds the DSP's ADC is coming from my AVRs preamplified outputs.
The AVR does all the digital processing and outputs to the speakers or the pre amplified outputs.
ANALOG AUDIO SIGNAL >> ANALOG to DIGITAL converter (ADC) >> DSP PROCESS >> DAC >> AMPLIFIER >> SPEAKER
The Analog Audio Signal is the NON DIGITAL source.
The AVR does all the digital processing and outputs to the speakers or the pre amplified outputs.
ANALOG AUDIO SIGNAL >> ANALOG to DIGITAL converter (ADC) >> DSP PROCESS >> DAC >> AMPLIFIER >> SPEAKER
The Analog Audio Signal is the NON DIGITAL source.
I still don't get it, maybe someone else does.
In any case, when converting a real analogue source, 96 kHz sample rate is best in theory. Whether the difference is audible to humans is a different matter, one that usually triggers flame wars on forums like this.
When converting a DACs output signal back to digital, you have to look out with clock harmonics getting converted to audible whistles. It could be that a sample rate that is no multiple of the DAC sample rate then works best, unless you can synchronize the clocks so they are exact multiples.
In any case, when converting a real analogue source, 96 kHz sample rate is best in theory. Whether the difference is audible to humans is a different matter, one that usually triggers flame wars on forums like this.
When converting a DACs output signal back to digital, you have to look out with clock harmonics getting converted to audible whistles. It could be that a sample rate that is no multiple of the DAC sample rate then works best, unless you can synchronize the clocks so they are exact multiples.
Last edited:
In order to exchange your ADC your new signal chain must look like: (you need a digital input):
ANALOG AUDIO SIGNAL >> new external ANALOG to DIGITAL converter (ADC) >> digital input on AVR >> DSP PROCESS >> DAC >> AMPLIFIER >> SPEAKER
//
ANALOG AUDIO SIGNAL >> new external ANALOG to DIGITAL converter (ADC) >> digital input on AVR >> DSP PROCESS >> DAC >> AMPLIFIER >> SPEAKER
//
Maybe I'm also confused. Please allocate the different functions in your flow to the different boxes that you have... and show the cables between... and what type of interface (ana/dig/Fs)
//
//
No.
My (most if not all) AVR does not have multichannel digital input.
I want to use my AVR pre amplified outputs and feed my DSP for proccessing.
For doing that I need to reconvert the signal to digital, process it and output with my DSP's DAC.
It's the same as if I would be using active speakers with my AVR. Most active speaker have integrated ADC --> DSP --> DAC --> AMP.
It's not that hard, I may be missing something to explain?
My (most if not all) AVR does not have multichannel digital input.
I want to use my AVR pre amplified outputs and feed my DSP for proccessing.
For doing that I need to reconvert the signal to digital, process it and output with my DSP's DAC.
It's the same as if I would be using active speakers with my AVR. Most active speaker have integrated ADC --> DSP --> DAC --> AMP.
It's not that hard, I may be missing something to explain?
You have an AVR/receiver/multichannel processor, which serves the purpose of processing a digital signal, i.e. a bluray movie.
the avr process the movie and (in the end) outputs to the RCA preamplified outputs.
now, you want to input that processed output to a DSP for EQ, active XO, etc.
How do you do it? you need to input that signal to the DSP (thru an ADC).
ADAU1701 have 2 stereo ADC which are low performers. My idea is to replace them with 2 better ones ADC to i2S and feed the ADAU1701.
OK, so you don't have a useable digital output on that first box (AVR) and as a result, you end up re-digitizing a signal that was converted from digital to analogue inside that first box.
It's presumably not possible to synchronize the ADC clock with the clock of the DAC inside the AVR, so you may get whistles due to the 2^n-th harmonic of the DAC clock aliasing to the audible range. I'd go for some ADC clock frequency that's greater than 48 kHz and no multiple of 48 kHz, maybe 88.2 kHz.
It's presumably not possible to synchronize the ADC clock with the clock of the DAC inside the AVR, so you may get whistles due to the 2^n-th harmonic of the DAC clock aliasing to the audible range. I'd go for some ADC clock frequency that's greater than 48 kHz and no multiple of 48 kHz, maybe 88.2 kHz.
Correct, it's not possible not possible to synchronize the ADC clock with the clock of the DAC inside the AVR.
So, this is also true when I use an active speaker with RCA/analog input?
For example, Kali LP6 have RCA input which feeds it's DSP. The ADC and DSP processing is probably made @48kHz. So If I am listen 48kHz music, the sound will be degraded by the fact that the active speaker use the same sampling rate?
So, this is also true when I use an active speaker with RCA/analog input?
For example, Kali LP6 have RCA input which feeds it's DSP. The ADC and DSP processing is probably made @48kHz. So If I am listen 48kHz music, the sound will be degraded by the fact that the active speaker use the same sampling rate?
It depends on the quality of the reconstruction filter of the DAC and the anti-aliasing filter of the ADC. I have heard clearly audible whistles with a computer sound card DAC driving a dirt cheap ADC. The mitigation was to insert an extra low-pass between them.
I mean a tone at a kilohertz or two. The frequency changed when I touched a pin of the ADC's crystal.
The ADC was a sigma-delta ADC running at 6.144 MHz followed by a digital decimation chain. The decimation chain suppresses any aliases except those caused by input signals close to multiples of the sigma-delta sample rate, so multiples of 6.144 MHz. There should have been an external analogue filter to suppress those, but the manufacturer had decided to cut costs by leaving it out.
The DAC was probably a sigma-delta DAC that also ran at some power-of-two multiple of 48 kHz. As its clock was not synchronized to the ADC's clock, their frequencies were close but not exactly the same.
For example, suppose there was some clock residue at 12.2886 MHz at the DAC output while the ADC's sigma-delta modulator actually ran at 6.1438 MHz. (I don't know if those numbers are correct, but they illustrate the issue.) You then get an alias at 12.2886 MHz - (2 * 6.1438 MHz) = 1 kHz. The extra low-pass filter had to suppress megahertz frequencies to get rid of a tone of the order of a kilohertz.
The ADC was a sigma-delta ADC running at 6.144 MHz followed by a digital decimation chain. The decimation chain suppresses any aliases except those caused by input signals close to multiples of the sigma-delta sample rate, so multiples of 6.144 MHz. There should have been an external analogue filter to suppress those, but the manufacturer had decided to cut costs by leaving it out.
The DAC was probably a sigma-delta DAC that also ran at some power-of-two multiple of 48 kHz. As its clock was not synchronized to the ADC's clock, their frequencies were close but not exactly the same.
For example, suppose there was some clock residue at 12.2886 MHz at the DAC output while the ADC's sigma-delta modulator actually ran at 6.1438 MHz. (I don't know if those numbers are correct, but they illustrate the issue.) You then get an alias at 12.2886 MHz - (2 * 6.1438 MHz) = 1 kHz. The extra low-pass filter had to suppress megahertz frequencies to get rid of a tone of the order of a kilohertz.
Last edited: