how to connect AD1896 to improve jitter rejection?
Is input 44.1khz and output 192khz a good idea? or the ratio should be integer?
Is input 44.1khz and output 192khz a good idea? or the ratio should be integer?
AD1896 will attenuate jitter. Big time. Hook it up like it says in the datasheet. Ratio makes no difference, and no matter how hard you try you won't get an integer anyway. The AD1896 (and other ASRCs) use a rate estimator to determine the average incoming data rate, and use the master clock to determine the desired output rate. Then they apply digital filters to complete the conversion. The output should have the same jitter as your master clock.
I don't see any reason to not go 44.1 to 192. Your DAC is probably going to oversample this to an ever higher rate anyway.
I don't see any reason to not go 44.1 to 192. Your DAC is probably going to oversample this to an ever higher rate anyway.
yejun,yejun said:
the oversampling is necessary for the digital filter, not anyway.
I suggest you if you use a sigma-delta DAC and no more than 44.1-48kHz source, to resample it to ONLY 48 kHz with the 1896 chip. This will leave more headroom for the digital filter, integrated in the sigma-delta, and it will use 8xoversampling /for example, see AD1853, AD1955 datasheet!/ If u resample to 192kHz, the DAC will be left with only 2xOversampling and the frequency and phase responce becomes uglier
😉
So you can see it is significant what kind of source and DAC you will use. Many people here hate Sigma-deltas, but they are a lot Cheaper and good ones have very good Digital filters inside and characteristics also.
Regards
Kaloyan
Hi All!
By the way, ad1896 wonderfully works in a synchronous mode, when it is equivalent to the ordinary digital filter with integer factor of oversampling.
From it is possible to receive 4Xoversampling for 44.1 or 48kHz.
In this mode it is comparable with digital filters SM5842/5847 and subjectively sounds better than SM.
Probably it explains in short length of the digital filter AD1896 (only 64 taps).
In an asynchronous mode, it certainly suppresses a jitter, but sounds considerably worse than SM from NPC.
But to apply a synchronous mode it is possible only in case of presence of separate clockline for synchronization DAC with transport.
This mode can be used as with R-2R DAC, and as with the delta sigma DAC as AD1955 and PCM1738 (i.e. there, where is possible to disconnect the built-in digital filter, in the other case, with ordinary D/S DAC here is no sense in it, good anything is not present in a double digital filtration, where one of them also with fractional factor).
For what to use it in a synchronous mode with R-2R multibit DACs (or inputs 1955/1738 in external digital filter mode), it is necessary to AD1896 on an output to put the shift register. It is possible to carry out synchronous reclocking at desire.
The main generator and divider is necessary for establishing in the DAC module, ad1896 to establish in slave mode (in/out) completely, and requires submission of a grid of frequencies from a divider.
From the same generator the transport / decoder should be synchronized.
Regards
Halyavshick
By the way, ad1896 wonderfully works in a synchronous mode, when it is equivalent to the ordinary digital filter with integer factor of oversampling.
From it is possible to receive 4Xoversampling for 44.1 or 48kHz.
In this mode it is comparable with digital filters SM5842/5847 and subjectively sounds better than SM.
Probably it explains in short length of the digital filter AD1896 (only 64 taps).
In an asynchronous mode, it certainly suppresses a jitter, but sounds considerably worse than SM from NPC.
But to apply a synchronous mode it is possible only in case of presence of separate clockline for synchronization DAC with transport.
This mode can be used as with R-2R DAC, and as with the delta sigma DAC as AD1955 and PCM1738 (i.e. there, where is possible to disconnect the built-in digital filter, in the other case, with ordinary D/S DAC here is no sense in it, good anything is not present in a double digital filtration, where one of them also with fractional factor).
For what to use it in a synchronous mode with R-2R multibit DACs (or inputs 1955/1738 in external digital filter mode), it is necessary to AD1896 on an output to put the shift register. It is possible to carry out synchronous reclocking at desire.
The main generator and divider is necessary for establishing in the DAC module, ad1896 to establish in slave mode (in/out) completely, and requires submission of a grid of frequencies from a divider.
From the same generator the transport / decoder should be synchronized.
Regards
Halyavshick
- Status
- Not open for further replies.