I designed a conversion board between DSDit SRC board and Ian's STATIONPI, so that the I2S output of DSDit SRC board and beaglebone black can be passed DSD signal to Ian's reclock via STATIONPI.
I will share my experience after I finish the test.
For DSDit SRC board, you can see the link below.
https://www.diyaudio.com/forums/dig...simple-dsd-src-beaglebone-20.html#post6706417
Upsampling will be good, you will know after trying an excellent upsampling algorithm.
Using this conversion board, in addition to using DSD Upscaling to obtain better sound quality, you can also use BBB with better I2S design, and the use is limited by the defects of Raspberry Pi.
The upsampling process does not "increase" the original music signal. Even if the sampling frequency reaches 192kHz, the bandwidth of the music signal can still only reach 20kHz.
Since there is no "increased" bandwidth, why does Upsampling sound better than the standard 44.1kHz digital signal? The difference is that the digital filtering will be affected before and after Upsampling. When we use digital upscaling to up to 192kHz, the center point of digital filtering is 96kHz, which is "the frequency that human ears can hear", and then the reference point of analog filtering will be far away from 96kHz.
Upsampled digital signal is by DA conversion, and in the "time range", it can more faithfully approach the original signal. In other words, the phase shift can be greatly reduced after upscaling, which is why the sound quality of digital upscaling is better than that of standard CD signals.
The built-in digital filter of the DAC will of course be upscaled and then processed. However, the limited performance of the DAC makes it impossible for its Upsampling algorithm to be better than the SRC4192 or AK4137, and it is even less likely to be better than the integrated SRC4192 and AK4137.
I will share my experience after I finish the test.
For DSDit SRC board, you can see the link below.
https://www.diyaudio.com/forums/dig...simple-dsd-src-beaglebone-20.html#post6706417
Upsampling will be good, you will know after trying an excellent upsampling algorithm.
Using this conversion board, in addition to using DSD Upscaling to obtain better sound quality, you can also use BBB with better I2S design, and the use is limited by the defects of Raspberry Pi.
The upsampling process does not "increase" the original music signal. Even if the sampling frequency reaches 192kHz, the bandwidth of the music signal can still only reach 20kHz.
Since there is no "increased" bandwidth, why does Upsampling sound better than the standard 44.1kHz digital signal? The difference is that the digital filtering will be affected before and after Upsampling. When we use digital upscaling to up to 192kHz, the center point of digital filtering is 96kHz, which is "the frequency that human ears can hear", and then the reference point of analog filtering will be far away from 96kHz.
Upsampled digital signal is by DA conversion, and in the "time range", it can more faithfully approach the original signal. In other words, the phase shift can be greatly reduced after upscaling, which is why the sound quality of digital upscaling is better than that of standard CD signals.
The built-in digital filter of the DAC will of course be upscaled and then processed. However, the limited performance of the DAC makes it impossible for its Upsampling algorithm to be better than the SRC4192 or AK4137, and it is even less likely to be better than the integrated SRC4192 and AK4137.
Last edited:
An externally hosted image should be here but it was not working when we last tested it.
Hi ! Ian
Can this circuit be used for SinPi ?
NH47M47LA(For Audio equipment)/High Precision Oscillator (OCXO / Twin-OCXO)/NDK
Because, if it isn't a sine clock (and especially in this case, since it specifically says in the image and datasheet that it an drive cmos directly), it doesnt need conversion?
Hello Ian,
You see there is already some serious interest for your new baby. And like i wrote you before i am interested too.
Especially if you wanna give the Sinepi or the two Italian boards a separate power supply there MUST be more than just a few solder connections to interconnect the boards because there are two SMA connectors that need a little '' torque '' And what makes it more tricky is that these cables go outside and could meet some challenges in real life if they end up on a typical wooden board with circuit boards all over.
Greetings, Eduard
You see there is already some serious interest for your new baby. And like i wrote you before i am interested too.
Especially if you wanna give the Sinepi or the two Italian boards a separate power supply there MUST be more than just a few solder connections to interconnect the boards because there are two SMA connectors that need a little '' torque '' And what makes it more tricky is that these cables go outside and could meet some challenges in real life if they end up on a typical wooden board with circuit boards all over.
Greetings, Eduard
That's correct.
"Bipolar Driver output that can drive CMOS-IC directly"
The output should be square wave based on the datasheet.
However, a SinePi can take both sine input and square input so there will be no any problem.
Square input could have even less additive jitter because of the higher slew rate under the same voltage rail noise. But have to confirm.
Ian
Last edited:
Andrea on FiFoPi and Power supplies
Some more review on the FiFoPi with Andrea Clock....
the-clock-reviews-4/
.
Some more review on the FiFoPi with Andrea Clock....
the-clock-reviews-4/
.
@dddac
Really appreciate for the sharing. Very interesting and useful experiences. Good job.
I'm still work on SinePi.
Actually the sine to square wave converter used to be part of oscillator. It can be considered as a high gain RF amplifier. This part of circuit is very sensitive to the power supply noise and interference. So everything has to be very carefully designed not only the circuit architecture but also the PCB layout. I have to use 4 layers PCB with impendence controlled traces to achieve better performance.
More improvements will be made. I'm looking forward to your evaluation.
Regards,
Ian
Really appreciate for the sharing. Very interesting and useful experiences. Good job.
I'm still work on SinePi.
Actually the sine to square wave converter used to be part of oscillator. It can be considered as a high gain RF amplifier. This part of circuit is very sensitive to the power supply noise and interference. So everything has to be very carefully designed not only the circuit architecture but also the PCB layout. I have to use 4 layers PCB with impendence controlled traces to achieve better performance.
More improvements will be made. I'm looking forward to your evaluation.
Regards,
Ian
Andrea on FiFoPi and Power supply sound impressions
Thanks Doede
First for posting here so we can comment on perceived sound pleasure without getting flamed on poor Andrea's thread, and
It confirms precisely what I predicted, but really would never have known unless you made these tests. Given the number of supercaps required I think I'll leave that one alone.
Thanks for taking the time to do this and report.
Thanks Doede
First for posting here so we can comment on perceived sound pleasure without getting flamed on poor Andrea's thread, and
It confirms precisely what I predicted, but really would never have known unless you made these tests. Given the number of supercaps required I think I'll leave that one alone.
Thanks for taking the time to do this and report.
SinePi and the NewClassD Neutron Star2 Audio reference clock
Neutron Star2 has a main high quality sine wave output in SMB coaxial cable connector though the built in sine to square converter output is not good enough.
SinePi makes the NewClassD Neutron Star2 Audio reference clock working great for FifoPi and ReclockPi. As well as all other sine clocks.
SincePi and Neutron Star2 can have three different configurations
1.One Neutron Start2 since clock
2.Two Neutron Start2 since clocks
3.One Neutron Start2 since clock and one local XO clock
SinePi_NeutronStar2 by Ian, on Flickr
Ian
Neutron Star2 has a main high quality sine wave output in SMB coaxial cable connector though the built in sine to square converter output is not good enough.
SinePi makes the NewClassD Neutron Star2 Audio reference clock working great for FifoPi and ReclockPi. As well as all other sine clocks.
SincePi and Neutron Star2 can have three different configurations
1.One Neutron Start2 since clock
2.Two Neutron Start2 since clocks
3.One Neutron Start2 since clock and one local XO clock
SinePi_NeutronStar2 by Ian, on Flickr
Ian
Thanks Doede
First for posting here so we can comment on perceived sound pleasure without getting flamed on poor Andrea's thread, and
It confirms precisely what I predicted, but really would never have known unless you made these tests. Given the number of supercaps required I think I'll leave that one alone.
Thanks for taking the time to do this and report.
Thanks for the feedback, yes my pleasure as I just like the hobby - I hope the flamethrowers will not hop over the fence and show up here....
I agree the Supercaps is not a practicable solution for everyone. May be when you build a larger chassis with only clocks and power supply inside. But as said, I consider it only as a small dot on the i .... but again, it is hobby, so....
@dddac The Neutron Star 3 is the latest version I believe.
I have a Neutron Star 2 @49.152 Mhz and upsample everything to 24/192 in my music server. Currently connected via a cheap SMA to u.fl connector to the Fifopi and it sounds great, but wow cant wait to try it with a Sinepi and use a quality SMA - SMA cable.
I have a Neutron Star 2 @49.152 Mhz and upsample everything to 24/192 in my music server. Currently connected via a cheap SMA to u.fl connector to the Fifopi and it sounds great, but wow cant wait to try it with a Sinepi and use a quality SMA - SMA cable.