Audial S5b paired with Matrix X-SPDIF 2

Quote from Pedja on 23 April 2025, 11:03 PMAt its USB input, the S5b can accept sampling frequencies up to 384 kHz, so it will not be the limiting factor in the sampling frequency domain.
Now, is over- or up-sampling advantageous, that would again come down to your preferences, source/upsampler quality, etc. I again prefer to send to the DAC the original files/source content unaltered, i.e., non-oversampling, but it is me (a non-oversampling devotee for more than two decades).
But if you like the upsampling with Primare, you should probably try the same with S5b, too. I am unsure if synergy would work the same, but it is worth trying. I know people using Audial TDA1541A DACs with software over-/up-sampling, and they report to like it. I fully respect that, and I believe it works well. Finally, the reason why I was probably the first to go with non-oversampling DACs above 44.1 kHz was not only to allow playing hi-res non-o/s, but also to allow software over-/up-sampling.
Anyhow, there is no technical requirement to downsample specifically for S5b.
At its USB input, the S5b can accept sampling frequencies up to 384 kHz, so it will not be the limiting factor in the sampling frequency domain.
Now, is over- or up-sampling advantageous, that would again come down to your preferences, source/upsampler quality, etc. I again prefer to send to the DAC the original files/source content unaltered, i.e., non-oversampling, but it is me (a non-oversampling devotee for more than two decades).
But if you like the upsampling with Primare, you should probably try the same with S5b, too. I am unsure if synergy would work the same, but it is worth trying. I know people using Audial TDA1541A DACs with software over-/up-sampling, and they report to like it. I fully respect that, and I believe it works well. Finally, the reason why I was probably the first to go with non-oversampling DACs above 44.1 kHz was not only to allow playing hi-res non-o/s, but also to allow software over-/up-sampling.
Anyhow, there is no technical requirement to downsample specifically for S5b.
Quote from PKay on 23 April 2025, 11:22 PMHi Pedja,
What is the upsampling frequencies that the S5B can accept on the coax and i2S inputs?
Hi Pedja,
What is the upsampling frequencies that the S5B can accept on the coax and i2S inputs?
Quote from gryphongryph on 24 April 2025, 10:07 AMOk, as I understand it, I can just let Roon send the original 16, 96, or 192 to both my rooms, my Primare will automatically upsample everything to 192 and my S5b will just not bother with anything over 16 bits and that should play at best sound quality on both my dac’s.
Ok, as I understand it, I can just let Roon send the original 16, 96, or 192 to both my rooms, my Primare will automatically upsample everything to 192 and my S5b will just not bother with anything over 16 bits and that should play at best sound quality on both my dac’s.

Quote from Pedja on 24 April 2025, 9:04 PMYes, that is correct. Regarding the sound quality, you can still try the possible upsampling options with Roon, and see how things change. It can be fun.
Quote from PKay on 23 April 2025, 11:22 PMHi Pedja,
What is the upsampling frequencies that the S5B can accept on the coax and i2S inputs?
Fs max for both S/PDIF inputs, so for both BNC coax and Toslink optical, is 96 kHz.
For I2S LVDS input and A-link input in I2S mode, the practical limit is also 96 kHz, assuming the source outputs the most usual 64-bit I2S frame.
For A-link operating in simultaneous data mode, with supposed and most probable 32-bit frame, the Fs limit is 384 kHz.
Yes, that is correct. Regarding the sound quality, you can still try the possible upsampling options with Roon, and see how things change. It can be fun.
Quote from PKay on 23 April 2025, 11:22 PMHi Pedja,
What is the upsampling frequencies that the S5B can accept on the coax and i2S inputs?
Fs max for both S/PDIF inputs, so for both BNC coax and Toslink optical, is 96 kHz.
For I2S LVDS input and A-link input in I2S mode, the practical limit is also 96 kHz, assuming the source outputs the most usual 64-bit I2S frame.
For A-link operating in simultaneous data mode, with supposed and most probable 32-bit frame, the Fs limit is 384 kHz.

Quote from Pedja on 24 April 2025, 11:14 PMAnd, before you ask... If a 32-bit simultaneous data frame can go up to 384 kHz, why can't a 64-bit I2S frame go up to 192 kHz?
The answer is that these two protocols are somewhat different, so the TDA1541A internal logic routes these signals differently as well. Nominally, for the TDA1541A, a 32-bit frame simultaneous data operation should be limited to 192 kHz (because of the claimed 6.4 MHz bit clock limit), but I found a hack that proved it capable of operating at frequencies twice as high. The hack, however, did not work for I2S, but only for simultaneous data mode.
If I ever find a way to make TDA1541A operate up to 192 kHz in I2S, I will let you know. 🙂
And, before you ask... If a 32-bit simultaneous data frame can go up to 384 kHz, why can't a 64-bit I2S frame go up to 192 kHz?
The answer is that these two protocols are somewhat different, so the TDA1541A internal logic routes these signals differently as well. Nominally, for the TDA1541A, a 32-bit frame simultaneous data operation should be limited to 192 kHz (because of the claimed 6.4 MHz bit clock limit), but I found a hack that proved it capable of operating at frequencies twice as high. The hack, however, did not work for I2S, but only for simultaneous data mode.
If I ever find a way to make TDA1541A operate up to 192 kHz in I2S, I will let you know. 🙂