You must consider the nature of the transmission path and any potentials for noise induced corruption of the data signals from network components and wiring path, etc… Noise on the power/ground being delivered to the network components have impact on the performance of those components… something as benign as a light being switched on, can cause spurious events… I’ve read where digital door-bells and security cameras on the network can precipitate interrupts…
What I notice with upsampling is that the first track always starts fine, while every next track has a slight hiccup. Sometimes the lock on the track won’t happen at all and replay stops completely.
@Antoine is there an option or software solution which might solve this? Dsd upsampling is quite addictive as the sound improves quite a lot: more air around instruments and depth and width I never heard so well.
Does this hiccup happen when modulating to DSD128?
128 and 256 are fine, only 512 has hiccups. Apart from the hiccups 512 sounds just a bit better.
My perspective on this is based on some salient realities…
A 16/44.14kHz file is limited by the Nyquist Fs of @22kHz where there is no recorded information available beyond the brick-wall filter applied on the original PCM ADC (Analog to Digital Conversion) encoding… The DSD output resolution/dynamic range will reveal low-level information that was captured in the original ADC that was masked by the 16 bit/44.14kHz PCM encoding noise-floor…
The dynamic range of the modulation of the PCM data to 1-bit PDM (DSD) is a function of the noise-shaping algorithm used in the DSD encoding processing… The Nyquist Fs of a DSD128 (5.6MHz) signal is 2.8MHz … A PCMxxx file modulated to DSD128 (5.6MHz) vs the same PCMxxx file modulated to DSD256 (11.2MHz) will not gain any more dynamic range or add any new information beyond that which was encoded in the original PCM ADC… Some will argue that the noise-shaping Nyquist Fs is pushed farther out in the 11.2MHz (DSD256) moduation and this makes an audible difference… This is just wishful thinking because if there exists any noise-shaping artifacts below the Nyquist Fs these harmonics will be so low-level as to be far, far below the threshold of hearing, and inconsequential…
There are limits of resolution in a PCMxxx ADC encoding that cannot be overcome in modulation to any 1-bit PDM sample-rate… it is fixed by the PCMxxx ADC sample-rate.
Now when we shift the focus to native 1-bit PDM (DSD) encoded recordings… The nature of the comparison changes as the sample-rate increases which translates directly into captured resolution (Again, Dynamic Range is a function of the noise-shaping algorithm in the 1-bit PDM encoding process and is the same for all 1-bit PDM encoding sample-rates)… So, recordings made at different 1-bit PDM sample-rates will have different sonic signatures… just as PCM recordings done at different bit-depth and sample-rates will have different sonic signatures.
In the case of modulated (up-sampled) PCMxxx to DSD256 or DSD512 it is highly improbable that we can subjectively perceive tangible sonic elements in juxtaposition to the same PCMxxx file being modulated to DSD128.
My take on your DSD512 playback hiccups is based on the computational demands you are placing on your Mac Mini when up-sampling/modulating to DSD512 and managing the UPnP processing and transport of such a dense amount of bits, in concert with the bidirectional latency demands of the transport scheme/network architecture.
I had a 2014 Mac Mini I5, I still have a 2018 I5 and now I bought a Mac Mini M2.
It is known that Apple computers have USB compatibility problems with various DACs. I had this problem myself.
However, now with the M2, and I believe with the M1 too, I noticed that there was an improvement in this issue of compatibility.
Could this not also happen for those who are using network file transmission?
Have Apple computers improved since the M1 for audio?
Thank you for your elaborate answer. I really appreciate it.
I can indeed see in activity monitor that with a 256 upsampling about 5,5MB/sec is uploaded and with 512 6,5MB/sec. Also the type of filter makes a difference. I’ll try to not hear the differences, this is made extra difficult because there is also a slight difference in loudness. What I do hear is a difference in harshness, depth and height listening to music for a longer time with a higher bitrate. What the reason might be I don’t know.
I’ll try out the filters for sound quality. I am listening for a while now to filter type C with dsd512 without hiccups. Before I used B8 with hiccups… not been able to compare them critically. I’ll report later.
I suggest using “BP” (Bit-Perfect) filter setting in the DAC and adjust r8Brain to suit your taste… It appears the NEO Stream does not utilize the analog FIR filtering of DSD signals available in the BB DSD1793 chipset… The four available digital filters in the NEO Stream are apparently applied to the PCMxxx signals, and are their own designs, bypassing the digital filters of the DSD1793 chipset. (This is an option of the DSD1793 chipset architecture)
From the Up-sampling Information guide in Audirvana:
For the R8Brain algorithm, you have access to the following settings:
○ Bandwidth: this indicates the limit of the low-pass filter as a % of the Nyquist frequency (half the sampling frequency). The slope of the filter is low at 74% and very steep at 99.5%. Make sure that this does not induce too much brightness (i.e. highs are too aggressive).
○ Cutoff Band Attenuation: this is the setting of the slope of the low pass filter expressed in dB per octave. With the maximum slope (218 dB), there may be too much brightness (i.e. the highs are too aggressive)
○ Phase: All low pass filters have some level of overshoot. The steeper the slope of the filter, the greater the overshoots. There are two types of overshoot: pre-oscillations and post- oscillations. The former are audible as a “pre-echo” arriving before the signal itself, and are the least natural to hear. You can choose between a filter setting with linear phase, but with an equal level of pre- and post-oscillation or a minimum phase filter, with no pre-oscillation, but non-linear phase distortion.
Computational speed in concert with memory bus speed will enhance all platform functions. The future looks bright for audio on Apple Inc. platforms…
What is your setup? Are you playing direct from Mac to DAC via USB, or something else?
Mac mini i7, 16gb connected to my Neo Stream via upnp, wired internet.
What’s your wired Ethernet spec? Edit: I mean what’s the spec of your Ethernet cables?
1Gbs, cat7 ethernet.
The Neo Stream has an optical connection for ethernet to the Neo.
Thanks. OK, plenty of network speed.
My DIY setup works without problems at DSD512, so I’m just left to wonder if it’s something with the NEO Stream. Sorry I couldn’t be of more help.
Thanks for helping! I appreciate it. Good to hear that it should work. I’ll try different settings, also WiFi, (474mbs) but that might give interrupts due to the use of other devices.
In my opinion, the ideal cable for audio is CAT6.
I think that CAT7 cables can bring some type of noise to the audio, due to the internal shielding.
I think it’s a subject that has already been debated.
My internet speed with CAT6
I made a mistake: I looked it up, it is cat8 cable. Should be good for 40Gbs… I specifically bought it for the streamer.
Cat6a would be best at the moment. It’s a newer spec than 7 and 8. Recommend the unshielded type, which doesn’t electrically connect the grounds at each end of the cable. But that’s just my recommendation - if you like what you have, go with it.
1 Like
I finally found the culprit for the hiccups. As it turned out the standard setting for audio buffering in the Ifi app is 30mb: counterintuitively I had to lower that to 2mb! I also tried the highest value of 64mb: nearly continuous hiccups. Who would have thought that! The audio buffer is supposed to help “to avoid chances of audio stuttering over the network” by increasing this value…
So finally undisturbed listening to dsd512 is possible. Thank you all who chimed in with ideas and suggestions for helping me out.
Hopefully both the warning for using the volume control and now the solution for hiccups helps other Neo Stream users as well.
4 Likes