New (was back in 2021) iFi Zen Stream device

Yea my ZS is in the living room in a TV stand so too hard to get to it to change the dial.

2 Likes

It’s arguably a bit of snake oil anyway. They already built in ANC and a data buffer into the USB port that should obviate any jitter issues anyway—and most decent DACs have hardware buffers in their USB inputs that prevent clock timing errors these days as well. Plus it’s a pain to have the GUI turned off and have to switch back and forth with the clumsy orange tool included every time you want to check for firmware upgrades (and I keep losing mine).

I also keep mine on AIO mode because when I switch my DAC input from USB to coax to play my SACD transport, occasionally when I switch back to USB, the ZS source setting reverts to S/PDIF, or my DSD Direct setting will change to DOP. AIO works great and I think the software jitter argument is a ruse extrapolated from the phenomenon that can occur with full-fledged PCs and sound cards.

2 Likes

You may not remember that but there was one thing that we noticed together - PCM 176,4 and higher was not working properly over Coax in AIO mode (frequent short breaks/pauses) but played fine in Roon-only mode.
That behaves now better with newer versions of firmware and/or Roon software. But there was a big difference between those two modes.

So switching all the sources off in AIO mode is not exactly Roon-only mode. The webserver & Volumio-based app is the obvious difference. Maybe something else uses ZS computing resources in AIO mode.

What doesn’t change the fact that Roon-only mode is unusable from practical point of view :slight_smile:

3 Likes

Yes I remember this issue and it doesn’t seem to be any better in the latest update.

That would be the main reason for using Roon only mode at the moment if you can get 24/192 over SPDIF, So maybe they developed this bug for you :grin:

2 Likes

This is just me, but the USB output on the ZS was one of its major selling points for me—I have a ton of files with higher sampling rates and I use Roon upscaling to DSD 256 for FLAC files, so USB connections are a must for me. Plus my coax input on my Topping is occupied by my SACD transport via an HDMI de-embedder.

Coax connections I’ve had nothing but problems with—I had to get the HDMI de-embedder for my SACD player for the same reasons you guys have mentioned. I was experiencing constant audio dropouts every few seconds, and a maximum sampling rate of 176.4.

Many of these issues are specific to the input receiver utilized by the DAC, in my case a CS8416 in the Topping that is the root cause of the audio dropouts and sampling frequency limits. It’s a widespread problem that extends far beyond Roon or the ZS, and in most cases firmware updates can’t fix it :cry:

1 Like

I’ve experienced this exact issue with coax in other situations. Please see my last post…

Easy solution for me was to set into 24/96 max.
But I do most of my listening on the Stream like you do over USB at DSD256.

1 Like

This is a response from John Yang, the engineer who designed the Topping, in regard to the audio dropouts on coax—it’s a known, pervasive issue across many platforms…

1 Like

It does seem that I have never really got SPDIF working reliably beyond 24/96. I have often struggled to get Toslink working reliably beyond 24/48 before the last few years.
So what is said there makes sense to me

1 Like

Every device has an input receiver chip, that is required to handshake with a gazillion different devices—CD players, streamers, TVs, audio components, the sky’s the limit. And if it can’t lock onto the device in question, a ton of jitter and timing errors get introduced and the audible result is dropouts. I tried four different HDMI de-embedders with my CD player before I got a solid signal with my Topping. Yang’s advice was to just try different players until I found one that was compatible with the DAC—obvious an impractical solution.

Coax and TOSLINK are unidirectional connections, unlike USB, which is bi-directional and much higher in bandwidth. The prior benefit for S/PDIF was full galvanic isolation and reduced noise compared to other connections. But since DACs introduced hardware buffers to override timing errors and associated jitter, USB has become the reliable standard and S/PDIF has become increasingly obsolete. Most components are including it only for compatibility reasons, and I doubt it will be around in coming years (it’ll go the way of S-Video)…

1 Like

Honestly that’s again something not in line with my experience.
If that’s a general Coax-DAC problem far beyond ZS then how change of mode from AIO to Roon-only could fix that?
For me that looks like a lack of computing power or inefficient implementation in Zen Stream. When you release some of the resources than ZS can handle higher bitrate.

It’s strange that it would be that tight.
Especially with nothing running, but it’s the reason they gave for the specific modes, to give all the power to the individual service selected.

I will go back to a previous statement, my aging ear’s can hear no difference between 24/96 and 24/192, besides on the Stream using SPDIF when the 24/96! sounds much better :face_with_hand_over_mouth:

1 Like

I get your point, but who knows what is going on from a hardware standpoint in that dial’s architecture. The only explanation for why the dropouts stopped is for the ZS connection to lock on to the clock of the DAC better with the hardware setting change—and that’s a phenomenon going on with your DAC’s input receiver.

As I said in my later post, the suggestion from the engineer who designed my DAC was to try different players until I found one that would successfully handshake with my DAC, because there was no firmware solution that would fix it.

The source selector dial on the ZS has an effect on the circuit board connections to your DAC, unlike the software toggles. Since dropouts are by design a function of the clocks syncing between the two devices, something in the internal connections in the ZS modified by that dial is affording you better synchronization with your DAC—and although I can’t say this definitively, based on John Yang’s explanation, it’s unlikely that a firmware change could impact that.

As to your suggestion that it’s a resource limitation in the ZS behind this, I can’t see how the ZS could process my DSD256 files without a hitch via USB and lack sufficient computing resources to handle 192 PCM via coax. DSD at that bandwidth is far more processor-intensive than PCM.

The issue has to be with the coax input receiver on your DAC handshaking with the coax output on the ZS—and if that varies with the dial on the back, then the changes in architecture modified by that dial are somehow affording your DAC’s input receiver to lock onto the signal. Nothing else makes sense.

I have read this thread since its inception, and I still have no clue if I want to spend the $400.

It looks good, no it doesn’t. This firmware works, this one doesn’t

I’m legit sticking with a RPi4 as it works. For now.

4 Likes

Nah, it works perfectly fine. But there’s nothing wrong with an RPI solution, if you don’t mind performing all the upkeep.

I bought the ZS for its ANC and anti-jitter tech built into the USB port, for ifi’s reputation with quality components and noise reduction, and its Roon integration. Despite some hiccups along the way, it’s far and away the best streamer I’ve found, particularly at this price point.

The stuff you’re reading here is no different than the struggles folks with an endless number of implementations encounter with any other new piece of gear—if you don’t believe me, head over to Head-Fi, search for the audio component of your choice, and grab the popcorn…

3 Likes

You guys are gluttons for punishment. I built a Pi streamer 3 years ago with RoPieee, and haven’t fiddled with it since. All the upkeep?

2 Likes

You have had better luck than I. I did nothing but re-flash SD cards and reboot devices (Allo USBridge Sig & RPI3 w/LCD) when I used RoPieee. I had better luck with DietPi on my USBridge until a major update killed the web GUI. Despite the ZS’s well documented issues and botched launch I haven’t had any show stoppers like with the RPI’s. Most of the issues are just cosmetic/annoyances or irrelevant to me since I only use Roon.

Looks like I’ll be keeping my ZS locked in mode 2 and running the Jan 2022 firmware indefinitely. Disappointing to hear that iFi has dropped the ball again. I wonder if we can petition Roon Labs to revoke Roon Ready certification for the ZEN Stream if the experience is crap…

1 Like

I use Roon Nucleus+, ZEN stream and Accuphase DAC DC-1000. With stable firmware 2.30.7, both PCM and DSD could be played nicely.

But, on stable firmware 2.31.7, only DSD cannot be played at all as on beta firmware 2.31.7. The issues of beta firmware 2.31.7 do not appear to have been corrcted.