I completely agree. Need both (and I don’t know of anyone here who has suggested anything different). Listening and subjectivity are more easily accomplished and reported by more of us, often with very little rigor around the process. I do this all the time as well. Impressions are easy. We need more measurements in my view to provide better understanding of what’s happening “in there.” Regardless of the (sometimes extreme) prejudices against any subjectivity over at ASR, some of the reported work is very informative. None of that keeps us from listening. But listening should also not blind us to the realities of psychoacoustics and cognitive bias. They are demonstrably real phenomena.
I believe we should strive for less polarization in this hobby. The subjective and objective are not incapable of balance and integration. They should inform each other, not stand in opposition.
Erno, my overall home network is a little complicated, but as far as my audio systems are concerned, the main Ethernet Switch is an 8-port all-SFP Mikrotik CRS309. From there, all audio devices are connected over fibre; to my Vivaldi stack with an independent Media Converter/LPSU (TrendNet/Keces P8, although I’m testing an integrated unit currently), and to my home-office Bartok & Lina stack with a Juniper EX2300-C. I use a range of SFPs, but mainly II-VI (the “Finisar” that most folks are familiar with) and Mikrotik/Juniper (actually FS & II-VI OEMs).
Elsewhere in my home there about half a dozen other Ethernet Switch/Routers, mainly Arista, Cisco, and Juniper; all Enterprise-grade platforms powered-on 24/7 (most haven’t been reset/power-cycled in years! )
I do have a couple of B.S “Audiophile Ethernet Switches” collecting dust in my storeroom. With the exception of a couple (Uptone and Innuos come to mind), the rest are all just cheapo $30-type Switches rebadged with better PSU and shiny cases (i.e. lipstick on a pig )
IMHO, the standard set of Audio Precision test repertoire that Amir goes through will not highlight some differences that we can actually hear. Especially those near the audible noise-floor or soundstage/imaging related differences - humans are much more efficient at pattern-recognition that an AP or Oscilloscope will only show as random differences or un-correlated measurements.
However, those kinds of audible differences are usually very very small, and require far more tests than the standard ones in order to be visible. The “huge” audible differences some people claim, especially the ones where they “don’t even need to do a blind-test to hear”, will be readily visible in standard AP tests.
I forgot my password into that specific Switch and it hasn’t be reset/power-cycled in 5 years! (I still get full line-rate performance without any packet loss out of it ). Albeit, I’ve been hunting for a similar fan-less unit with higher port count.
My broadband is also 10G, so the uplink from the Mikrotik to my Internet Router is via a 10G SFP+ (Copper). I’d definitely recommend that Mikrotik, it’s low-cost, bullet proof, and relatively easy to manage as a layer 2 switch.
No actually. The Bartok HDAC/non-Apex + Rossini Clock pairing which I’ve had since launch, and now a full Lina stack beside it. Eventually one set will stay and one will go. I haven’t decided which quite yet
Mine’s the Huawei HN8254; it’s a 10G-PON ONR (Optical Network Router). Like most ISP demarcation terminals/routers, consumers don’t have a choice on these 10G ONRs, they’re specifically supplied-by the Broadband Service Provider as they’re usually OLT dependent.
The Mikrotik CRS309 Ethernet Switch that I mentioned previously is a 10G Switch and it connects directly to the ONR’s 10G port via (Copper) Ethernet.
Thanks, Anup. Do you use another router behind the ONR or can you rely on the ISP’s router for NAT/DHCP etc? We have an ADVA, which is locked down, and a second router is required to handle NAT etc.
Nope, I don’t. Being an ONR it has full Routing, NAT, Filtering, IP address management etc. As opposed to an ONT (Terminal) which typically are pure layer 2 demarcation devices only and would require a router behind it. I’m assuming your ADVA is L2 only? What’s the model number? (by the way, ADVA was acquired by ADTRAN in '21).
The 10G Mikrotik CRS309 that I use is in fact is capable of being both a pure layer 2 Ethernet Switch (which is how I use it), as well as a full blown layer 3 Router; you get to select between two Operating System modes - SwOS (pure Layer 2), or RouterOS (full Layer 3, with a very rich set of Enterprise-class Routing capabilities, including MPLS, VPLS, BGP, Virtual Routing etc).
I was provided an FSP150-GE102PRO, but believe another model is going to be required when the line is upgraded. Interesting that the Mikrotik can run as a switch or a router. I have a CCR1009 lying around, but the ports are limited to 1Gbps. For a lot of people, the Mikrotik is probably a great way forward
PS I realise I should probably move this to a PM as I’m going OT!
Back on the topic of this thread somewhat, but specifically on Ethernet Streamers. As an side, I happened to come across a review of the Holo Audio Red Streamer which, similar to the Metrum, uses a Raspberry Pi Compute module, and has both AES, SPDIF outputs, along with I2S, and USB.
However, its just over half the price of the Metrum Amber, and while it doesn’t have a TentLabs Clock and I don’t see any opto-isolation, from an internal design standpoint, frankly, puts the Metrum Amber to shame.
Additionally, it’s SPDIF output was measured with an AP analyser for jitter (FFT on the Jitter spectrum).
No comparisons were made using different Ethernet Switches or anything, but considering USB and Ethernet measured the same from a noise and jitter perspective, I certainly don’t expect any impact from different Ethernet Switches.
Additionally worth noting that performance did not change regardless of whether USB input or streaming input was used.”
The Holo RED is the lowest jitter and one of the lowest noise digital sources I’ve tested. Absolutely incredible performance that sets the current benchmark for a streamer or DDC. The fact that this level of performance is achieved at under $1000 is honestly incredible.
)
)
