Dante AVIO Review (streaming audio interfaces)

[PeteL]

Not sure about the accuracy of this. May be I misunderstood but last time I looked at Dante specs, the latency "handshake" was designed to manage transceiver processing latency (not network latency). If a transceiver at the other end cannot process packets faster than every X ms (including buffering), there is no reason to send packets faster than that. So that is the minimum latency requirement as far as that transceiver is concerned. This is hardwired into the transceiver (which knows what latency happens inside its processing) and is the same on any network (topology). The sender could decide to send packets at a slower speed than that.

If there is a peer-peer direct connection between them then that is the latency for transmission.

However, if it is on a shared IP network, then I am not sure how the network latency is automatically figured out. PTP is just for clock synchronization as far as I understand it.

Then there is the issue of variable latency in a shared IP network unless the network is purpose-built with QoS guarantees for latency.

I don't see this protocol as a simple plug-and-play for a consumer application except for point to pont (or you have network engineering expertise) in which case an USB will do just as well without the transceiver complexity.

In studio settings and local or wide area distribution, absolutely.

Having said that, I don't think the latency is an issue with most simple topologies on private networks but it could make a difference in the ability to do live recording/playback from multiple sources (unless the latencies are as low as the typical audio interfaces).

The latency of the network will be the same from any input to output and vice versa. that’s based on pinging for worst case scenario. that’s post adc and pre dacs, you plug a laptop in the switch with dante controller application, and it will tell you what this latency is, along with many infos

 

[ctcwired]

I believe the main reason these models only go down to 1ms and not lower is due to the PHY rate latency of 10/100 ethernet naturally being too close / risky to handshake at 500μs. That or it's just a limitation of the Ultimo chipset in general. Either way almost all Dante hardware defaults to 1ms.

 

[Vasr]

The audio bitstream is typically RTP (or a proprietary flavor of it) over unicast, or optionally multicast. Dante supports IGMPv2 and DSCP flags to assist with reliability on shared infrastructure.

That just means you reduce packet overheads and can participate in QoS features of intermediate nodes but someone needs to have engineered/configured them correctly for that network. This is fine in studio or pro-audio settings (or a professional installer in a house for zone distributions with well-configured routers).

The latency of the network will be the same from any input to output and vice versa. that’s post adc and pre dacs, you plug a laptop in the switch with dante controller application, and it will tell you what this latency is, along with many infos

You need to be more specific than that on what latencies you are talking about. There is processing latencies within each transceiver. The lowest latency it can handle. When you connect it to a network the Dante Controller on either end "handshake" to agree on this minumum latency but I am not sure where the network latency is figured into this. These transceiver latencies are fixed. That is not the same thing as network latencies between a Dante source and a sink on a network.

Any issues with network topology induced latencies are specific to the network and if not engineered correctly for QoS, can vary. Imagine just connecting your Dante-enabled pre/pro to your local Wifi router usinga WiFi dongle with a Dante-enabled media server on your network. No amount of handshaking is going to make it work right when someone starts downloading a Netflix 4k movie on the same network. On a well-engineered and configured network (as in a studio or a professionally installed home audio network), these could be managed. Just saying this is not for just plugging a Dante device into your WiFi or nearest Ethernet jack at home.

VoIP systems have suffered through this for a while and they don't have the high-fidelity requirements.

May be I am missing something in some part of Dante completely.

 

[ctcwired]

Meanwhile Dante Virtual Soundcard handshakes at either 4ms, 6ms or 10ms (it's so high because it has to cope with non-realtime network stacks in Windows / macOS). Add ASIO or WASAPI buffers to that, then double for roundtrip. It also requires at least one hardware device for clock source. It's mostly meant for basic recording & playback.

They do sell a PCIe card with an ethernet port powered by their fancy FGPA, and speaks ASIO or CoreAudio to the host PC if you desire lower latency. It basically pretends to be a normal soundcard. Figure the Dante handshake + ASIO buffer, you can probably get roundtrips of 4 - 8ms through a DAW depending on sample rate and how risky you're feeling with the ASIO buffer.

 

[ctcwired]

These transceiver latencies are fixed. That is not the same thing as network latencies between a Dante source and a sink on a network.

Any issues with network topology induced latencies are specific to the network and if not engineered correctly for QoS, can vary.

This is exactly the problem they aim to solve. By constantly monitoring/testing, nudging tick rates up & down, and maintaining a buffer, the Dante devices can survive the natural "weather" of a typical switched network. The lower you set the handshake setting the higher the risk of it getting disturbed or having underruns, thus the setting is available to be changed. They also provide tools to watch how consistent the network latency is over time with a graph & histogram, which can help test how much risk you have in a given environment.

 

[ctcwired]

Imagine just connecting your Dante-enabled pre/pro to your local Wifi router usinga WiFi dongle

Yeah this is already outside the scope of the target market. Their documentation basically says "WiFi is very not recommended." It probably can be done but you'd want to select the highest handshake setting or risk lots of errors / underruns / handshake failures. I believe they have a specific error code for "network latency is already higher than desired handshake value."

It's very valid indeed to say Dante is not intended for home consumers.

 

[PeteL]

That just means you reduce packet overheads and can participate in QoS features of intermediate nodes but someone needs to have engineered/configured them correctly for that network. This is fine in studio or pro-audio settings (or a professional installer in a house for zone distributions with well-configured routers).



You need to be more specific than that on what latencies you are talking about. There is processing latencies within each transceiver. The lowest latency it can handle. When you connect it to a network the Dante Controller on either end "handshake" to agree on this minumum latency but I am not sure where the network latency is figured into this. These transceiver latencies are fixed. That is not the same thing as network latencies between a Dante source and a sink on a network.

Any issues with network topology induced latencies are specific to the network and if not engineered correctly for QoS, can vary. Imagine just connecting your Dante-enabled pre/pro to your local Wifi router usinga WiFi dongle with a Dante-enabled media server on your network. No amount of handshaking is going to make it work right when someone starts downloading a Netflix 4k movie on the same network. On a well-engineered and configured network (as in a studio or a professionally installed home audio network), these could be managed. Just saying this is not for just plugging a Dante device into your WiFi or nearest Ethernet jack at home.

VoIP systems have suffered through this for a while and they don't have the high-fidelity requirements.

May be I am missing something in some part of Dante completely.

I believe @ctcwired answer the point about tranceivers latency, for QOS, yes sure your network should be configured properly, and on a seperate LAN.

 

[Vasr]

This is exactly the problem they aim to solve. By constantly monitoring/testing, nudging tick rates up & down, and maintaining a buffer, the Dante devices can survive the natural "weather" of a typical switched network.

Do you have a reference link for which part of Dante protocol manages this specifically (not the pointer to the entire Dante documentation which I have). I mean this adaptive capability during play if it exists (as opposed to initial setting at connection). Just curious. I am certainly not an expert on this and last time I looked, I didn't see anything that talked about the above capability. The documentation is quite daunting.

They also provide tools to watch how consistent the network latency is over time with a graph & histogram, which can help test how much risk you have in a given environment.

Which should keep our home audio warrior busy with more graphs to look at in addition the room eq graphs.

It's very valid indeed to say Dante is not intended for home consumers.

That was exactly my take away the last time I read it and yet I see people wishing for Dante interfaces on their home audio equipment.

 

[ctcwired]

Do you have a reference link for which part of Dante protocol manages this specifically

Apologies as this is beyond even their "Level 3 certification" course.

https://dev.audinate.com/GA/dante-controller/userguide/webhelp/content/clock_synchronization.htm

On that page they do mention: "A Dante device set to 'Enable Sync To External' will use the external word clock from its host equipment to tune its onboard VCXO."

I assume this is hinting that devices continuously monitor and adjust for drift over time to follow a master clock, but I don't see them explicitly state that. I don't have access to OEM documentation. One way to test would be to get multiple input modules to ADC the same signal, record it, and null test how aligned the samples produced by the two units stay over time. I'm guessing they would drift back and forth by a sample or two over the course of a few minutes. I might test this and let you know.

 

[PeteL]

Apologies as this is beyond even their "Level 3 certification" course.

https://dev.audinate.com/GA/dante-controller/userguide/webhelp/content/clock_synchronization.htm

On that page they do mention: "A Dante device set to 'Enable Sync To External' will use the external word clock from its host equipment to tune its onboard VCXO."

I assume this is hinting that devices continuously monitor and adjust for drift over time to follow a master clock, but I don't see them explicitly state that. I don't have access to OEM documentation. One way to test would be to get multiple input modules to ADC the same signal, record it, and null test how aligned the samples produced by the two units stay over time. I'm guessing they would drift back and forth by a sample or two over the course of a few minutes. I might test this and let you know.

I'm not sure I saw a form of adaptative tick rate neither. I am level 3 certified too but not an expert in networking specifically, so in exact terms I don't have the maybe exact vocabulary, but there is a subscription an a negociation process between each source and syncs, and there is a scheduler (time stamps?), then if you set a latency that the network can't provide, taking into account QOS, there will be a scheduler error, but I don't think there will be changes during playback, as far as I understand.

 

[PeteL]

I don't see this protocol as a simple plug-and-play for a consumer application except for point to pont (or you have network engineering expertise) in which case an USB will do just as well without the transceiver complexity.

If nothing else, this product here will allow you to travel the distance, that USB can't and it makes a very affordable music streamer if you don't need wireless, and arguably easier to setup than a Pi, that don't have a great dac in it's stock form neither anyway. Unfortunatly the performance is not great, I'd like to see the Amphenol version maybe. There is no reason to believe it would sound worst than something like a chrome cast. Granted, these device offer maybe less flexibility. Yes it's aimed at pro, but if a AV Receiver do have a Dante Input, there is nothing really wrong in using a PC as a Dante virtual sound card and stream to it, personally my Blusound streamer has an Ethernet connection already rather than wifi, and it's not as cheap.

 

[Blumlein 88]

https://www.rdlnet.com/Dante/

Another supplier for Dante things. More expensive than the reviewed item here or the Amphenol offerings.

 

[ctcwired]

Unfortunatly the performance is not great, I'd like to see the Amphenol version maybe.

It is unfortunate. I suppose the economies of scale aren't the same for products like these compared to something more popular like a Focusrite 2i2 or a MOTU M2.

Slowly, more products like this are starting to appear: https://www.bhphotovideo.com/c/prod...nt22x_planet_22x_professional_2_in_2_out.html

I used one on a gig for a bit and it was fine, but wasn't impressed with the build quality. How much you wanna bet it uses the same reference design internally as these AVIO units? It's a good attempt but the price for what you get is still too high.

Focusrite's lowest end full I/O RedNet interface is this: https://www.sweetwater.com/store/de...x2-ethernet-audio-digital-io-with-mic-preamps

I'm sure they're decent if you can get past that price tag for what is effectively a 2i2 with but with Dante. It might not seem too bad...

Until you realize that for just $150 more you might as well get one of these chonkers: https://www.sweetwater.com/store/detail/Tio1608D--yamaha-tio1608-d-digital-stage-box-with-dante

 

[Jimster480]

Very interesting product, so this actually puts audio over an IP network not just over ethernet cables? that is totally different than some products which I have tested through vine; which basically use Ethernet to extend a variety of connections from HDMI, USB to RCA & 3.5mm...

 

[Rja4000]

One should not confuse the performance of this cheap after-thought converter (mainly meant to solve a secondary last minute problem or to allow easy and headacke-free sound distribution on some remote area) and the performance capability of the Dante system as a whole.

Most big sound enforcement events worldwide are based on Dante nowadays.
Make no mistake: Huge amount of IOs, easily sharing signal for multiple simultaneous purposes (FOH, monitors, broadcast, recording,...) :
Ethernet-distributed sound is a revolution.

I personnaly use it for more than 10 years to distribute sound betwern Mixer, preamps, effects and PC(s). (I'm not pro, though)
This system is simply great.

Now, those cheap adapters are speced as we see here: 0.01% THD is not what we're looking for here, for sure.
For sure, we whish they set a target to cover at least 16 bits.
Use a serious DAC behind an Dante-to-AES converter though, and you'll get all the performance it's capable of, for sure.

 

[Vasr]

If nothing else, this product here will allow you to travel the distance, that USB can't and it makes a very affordable music streamer if you don't need wireless, and arguably easier to setup than a Pi, that don't have a great dac in it's stock form neither anyway. Unfortunatly the performance is not great, I'd like to see the Amphenol version maybe. There is no reason to believe it would sound worst than something like a chrome cast. Granted, these device offer maybe less flexibility. Yes it's aimed at pro, but if a AV Receiver do have a Dante Input, there is nothing really wrong in using a PC as a Dante virtual sound card and stream to it, personally my Blusound streamer has an Ethernet connection already rather than wifi, and it's not as cheap.

If the AVR has ethernet/wifi, then you can just stream the content as a file over ethernet and use the AVR's DAC. This is built into the AVR. There is no need for this dongle. Enabling Dante won't provide anything additional for this use case nor would you need to bundle a DAC with a Dante interface. The AVR already has the buffers necesssary for its ethernet ports.

If the AVR has no Ethernet capability, Chromecast or other dongles attached to the HDMI ports also supply a digital input to the AVR from the streaming which can use its own DAC (and DSP for room eq etc without additional ADC).

So, I just don't see the value in an AVR or pre/pro for providing Dante functionality as a sink.... Perhaps if the AVR is required to be used as a source for zone distribution in a home audio network and it can output to the Dante device to unicast or multicast to one or more sinks.

The use case for a dongle like this is really to hook up a IP enabled source to an analog mixer simply as an ethernet enabled DAC. This assumes the source is also Dante enabled and you have multiple sources to justify it. Studios can use AES/EBU instead for long runs with DACs that can take that input. But that requires a pair of cables for each source. The Dante audio over ethernet can simplify this with a single cable switched/routed network for all sources. This is the only value I see.

 

[ernestcarl]

These kind of remind me of the Lyxpro XLR network breakouts (pure analog).

I use a pair of something similar with only two outputs to send audio to the garage and upstairs bedroom. Works perfectly fine and I can't really hear any significant or additional audible noise. However, the amount of noise will depend in the house's network wiring & shielding. In our much older house through a 100ft cabling, I could hear significant crosstalk.

 

[pierre]

It looks like with a PoE network switch you could connect multiples of these and have some matrix control with Dante Controller.

I have a few dante to AES3 from Audinate and yes you can aggregate them easily. That's how I did my atmos speakers: i did put a small 8 ports switch on the ceiling and then went to all speakers and used AES in.

 

[jones]

The measurements done here are a disappointment at first but on the road that should be fine for noncritical applications.
For me the most important thing with such devices is that the SNR ist high enough to not turn your PA into a noise generator.
This seems to be fine here.

The equivalent modules from Amphenol problably perform the same because they most likely use the same boards like the AVIO devices as Audinate also sells them to manufacturers. see here
All other Dante solutions are also a lot more expensive so there is not an option really.

I like to use Dante at events a lot. Less cables are always good and staying in the digital domain between input and speaker becomes much easier.
That said: Are there plans to measure more Dante equipment? I would enjoy that a lot

 

[tifune]

The audio bitstream is typically RTP (or a proprietary flavor of it) over unicast, or optionally multicast. Dante supports IGMPv2 and DSCP flags to assist with reliability on shared infrastructure.

Hey finally a language on this board I can speak fluently! In medium-large setups, are there admins initially brought in to setup QoS, etc? Or maybe Dante just has a list of recommended switches proven to work out of box?