Bo Deadly
Well-known member
I'm thinking about getting a new audio interface for all things home audio (TV, music, recording all runs through a MOTU Traveler right now). Clearly Ethernet audio protocols are starting to really catch on which is fantabulous for a number of reasons. Unfortuantely it seems there are at least two major different protocols (Dante and AVB) which means I need to pick which one to bet on. Anyone want to buy a Betamax player?
Protocol-wise Dante and AVB technically look very similar but there are some important differences:
Dante works with what is currently conventional Ethernet. And Dante has a lot of market share. AVB requires special switches and adoption is relatively small for at least pro-audio.
However, looking at this closer, I don't understand why AVB is not doing better.
AVB may require special switches but my understanding at this moment is that the "special" bits are just new Layer 2 IEEE 802.1 protocol standards that will eventually work it's way into networking chips. Intel has a PCIe card that you can buy right now (that works w/ Linux and they contribute significantly to Linux development). Recent Macs have AVB Ethernet ports. So over time this requirement should dissappear (at which point Dante will actually be harder to use with your PC or Mac because you have to buy and install their virtual sound card).
I have a slightly nutty theory as to why Dante is more popular. Audinate sells modules that implement the various hardware and software bits. That makes the bar very low for manufacturers to add Dante support by just adding a slot that does i2c and connects various pins. But again, this seems like a superficial advantage because it will dissapear in the long run as switches support AVB and mfgs implement the code and other stuff to support AVB in their gear.
So that leads me to what I think is the major disadvantage of Dante: it's proprietary. I actually don't have a problem with things being proprietary. Companies need to protect their investment so that they can develop the technology. However, it has to be supported on all applicable platforms and in this case that could be difficult. Linux for example is a bit of a moving target so it can be painful for a company to be expected to recompile their code for whatever version of Linux. The only solution IMO is to actually open source the driver (although making a firmware blob available for free distribution seems to satisfy most folks too). Linux already has AVB code in the works. But Audinate are doing something else that is a major issue with proprietary solutions: They are nickel-and-dime-ing people at every turn. Many Dante products are just a hardware shell around whatever module and yet prices seem a little high. For example, clearly the Amphenol RJD1212 Dante adapter is just a Dante Output Module dipped in rubber. But it costs almost $300 USD? And that's super cheap compared to most Dante products. MOTU is all AVB and their products are a little cheaper (although I suppose they have always targeted that market). I cannot help but wonder what the Audinate royalties are. Margins must be pretty good. Someone should make little AVB modules and charge a license fee for supporting, curating and burning code onto the chips.
So does anyone see AVB making headway in the long term?
Why is Dante more popular? Is there some technical limitation I'm missing?
If Intel and other network chipset mfgs eventually add AVB support to their chips such that PCs and Macs and routers have AVB capable ports, then do I really want to be investing in Dante products now? Why?
Protocol-wise Dante and AVB technically look very similar but there are some important differences:
Dante works with what is currently conventional Ethernet. And Dante has a lot of market share. AVB requires special switches and adoption is relatively small for at least pro-audio.
However, looking at this closer, I don't understand why AVB is not doing better.
AVB may require special switches but my understanding at this moment is that the "special" bits are just new Layer 2 IEEE 802.1 protocol standards that will eventually work it's way into networking chips. Intel has a PCIe card that you can buy right now (that works w/ Linux and they contribute significantly to Linux development). Recent Macs have AVB Ethernet ports. So over time this requirement should dissappear (at which point Dante will actually be harder to use with your PC or Mac because you have to buy and install their virtual sound card).
I have a slightly nutty theory as to why Dante is more popular. Audinate sells modules that implement the various hardware and software bits. That makes the bar very low for manufacturers to add Dante support by just adding a slot that does i2c and connects various pins. But again, this seems like a superficial advantage because it will dissapear in the long run as switches support AVB and mfgs implement the code and other stuff to support AVB in their gear.
So that leads me to what I think is the major disadvantage of Dante: it's proprietary. I actually don't have a problem with things being proprietary. Companies need to protect their investment so that they can develop the technology. However, it has to be supported on all applicable platforms and in this case that could be difficult. Linux for example is a bit of a moving target so it can be painful for a company to be expected to recompile their code for whatever version of Linux. The only solution IMO is to actually open source the driver (although making a firmware blob available for free distribution seems to satisfy most folks too). Linux already has AVB code in the works. But Audinate are doing something else that is a major issue with proprietary solutions: They are nickel-and-dime-ing people at every turn. Many Dante products are just a hardware shell around whatever module and yet prices seem a little high. For example, clearly the Amphenol RJD1212 Dante adapter is just a Dante Output Module dipped in rubber. But it costs almost $300 USD? And that's super cheap compared to most Dante products. MOTU is all AVB and their products are a little cheaper (although I suppose they have always targeted that market). I cannot help but wonder what the Audinate royalties are. Margins must be pretty good. Someone should make little AVB modules and charge a license fee for supporting, curating and burning code onto the chips.
So does anyone see AVB making headway in the long term?
Why is Dante more popular? Is there some technical limitation I'm missing?
If Intel and other network chipset mfgs eventually add AVB support to their chips such that PCs and Macs and routers have AVB capable ports, then do I really want to be investing in Dante products now? Why?
