I'm "this" guy who reverse-engineered the Aviom ANET and built a digital injector for my own personal use and thus understand really how it works. I also used various Personal Monitoring systems both on Stage and Studio. Before going into Aviom's architecture let me sum up my experience with the various products I used:
Paramount of importance is accessibility to make changes fast without thinking. Especially singers don't want to be interrupted while recording a take just to increase the volume of the playback. All systems that don't offer any haptics or were designed to be menu-driven fail in this regards. A simple pre-selection of the to-be-adjusted channel with a DIRECT ACCESS TO A PHYSICAL
BUTTON (with haptic properties) are OK for most artists however direct access to each channel's volume knob (such as the Elite Core or analog mixers) is better.
A small note on digital systems in general:A properly designed ADC system can sound as glorious as full blown analog. The trick is to get it right and not to do any shortcuts on the design - especially on the analog front-end and the power supply and PCB layout. I'm an engineer and studied a lot of pro-audio product designs and I can tell you that most of them are just too simple and there is no "analog love" (as I call it) to make perform the ADC in the best way. Let me give you an example here. Cadac desks play in the league of SSL, NEVE, TRIDENT etc and are well known for their glorious EQ and pre's (when comparing with non-tranny designs). They use massive power supplies, internal bussing is balanced and with CMRR feedback for what's called virtual earth summing, every critical Opamp is hand tuned for CMRR. components are best in breed etc... The list goes on and on. Interestingly Cadac moved to digital and never sacrificed on sound. In fact, using a properly built ADC and DAC front end, a lot of issues of the analog interfacing, reliability of contact quality, temperature drifts etc is not an issue any longer. Cadac wouldn't have gone to digital just because of the workflow, routing and configuration advantages of digital if the sound wasn't right and virtually indistinguishable from the analog desks. It can be trusted, but it must be done right. To blame all digital systems for bad sound, simply doesn't work. However, I can understand how this evolved and can relate to such opinions.
Coming back to the personal monitoring now ;-)
I judge the quality of a personal monitoring system (or any other generic system) in the following dimensions:
Quality describes simply the quality of the sound, the hardware (enclosure, cables, components used etc) and other aspects such as quality of documentation. We can all rely to this I think.
Availability describes how easily the system connects to each other and is made ready for operations. A problem here can be AVB based design (or any L2 based protocols) because they rely on proper networking, some of them being very sensitive to VLAN, IGMP, PTP, SRP, Multicast settings etc... That's why AVB based transport typically refers to some known-good switches because the need of the configuration on the switch itself. Another aspect is power supply: Is the design made in a way to be OK with daily power-off/on over a lifespan of almost 10 years? What happens if one power supply fails, do all ports fail, just one mixer, is the injector redundant, what about ground loops that may occur on shielded Ethernet cables especially when moving to unknown locations, things are "sometimes" working, sometimes not... And those PoE based mixers: How quiet is the switched mode supply on each mixer which is very close to the audio part, is it ground-free, what happens when somebody connects a wedge monitor (instead of "earth-free/floating" headphones) will the noise then be heard? What happens then a hi-impedance in-ear system is used which want's to be driven by voltage rather than current? Behringer's noise is a no-go to me and the Aviom A-16 is another one. Aviom A-16II is much better here! Actually, I'm impressed by the quality of the A-16II in terms of audio quality. One reason is that the nature of A-NET is Layer 1 only, another one is that each mixer have it's own main power supply (open the Hub - you will see 9 power supplies!) - a really good idea to decouple each "consumer" from each other and avoid intermodal distortion and short circuit protection and continuous operation - also preventing ground loops as they are isolated.
Accessibility describes how easy it is to perform manipulations - some points I have covered above. There is a solution called MyMix which really doesn't work: a nice idea but totally not based on real-world experience: selection of a channel is done by scrolling (!) through a big knob (the incremental dial that we know will fail and change scrolling direction because of the nature of grey code), then pressing the button (and as you press, you scroll one more) then you are entering "Volume mode" and you scroll again.... c'mon.... used that on stage - doesn't work. It has nice (meant-to-be) features such as multitrack recording to a SD card and local injection of analog signals which can then be sent in the AVB based network - nice! too bad the gain knob of the analog inputs is so badly placed that whenever tweaking something, the gain is off and every musician will blame the sound engineer, just to find that the gain knob (I should say "trim pot style"), was altered - thank you very much when changing the gain in a +20dB scale while wearing closed in-ear systems. Elite Core's system is great in this aspect (albeit the diagonal layout might be confusion and the pan knobs might look like volume knob) - a signal presence led helps to be sure that the channel is the channel of interest (for the system without a screen). Avioms take on this is OK but a signal presence led (or a mini meter bridge) would have helped along with a channels names. The newer units offer that by putting an iPhone on top of the mixer but the physical layout will not match and administering the channel names, groups etc, is a pain. Hearback is nice - one knob - finished! However the Limiter knob or the stereo channel knob can look like a regular channel, but it's not. Hearback is actually very good on audio quality, takes analog and ADAT (44.1k and 48k) and is incredible in driving low-impedance headphones - this thing is loud, and dynamics are very good. Price is excellent. However, cabling is not (no Ethercon connectors which (including AVIOM A-16 series) is a pity these days).
Mobility is the ability to move and change something fast without hassles. Here, Ethernet-based products have an advantage as networking is easy to expand. AVIOM's design is harder to do because they require HUBs which also power the mixers - so they must be present and interlinked by a dedicated Cat5 cable to a previous hub in topology. The bulkier the box is, the more difficult mobility gets. A heavy and big mixer (such as the Elite Core which is very robust) has the disadvantages that on the new location space or boom stands must be bigger/heavier than a Behringer, Aviom, Pivitec product.
Retainability talks about how good something is "help up" AFTER getting it initially. Changing the sample rate from 44.1k to 48k (on digital injectable systems) while the entire band has volume turned up is something you definitely want to avoid because they will produce a loud "SHHSHSHSH" until the new stream is phase locked. Speaking of 44.1k - not all systems take 44.1k on ADAT for example. The ADAT injector of Elite core is one example - it's fixed at 48k (and is built on XMOS AVB and ADAT IP cores). Hearback was fine on both, AVIOM's AV-M8 (the only) 8ch (!!!) ADAT injector (which is fixed on CH9-16) is fine on both as well.
Some designs use iPads, iPhones and alike: Ok, but besides the costs, how healthy is the battery in 3 years, what about software updates and iOS releases backward-compatibility, what about mounting options, what happens when the customer drops it and the glass is broken, screens are too bright on low-light stages and too dim in open-air and sunny days. somebody accidentally uninstalled the application and you're on the field without any connectivity to reload the app and what about the lock code? Purpose-built is still king in my view!
If I were to design a personal monitoring system it would have:
- AVB based transport with PoE. Special care to be taken to meet the audio performance of the step-down converter for both hi and low impedance headphones. Being earth free of course.
- Ethercon connectors
- 16 volume dials with a signal presence led and an OLED scribble over each knob to show the channel name which is set by using REST API's over the AVB network. another bank of 16 channel can be made accessible by banking up/down.
- a big, semi-weighted master volume (on the analog domain: to avoid truncating, noise-floor, and giving confidence that turned down is really mute)
- completely stateless - nothing can be saved or restored. Forget those functions - I never saw somebody using it really. Instead, grouping (and not just odd-even pairs) must be done on the mixer directly.
- a volume knob with logarithmic taper curve (Elite Cores are linear, which is unusual), the OLED showing the bar graph of the set volume.
- functions such as lower the entire mix to create more headroom etc...
- no solo or mute, or if needs be then very simple and intuitive. Don't like the Avioms implementation of Solo/Mute access (pressing a button WHILE pressing another and/or double-tap a button etc...)
- a talkback mic which can be used to inject ambiance sound (Elite Core has this!!!)
- a limiter to protect hearing (yes for Hearback, Elite Core and new Avioms, Behringer (but poorly designed), no the old Aviom).
- both sizes of jack outputs
- TRS outputs for the stereo mix plus 2 direct outs (useful for sending the click to the drummer...)
- 3 Band EQ on the master (but not on the channels)
- Injector natively supporting OSX's AVB talker or a 19" rack unit that takes analog in (with plenty of option for sensitivity), ADAT and optionally MADI.
- configuration done on the web server of the injector (channel names etc...)
Systems I used and liked:
Aviom A-16II, Elite-Core, Roland, Hearback - all despite the weakness
Systems I used and didn't like:
MyMix (operation), Behringer (noise, dynamics and limiter)
Systems I know but didn't use and will probably not like:
Klang Technologies (cost and accessibility and mobility), Pivitec
Systems I don't know but look fine:
dbx, Allen & Heath
Some words about ANET. I can't disclose the details of the protocol as I might violate the IPR of Aviom Inc. But what I can tell you is that it is too simple to be true. Install wireshark, disable all protocol decoding (so you only see raw ethernet frames) and play around. It's similar to TDMoIP and so basic I doubt that this can be really IPR protected.
The packets are sent every 2 samples which lead to very low latency and still offers plenty of time to so ASRC and creation of the master clock. I created an XMOS based project that takes 16 channels from ADAT or I2S codecs and sends it in ANET form to make the receiver happy (44.1k and 48k). Meta-data (channel group etc...) is being set by an internal web page - so the hardware unit is a simple gateway without any user operatable components. I'd like to add MADI - but that's another project for it's own.
The simplicity of ANET also has a huge strength - it's guaranteed that it will work once the stream is there and latency is very low (around 50uS on the transmission side) and audio is Bit-transparent / bit-perfect. There is no reason why ANET can sound bad (even jitter will not harm it if in a certain window)). Behringer's Powerplay just send two streams of 8 ch AES50 and jitter is an issue here. I see AVB as the next generation transport technology. EtherSound and friends will disappear. Possibly DANTE also (closed-source and hardware dependency will make it hard to be adopted my innovative start-ups and the licensing model is just off).
Sorry for the lengthly post, I got asked to write something like this from a friend - I hope you find it interesting. Feel free to agree or disagree.