by Pablo Bellinghausen –
In our previous instalment we saw the basics of what constitutes an audio interface, or sound card, as well as the kind of audio connectivity they can offer. This time we will look more at the digital side of these devices, to see how they interact with a computer.
Digital connection and compatibility
The first models in the market were meant to go inside the CPU of a computer tower, plugged directly into the motherboard. This was due to the relatively high bandwidth of the old PCI standard compared to USB 1. With the advent of USB 2 and 3 as well as Firewire and now Thunderbolt, speed is rarely an issue, although the largest interfaces will still tend to be PCI Express. The Apogee Symphony I/O range and the industry-standard Pro Tools HDX systems have a configuration of up to 64 input/output channels at the highest sampling rate, with the largest Pro Tools systems maxing out at almost two-hundred channels at a lower rate. Maximum track counts will in general depend on the sampling rate, since 192 kHz audio will require over four times more bandwidth than the CD-quality 44.1 kHz rate.
Symphony I/O converter box and internal card: some of the best sound available
Thunderbolt is the latest connection to come onto the market and the best external choice whenever very large track counts are required with the smallest possible latency, since its speed is comparable to the best internal connections. It is however an expensive option for more modest requirements and therefore an unusual sight on small devices, although ultra-low latency is one of the selling points of the Zoom TAC-2.
Another option (although a rare one) is to use the Ethernet port as a digital audio connection. The Focusrite RedNet systems can be plugged into the RJ45 port directly, but internal speeds can vary wildly depending on the type of port the computer’s motherboard uses, and can be unstable or add latency in some models; this is why for larger systems the RedNet system requires an internal PCIe card to send their proprietary Dante protocol through an Ethernet cable for a maximum of 128 channels of I/O at 96 kHz at very low latencies.
Before the advent of USB 2.0 the best external connection used to be Firewire, which is still theoretically the fastest and most stable choice out of the two. The chipset does many of the calculations required to route the audio so Firewire will use less processing power than USB, although that isn’t an issue with the latest generation of computers. The connectors come in three sizes: 4-pin and 6-pin Firewire 400, and 9-pin Firewire 800, which as the name implies has exactly twice the bandwidth. All three are inter-compatible by using the correct lead.
The 4-pin connection, which doesn’t carry any power, is often used in video cameras but very rare in the audio world, although it can be seen on some PC laptops. The 6-pin version is the most common one and can be seen throughout the Focusrite Scarlett range for example. Some higher-end models such as the RME Fireface 800 offer both 6-pin FW 400 and 9-pin FW 900 ports, but this is mainly for convenience since they still only work at 400 speeds.
The vast majority of models in the market are currently USB 2.0, which works great on Mac and PC systems alike. USB wasn’t originally designed for time-critical, constant data, but manufacturers have worked around the limitations and the latency and stability of their drivers can now be up to par with Firewire.
Some USB 1.1 interfaces are still being manufactured but the protocol is too slow for even one stereo input/output at high sample rates, so they are now pretty rare. USB 3.0 is the latest version, and some interfaces such as the RME MADIface XT do use it for very high track counts at a relatively low cost, but it is still a rare option since it lags behind Thunderbolt in reliability.
The last thing to consider with regards to the interface’s connection type is how the device will be powered; although USB 2.0 carries 2.5 watts of power at 5 volts, internal circuitry can drive up the voltage to provide a couple of channels of 48 V phantom power without having to plug the device to the mains. Most 2-input cards will indeed be bus-powered, but in order to get the most reliable and clean power, high-end preamplifiers used in some premium interfaces like the Focusrite Forte or the Audient ID22 will require to be plugged to the mains, at least when providing phantom power.
The Audient ID22 needs mains power to operate, but the sound quality is more than worth it
It is important to note that at the very bottom of the price range some interfaces will not always provide full 48 V of phantom power, some sending 24 or even as little as 12 V from the preamps, often despite their marketing blurb. Whilst many condenser microphones will still work, the sound quality can be compromised, and in a worst-case scenario you might even hear a high-frequency whine or distortion. The power supplied by USB ports can be unreliable and is susceptible to electromagnetic interference, which has been known to add unpleasant digital noise to audio signals. This is a rare issue but when it does crop up it is almost impossible to fix it, short of using an AC adapter or a powered hub.
To DSP or not to DSP
One of the features that can have the biggest impact in the workflow, should you choose to go that route, is the inclusion of a processor with built-in DSP (digital sound processing).
The basic models offer monitoring effects for musicians without straining the computer and with the lowest possible latency. Models like the Tascam US-366,the MOTU Track16 and the Focusrite Saffire Pro 24 DSP allow you to select compressor, EQ and reverberation effects to make the performer as comfortable as possible; this is particularly useful for vocalists, who often prefer a bit of reverb to hear themselves better, which can help their intonation and overall performance. Although these effects can sound pretty nice, they aren’t really meant to replace the internal software plug-ins and the signal is recorded without the applied effects.
The Tascam US-366 offers on-board digital effects for monitoring to help singers perform at their best
Another type of sound card uses DSP to add some flavour to the recording before it reaches the computer; for example the Focusrite Liquid Saffire 56 has two “liquid” mic preamps, which emulate famous and venerable preamps in order to creatively shape the sound.
High-end DSP however is meant to replace the computer’s processing, and can use some of the best-sounding digital algorithms currently available; great examples are the aforementioned Pro Tools HDX systems, as well as the Universal Audio Apollo range, which boasts critically acclaimed sound effects as good or even better than the top native effects; the processing is done inside the unit but the front-end will work exactly like a native plug-in.
The Universal Audio Firewire/Thunderbolt card with DSP: a powerful combo and a fantastic sound quality
As computers have become faster this option has become less of a requirement for large studios as it used to be; the latest computers are so powerful that the most expensive processors are often overkill, and even a standard laptop can deal with very large and complex projects. This has seen a shift from external DSP to computer-based plug-ins, which is why Avid (formerly Digidesign) has now issued a fully “native” version of their HD range of interfaces, appropriately called Pro Tools HD Native. All the processing happens inside the computer but the digital algorithms are basically identical to the external card effects.
Avid HD OMNI and HD Native Thunderbolt interface: a new take on the industry standard
Stay tuned for our third and last instalment on audio interfaces, in which we will discuss what a higher price tag will get you in terms of sound and build quality, as well as discussing mixers and some of the myths that surround audio interfaces with mixing facilities. Until then, happy recording!