Do I really need an audio interface?

There are (in my humble opinion) three main considerations when speaking about audio interfaces. Those are: noise, distortion, latency. Many PC-inbuilt devices (e.g. AC'97 and relatives) are just awful in the first two disciplines. However (although without speaking from personal experience) Mac audio interfaces are supposed to be a whole lot less bad. One probably still benefits from a good audio interface, but maybe not as much. On a PC notebook, it is usually not even a question, though...

Latency is largely not an issue anymore, IMHO, at least when only playback is concerned. That is, roundtrip latencies (recording to playback) of less than 50ms are nowadays achievable even without special drivers like ASIO. The newer Windows architecture(s) as well as the Mac OS audio architecture and ALSA on Linux allow for reasonably low latencies; in fact this now works so well that for Windows an ASIO emulation layer exists that sits on top of WDM kernel streaming (ASIO4All) and makes native ASIO drivers obsolete most of the time. The lowest achievable latencies differ depending on the device and its drivers, but values of 5-10ms (or 10-20ms roundtrip) are possible in most cases, and this is usually squarely in 'fast enough' town.

Not quite true. There is software latency, driver latency and hardware latency. For example, I normally use a buffer size of 128 samples at 48kHz in Pianoteq. This corresponds to a software-induced latency of 2.7ms. The total latency of my setup (Zoom R24) is closer to 10ms, however, leading to a total round-trip of ca. 20ms when used in a DAW setup. The reason for this is that the driver increases the latency somewhat, and also the R24 itself has an internal playback buffer (of probably about 128 samples). This is necessary since e.g. USB is no real-time bus protocol and introduces jitter and latency of its own. For the same reason the really, really high-end stuff used in scientific setups (e.g. for accoustic phase measurements, where absolute timing consistency and lowest latency are vital) is still PCIe, as far as I know.

(Side note: the limited and often fixed internal buffer is also in my opinion the reason most USB audio devices don't really play nicely with USB-hubs. For example, I currently can't plug the R24 in directly, and this leads to crackle-free 96kHz operation being more or less impossible even when using huge buffers on the computer side. The probable reason for this is USB timing jitter reaching the ms range.)

Easiest way to get the right answer to your question:

. . . Buy the necessary cables to connect your speakers to the Mac headphone output,
. . . and plug them in.

If the QSC 10's sound as good as the headphones (which make you happy with their sound), you will have good reason to believe that you _don't_ need an outboard audio interface.

If you say "Yuck! They sound awful!", you'll know you _do_ need an audio interface.   Or at least, that it's worth trying one out.

I think that Mac's have a reputation for better "built-in sound" than most PC's.   

.           Charles

Sound quality is not just noise. A good audio interface can also sound differently due to less distortion. It will probably show less aliasing and intermodulation artifacts as well as lower latency and better stability.

PS: WASAPI shared mode and latency being "not an issue" is something I find quite hard to believe, actually. Whenever I tried shared mode in any software, a noticeably increased latency was the result, regardless of the actual interface used. Also shared mode sets the audio interface frequency to a fixed value, which can lead to increased aliasing distortions (I seem to remember there was a bug in the software mixer at least up to Win7).