... do speaker cables and interconnects need to be the same brand?
To add to the previous responses, with which I agree, the idea that the cables in a system should all be of the same make is often referred to as "loom theory." While a goodly number of audiophiles subscribe to that belief, and many of them have achieved satisfactory results following that approach, I daresay that the majority of audiophiles have achieved satisfactory results without following that approach. And as I see it "loom theory" is fundamentally flawed in that it fails to recognize that the sonic effects of a particular cable are dependent not only on its intrinsic characteristics, but are also dependent (and in many cases are more dependent) on the technical characteristics of the components the cable is connecting.
For example, the sonic effects of an analog interconnect will depend to a significant degree on the output impedance of the component driving the cable; on whether the cable is balanced or unbalanced; on the susceptibility of the components that are being connected to ground loop effects; on whether the signals being conducted are line level or phono level; if they are phono level on whether the cartridge is a moving magnet or a moving coil; and on other variables. While the sonic effects of a speaker cable will depend to a significant degree on the impedance of the speaker; on how the impedance of the speaker varies as a function of frequency; on the criticality of woofer damping to the particular speaker; on whether or not the amplifier utilizes a feedback loop from its output; on the sensitivity of the amplifier to spurious energy that may be introduced at its output and couple from there into that feedback loop; and on other speaker-dependent variables.
As evidence of the dependence of the sonic effects of a cable on what it is connecting, here are a couple of relevant examples:
1) If an interconnect having relatively high capacitance is compared with one having relatively low capacitance, and if everything else is equal, the higher capacitance cable will produce a duller and more sluggish response in the upper treble region if used as a line-level interconnect (especially if it is driven by a component having high output impedance), due to the interaction of cable capacitance and component output impedance; while in many and probably most cases the exact opposite result will occur if those same two cables are compared in a phono cable application and driven by a moving magnet cartridge, due to the interaction of cable capacitance and cartridge inductance.
2) It is easily possible for digital cable "A" to outperform digital cable "B" in a given system when both cables are of a certain length, and for cable "B" to outperform cable "A" in that same system if both cables are of some other length. The happenstance of the relationships between cable length, signal risetimes and falltimes, cable propagation velocity, component susceptibility to ground loop-related noise, the happenstance of how closely the impedances of both components and the cable match, and the jitter rejection capability of the DAC, all figure into that.
Given these kinds of component-dependent variables, happenstances, and dependencies, it is hard to conceive of how, as a general rule, a single-manufacturer loom would necessarily stand a greater chance of being optimal than a mixed set. And for that matter, given that the components in a system perform completely different functions, are very different in design, and are usually produced by different manufacturers, a mix-and-match approach, if pursued with reasonable thoroughness, would seem more likely to do so.