That's correct. For optical telescopes, you have to physically combine the signals in real time to do interferometry. The VLT can do optical interferometry, and can create baselines as long as 200 meters (reaching nearly 100x the resolution that the Hubble Space Telescope achieves).[1] If you just combine the images from different telescopes, without doing interferometry, you only get one of the benefits of a larger telescope (you can see fainter objects), but you don't get all the benefits (e.g., higher resolution).
Interferometry is much easier with radio telescopes. You can record the waveform at each dish, and then digitally combine the signals afterwards (there is specialized "correlator" hardware that is purpose-built to do this very efficiently).[2] That means you can put receivers anywhere (on Earth or even in space), as long as you are able to synchronize the timestamps in the data well enough to digitally combine the signals afterwards. One of the major technical challenges faced by the Event Horizon Telescope, which imaged the black hole at the center of M87,[3] was synchronizing clocks at stations on opposite sides of the Earth. They had to physically bring atomic clocks from one location to another. After they took their observations, they flew hard drives with the recorded signals to a centralized location to do the correlation.
The article on interferometry implied that this is "just" a matter of computational difficulty - the higher frequency means that more accuracy and computing power is needed to resolve an image. Is that right?
If so, surely this is a Moore's Law problem, solved by waiting a few years for the needed computing power to be cheap enough to use?
By higher frequency, do you mean optical frequencies? If that's what you mean, then it's not just a matter of computing power. It's a matter of optical detectors not actually measuring the phase of the incoming photons. If you lose the phase information, you lose the ability to do interferometry.
CCDs just count photons (incoming photons kick electrons into the conduction band, and you count electrons after each exposure). There are more advanced detectors that tell you the energy of the photons, but not their phase. If you can't measure the phase of individual photons, you can combine light from different telescopes directly, and let nature do the correlation for you. That's what optical interferometers do.
Interferometry is much easier with radio telescopes. You can record the waveform at each dish, and then digitally combine the signals afterwards (there is specialized "correlator" hardware that is purpose-built to do this very efficiently).[2] That means you can put receivers anywhere (on Earth or even in space), as long as you are able to synchronize the timestamps in the data well enough to digitally combine the signals afterwards. One of the major technical challenges faced by the Event Horizon Telescope, which imaged the black hole at the center of M87,[3] was synchronizing clocks at stations on opposite sides of the Earth. They had to physically bring atomic clocks from one location to another. After they took their observations, they flew hard drives with the recorded signals to a centralized location to do the correlation.
1. https://www.eso.org/public/teles-instr/technology/interferom...
2. https://www.aanda.org/articles/aa/full/2007/05/aa4519-05/aa4...
3. https://eventhorizontelescope.org/press-release-april-10-201...