The universe is full of merging black holes, but in the vast majority of cases, the product is too weak for us to detect. However, two physicists think they have found a way to enhance the power of existing gravitational wave detectors to pick up faraway mergers we currently miss, unleashing a new wave of discoveries about the nature and origins of black holes.

In 2015 we detected gravitational waves from merging black holes for the first time. The collision caused an eddy in the fabric of spacetime that spread out through the universe until it gently shook the Laser Interferometer Gravitational-Wave Observatory (LIGO) on Earth, almost a billion light-years from the source.

It was one of the discoveries of the decade, until it was overtaken by last year's merging neutron stars. In the 30 months since we have detected a total of six. All of these were quite close by, astronomically speaking, and in terms of the size of the universe, it is estimated that mergers occur once every 200 seconds.

When black holes merge they produce a set of waves with distinctive frequency patterns, which, Dr Eric Thrane of Monash University told IFLScience, could be converted to sound waves to produce a characteristic whoop. However, unless the merger happens quite close by, cosmologically speaking, the waves are so subtle they get lost in the gravitational noise of the universe, like trying to pick up a signal on a radio dominated by static.

Extending the analogy, Thrane asked us to imagine listening to a radio when beyond a station's normal range. “You might not be able to work out what song was playing, or even the genre, but if you listen long enough you'll pick some things up and be able to work out the style of music even if you never hear a complete song,” he said. “If your brain can do this there must be an algorithm, so we came up with a statistically rigorous method for listening to the radio of the universe.”

After feeding the algorithm with enough test data, Thrane and colleague Dr Rory Smith were able to detect simulated mergers, and now they are keen to apply this to real observations from the network used to pick up those we have already found.

In Physical Review X, the pair predicts their new method should be a thousand times more sensitive to distant mergers than existing ones.

The black hole mergers LIGO can detect are those with masses a few times that of the Sun to 100 or so solar masses. Larger black holes, such as those at the center of galaxies, would produce a lower frequency when merged, one far too deep for current instrumentation to detect.

We still don't know the reason for these events, Thrane and Smith added. Most astronomers think they come from very large binary stars, where first one and then the other experiences a supernova explosion and turns into a black hole, with these two circling each other until their orbits decay.

However, an alternative theory holds that most mergers come from globular clusters, where stars are packed so tightly that, after the larger ones turn into black holes, random movements can induce collisions. An even more exotic theory is that primordial black holes remain from before the first stars, and these sometimes run into each other.

Thane and Smith's work may give us a sample size large enough to determine which of these processes dominates. They plan to seek this using the recently launched OzStar supercomputer, which replaces central processing units with graphical processing units, hundreds of times faster for these applications.

Simple adjustments to the technique should make it applicable for finding more neutron star collisions, the authors believe. Going further, they expect it could eventually help us witness the waves released at the birth of the universe. “Although the gravitational waves from the Big Bang would include some wavelengths the size of the universe, the spectrum should be very broad,” said Smith. “We can say with some confidence that when the period of inflation ended it should have created a wide frequency range, including at those we can detect.”

This would be, as Thrane and Smith say, “one of the most amazing things to detect.”