It's not cheap though, all the bloody carvers get it. Not that any wood is cheap at the moment.
I think they might have just been highlighting basswood because it's so soft —softer than many softwoods— and so the outcome shows a much bigger improvement.
That was my first thought too - maybe basswood is cheating. It does look though after some research that it seems that the process can be applied to other woods as well, since the process consists essentially of removing the non-cellulose component of the wood (hemicellulose and lignin) with heated chemical solution, then heat-pressing the resulting cellulose-only matter into hard block material.
Naturally, the harder of the hardwoods (IPE, Brazilian Teak, Ebony) are also relatively low-lignin wood types (that also grow relatively slowly, sensitively or in unfavorable geographical regions for logging and transport) and the result of removing the non-lignin would logically yield a lower improvement factor (and possibly take longer).
In general, the difference between softest readily-available lumber (such as basswood/spruce) and hardest (IPE) is about one order of magnitude. The result of this study at 23x means basswood can be made more than twice as hard as the hardest hardwood that is reasonably available. It would likely take a lot of lumber weight input though (explained below).
To your comment: Most common timber/lumber woods (softer: spruce, red pine, fir, chestnut, tamarack/larch and medium: cedar, maple, oak, birch) are rapid-growing and have established forestry industry around them. If you were to take white pine or spruce it should yield similar results to the study since you're basically condensing it to cellulose and they have similar weight densities. You would need to also factor the density of the wood since yield would be ratio of cellulose * weight of the source wood.
Since this is a high-waste process (only 40% of the weight is kept in the final product which is then compressed to the target density of 10000lbf or so) it would probably make most sense when using waste-wood as input (wood chips, sawdust, offcuts, recycled wood) and not on viable timber. This is similar to LVL and OSB (although it uses glue for binder)
Some composition comparisons:
Nordic Spruce: 39.5% Cellulose [1], 0.43 kg/m3 [5] (the poster child for engineered lumber construction in europe)
I think they might have just been highlighting basswood because it's so soft —softer than many softwoods— and so the outcome shows a much bigger improvement.
Show me pine/spruce, poplar and oak.