The Yamaha drive units are excellent, but the dividing network is not (bipolar caps! Ugh!).
These speakers will soon be forty years old. So when I got mine I opened them up, took the crossovers out, disconnected the caps and measured them. They're fine - unusually accurate capacitance values and very respectable ESRs. There's nothing wrong with them. But if you object to them on religious grounds or because you don't like the way they taste (Ugh!) then you will definitely need to replace them.
I have heard the NS-1000 that Troels Gravesen made a new filter for! Works MUCH better than the originals.
Here's the speakers' behaviour as measured by Troels before and after he made the crossover changes
You can read the full story on his website
http://www.troelsgravesen.dk/Yamaha-NS1000.htm which is where I pinched this graph from.
I'm afraid I can't find anywhere which curve is 'before' and which is 'after'. I tried to compare them with other curves he shows, measured under different circumstances, but I'm afraid the difference between the crossovers is smaller than the difference made by moving the speakers or by putting them on stands or by moving the measuring mike. Whether you call this difference MUCH better or not is again a matter of taste.
The beryllium drivers are fabulous, but the woofer is a weak point as there is not much going on below 50hz.
Well, Troels says it's 3dB down at 50Hz. But it's not dead. He also points out that you can compensate this by room positioning which I've done with mine. I agree that the woofer isn't nearly as special as the other two drivers though. It's also worth pointing out that the woofers in the NS1000 (which I've got) and the NS1000M (which, from the pictures, seems to be what Troels worked on) have different Yamaha part numbers. One obvious difference is that the M versions have a grille whereas mine don't (Troels grumbles that the grille made it very hard for him to make near-field bass measurements). But there may be other differences too. The M version of the speaker is 8kg lighter than the non-M. Some of that will be due to the cabinet. But maybe some is due to the woofer magnet too (pure speculation though) ?
VB