That's it. The graph scaling is awful, but you can see that it does .006% at 1kHz, with a small rise to about .03% at 20kHz, at 30W. At 100W, .003% and .018%. At full power it is even less, at least at 1kHz. Thus, it seems that distortion keep falling as power is increasing, which is what you want to see. Yes, 20kHz THD could be a little lower, but with over 350W available per channel, there is not much to complain about. This is better performance at 20kHz than probably 80% of the "high end" products out there, meaning that the design is more linear than a great many of them. Also, the fans apparently don't come on unless it is driven extremely hard. And if you really want lots of power, buy a p5000s or a p7000s.
There, amp shopping done. I am not sure if their newer series do as well. I haven't looked at the design. Many of the manufacturers are going all in for the weight savings and sacrificing distortion performance in the process. It's also worth noting that most pro products do not perform this well. Don't go thinking you can snag an old QSC or the Behringer copy and have behave this well. You can't. Products that ever performed like this were rare, and most are no longer in production since the landscape has shifted to higher power and lower weight. People went from schlepping 90 pound amps, to 40 pound amps, to 10 pound amps.
I decided to follow this up with a little hint to a pro amplifier even better than this.
One thing that makes the Yamaha perform is that it uses, if I recall, a discreet dual differential, fully complementary front end, which is not all that often seen in cheap PA amps. I suspect it might be let down some, though, by its use of a rather overly complicated output stage, Yamaha's "EEEngine" which is sort of a Class D style amp layered on top of a Class AB amp. The idea is to make the rail voltage track the input signal, like some of the old Carver amps. Weight reduction.
If you just want a standard Class AB amp, there is one (and perhaps only one) that is inexpensive and has a state of the art traditional AB design (that isn't really spendy). Sometime in the mid 1990s, a rather well known pro sound company decided they wanted to sell power amps. So they let one of their designers run wild. Instead of copying the fairly cheap to produce Crown or QSC designs, they engineered what was arguably one of the more technically advanced products of its time. This was before the "phone an order to China" days, and the products were USA designed and assembled. Compared to the cost of other products at the time, this company still managed to do it cheap by using surface mount and plastic output devices, which legacy amps of the time did not. Somehow, I mentally filed away an old Arny Krueger (RIP) rec.audio post from 20 years ago praising these amps, which I remembered a couple of years ago when looking for cheap amps that might perform well. I pulled the schematics and service manuals, and ... there it was. An amplifier that had no reason to exist, that shouldn't exist, and would never be designed today, but for some reason made it past the bean counters.
Some fairly simple analysis reveals that this amp has perhaps the best low distortion circuitry ever designed and put into any mass market pro audio amplifier, and it bests most "high end" products to this day. It is chock full of distortion reducing principles which few high end amps follow, and almost no pro amps do. Dual differential, fully complementary front end, with an extremely good performing transimpedance (or VAS) stage, lots of cascoding, a very clever compensation scheme, all using parts that were among the absolute best ever designed for the purpose.
At 1kHz with these amplifiers, you should see at or under .002% THD at full power, with a rise to maybe .007% at 20kHz,
at full power of 250w. Back it off a little and that should fall to .004% or less. If that doesn't sound that impressive, look through every measurement of every amplifier on the entire internet. Almost none come close to that level of linearity. Bryston, Cambridge, Benchmark, and perhaps Arcam are the notable exceptions. Even at 20kHz, the line for THD+N just keeps heading downward along with the noise floor, and only starts to flatten on when the amp is about to clip. Happily, the company included Audio Precision measurements for some of the amps that show this textbook performance (probably taken on an old System 1). Today, you can buy these things for $200 used all day long from Guitar Center or Music Go Round.
You'll need to do a low speed fan modification, since they do have a fan. Also note that the noise performance at low power isn't as good as some of the new Bryston stuff. But that could be the input circuitry, since there is effectively a preamp circuit in line at most times. If you tacked on a pair of RCA jacks directly feeding the amp section, I'm sure things would improve.
Oh yeah. The amps are the dime-a-dozen Mackie m-1400. The manual specifies less than .012% THD at any power level or frequency into 8 ohms. That was conservative. Here's a link to a dozen of them for sale for about $220, shipped.
https://www.musicgoround.com/products/PSPA/power-amps?search=mackie m1400&Brand=mackie&page=1. Just get some 1/4" jack to RCA adapters or use the balanced inputs and you're off to the races, since they do have binding posts instead of Speakons. I haven't actually measured one, but there's no reason they shouldn't perform as specified, and they probably won't burn up or blow up after this many years.