I've noticed with a both a needle-based Korg tuner, and with Tunelab97, that when a string is played on my piano, that the frequency which is read (be it the fundamental or some partial) drifts slightly through the sustain. From the beginning of the reading, to three or four seconds later, the frequency of most keys drifts up about 2 cents. This happens on many of the keys. Some are much more stable, and on a very few, they drift down. The up and down drifts can happen on neighboring keys. So it's not a case of the down-drifters are confined to one part of the scale and the up-drifters are at another.

The strings are in good visible shape, about twenty years old. No rust or any other form of apparent problems. But I've been noticing this for about as long as I remember. I always thought it was a problem with my Korg device. But I just started using Tunelab97 and it's showing the same thing. So I believe it's real. Is this a common trait of piano strings? Is it indicative of some known malady of the particular strings on my piano? Has anyone else observed this? I don't tune many other pianos, but I have seen it on others as well.

Did you get the same result when you pluck the string. Use a guitar pick.

I liked your work but it looked as if the hammers were raked back way too much. Most grands only rake the hammers one or two degrees.
Was it the picture?

That´s normal. On every piano. Particularly in the bass. And there is a difference when you pound the key hard or soft. Pounded hard there is more deviation.

Gregor

Keith,
I haven't tried plucking the strings. I will. But then Gregor say's the drift is normal. So maybe the drift is not really indicative of something wrong.

On the hammers being raked back - You're probably referring to the below picture:



Don't put any stock in that. It was just a spare shank and a spare hammer I had laying around that I threw together just for the photo not realizing the slant on the hammer. It was definitely bored incorrectly causing the apparent rake. That's why it was a leftover spare. I realized the rake in the photo after I had put all the equipment away and didn't want to go through the hassle of taking it all out and setting it up for the photo again. The hammers I actually installed in the piano were carefully checked to ensure that they were bored accurately and then installed at 90 degrees using Spurlock's grand hammer hanging jig. Let me know if there was some other reason you thought the hammers were raked back.

I think a certain level of drift is normal and I think it has to do with the mass of the string. When you pound a thick bass string very hard, then it takes a little bit of time until the string "calms down".

But in some pianos there is a lot of drift even in the middle section or in higher notes and on some keys more than on the neighbouring keys. No idea about that. Perhaps string seating helps?

Gregor

Had me fooled.

If for some reason the hammer isn't rebounding nicely, it could inhibit the initial sound and there is a delay as the string energy comes into focus. Always determine if the problem is the string and terminations or the blow delivered to the strings.

Of course plucking the strings is the easiest and quickest way. Does anybody have any other tests?

Keith,
Piano strings are fairly linear devices. Initial transient effects in linear systems tend to die out pretty quickly. As a pretty neat demo of this, here are some waveform images of some of the keys in the temperament of the same piano I captured with Reaper, which I described in my discussion of tuner training software. Note that these are recordings of true single piano strings, not simulated. The name of each track, at the very left of the track just above the volume adjustment sliders, are the names of the notes represented in each track.



The top of the image shows the time axis in seconds, so the whole graph is only about 2.4 seconds long. You can see that the transient effects vary from string to string, which I don't understand. May be poor rebound from the string on those notes, as you mentioned. But those effects, whatever they are, are only less than a tenth of a second. The worst shown here is on F4, the 6th trace from the top. You can also see from the traces that the higher strings decay at a much faster rate than the lower strings. Pretty dramatically so.

And just for ya-has, heres some of the waveforms zoomed way in, so that you can see the actual time representation of what the microphone picked up.



You can see from these that some of the keys have a lot more higher partial content than the others. Possibly because I've never voiced the hammers since I installed them. Some are significantly quieter than the other as well. I noticed this happening when I was recording them, although I couldn't detect much difference in volume through my ears. Don't know what to say about that.

The bottom set of notes are not all the same as the top set but most of them are. Of the ones that are common to both sets, you can see that the ones with the strongest transient effects in the top also appear to be the least well-behaved ones when you zoom way into them in the bottom set. Pretty cool, huh? I wonder if the ones with the stronger transient effects could be because of poor string terminations, or perhaps loose hammer flanges, or who knows what?

I'd say that there is probably a lot that could be learned by studying these things. I just started looking at them in the past week or so. It will be interesting finding where this brings me, if I stick with it.