Відео

Hi Mr. Koenig, I am an IB Physics student trying to learn more about spectral analysis. May I ask what program you used to analyze sounds for this series? Thank you!

Is there a plug in you can recommend to get the pitches for bells?

Hi Maggie. It will be easier if you reference the PTJ article. I can answer you mote clearly that . ..and it will be easier if you use my email which i think i sent you.

Great video! Since piano techs are your audience, maybe explain that the "second partial of y" is a partial derivative, not a partial of the piano note (fundamental, second partial, third partial, etc.)

This is blowing my mind. When I tune the bass aurally, I tend to not only follow the railsback curve, I tend to push it even more flat. That's because the 2 mentors I admire the most taught me to keep the beat rates closer to the same as I move down instead of getting really slow. They still slow down, but in a barely noticeable way. Am I tuning my bass way too flat?!? I think I don't understand the difference between the "first partial" and "pitch". Pitch of what?

1/9 to 1/7 is the projected ideal for strike point in order to minimize those partials.

Any recordings of this beautiful piano?

We'll never solve this problem. In tune perception versus resonances is impossible. You have only three options. 1) make all strings of the same diameter and mass per lenght, which means your piano will be about eight meters long. 2) build a sound board that compensates for equal tuning. This is science fiction for now. 3) Make a compromise. Which also means to tune the piano according to the repertoire. Option three being the only reasonable one, it is a pain, but it somehow works. How lucky are these fretless instrument's player, as they can compensate (at the price of playing only a couple of notes at the same time). Reminds me of my comment to a sax player : "come back to me when you can play a chord".

Very clear.thank you

This video is grate for understanding Autocorrelation! I will use it in a projekt.

very good.

But why is the highest string struck so near the top end of the string?

So basically a string is excited with a triangular function and rings with a triangular wave?

Amazing, wow, just what I was looking for.

this is simply amazing

Thanks for busting a myth.

BTW - Does not detract from the "Angel Shot Voicing". As usual with pianos and piano technicians, the techniques used to improve tone are quite practical and useful even if the explanations as to how they actually work miss the mark from a physics standpoint.

Great question from Ed and clear response.

Clear. Well done and explained David

Thank you so much! As a "merely interested" random internet person (officially computer science student), this was extremely helpful! I thought "huh, cool" after the last video, but I feel like this has given me an actual understanding of many details and nuances in it. Looking forward to the next "lesson"!

Could you please share the audios used in the video? For example, the A4 tone. Thank you so much!😀

Seems like the force exerted on the string by the motion of the bridge would be an important inclusion. Are results different if the string is off centre from the bridge?

I do not understand your question. Try email: spectraldmk@gmail.com

This happens on any sound wave? The harmonic series is increasing over the spectrum?

Hi Scctt. Yes, I tried that. There just isn’t enough data.

Good David. Perhaps (if there is enough data) you can also check the spectrum of the precursor section to see if the frequency is a sensible match to your frequency calculation and if, in fact, it matches the spurious peaks on the overall spectrum of the string. May work better lower in the scale.

Very good David. Would not be unusual to not see a lot of C1 fundamental as this should be lower than the 1st modal frequency of the fully strung soundboard and thus not transmitted well. Even more so on the 2nd C1 sample. I assume that second sample is from a much smaller piano. Suggestion to record a rubber hammer hit or even a fist hit on the center of the soundboard and check the spectrum to find this fundamental board frequency of each of the instruments. Then you could also check the notes near this fundamental board frequency to see how if affects those notes.

Thank you!

Really stunning illustrations and explanations. Superb analysis. Thanks!

Great video! I'm looking forward to part 2.

Excellent introductory video, David. Through my speakers, the sine waves you played don't really sound like pure sine waves and they were louder than your voice. Thank you for all your work. Looking forward to the next videos.

Thanks for the comments. Glad you enjoyed the video.

Thanks, that was very interesting. So did the partialsounds come from the ripple effect?

can you send a link of the all the recorded musical notes for matlab I couldn't find it on the facebook page

Please I want the file Matlab that you used in this video and thank you

Personally, I try to tune a piano to a Railsback curve, which BTW gets steeper the smaller the piano because inharmonicity is directly related to string stiffness, and the only way to get decent low tones out of a small piano is with really thick strings which are then also stiff. However, there are so many soundboard resonances under ~200 Hz that sometimes notes need to be tuned to avoid either vanishing on a frequency the soundboard doesn't care for (at least from that particular location), or to avoid jumping out as being louder than the others around it. This is why the bass notes deviate significantly from the Railsback Curve, even when attempting to follow it.

Helpful script for data collection

Does the bridge in this model have additional mass and stiffness or is the "bridge" simply referring to the point on the board that the string is driving ? So, for example, do you make the board thicker in the area that represents the bridge?

Wow this is mind-blowing!

Hello David, I have a sound track from an experiment and there is a constant hmm noise in background, how can I remove that noise from my tracks using matlab.?

Jo

where is the script that you are running (AnalyzeAWavefromFile.m) ????

Hello David!. These scripts are really great!. Do you know if there is a way to analyze every harmonic separately in order to check the amplitude in time of each one of them?. I am trying to emulate sounds by additive synthesis and knowing that would be awesome. Thank you very much, this video was really helpful! Greetings from Argentina

Isn't that the lower frequency components travel faster?

Nice work! Looking forward for more sound examples. However, i think comparing real piano with only a string model is not fair at all. In real piano you have body response and many other coupled strings that affect each other, strings interacting directly with air (not through bridge), etc.. But i think those videos really helpful to understand what's going on here from mathematical point of view. One note - autocorellation peaks are not equal to pitch. It may be related to precieved pitch, as for harmonic signals, but in case of inharmonic signals AC peaks are much less useful. Unlike fourier, it's also nonlinear, which only complicates comparisons with real strings.

Well done David. Not sure what you plan next, but it would be informative to see an expansion of what happens around the 14 minute mark. You could repeat the comparison of two situations theme, but do two versions with different hammer elasticity to explain how that changes the force profile on the string, wave shape and thus harmonic content. Same with hammer mass.

Do you impose a time envelope on the sum of pure tones?

A fantastic contribution to our understanding the underlying mechanism of inharmonicity.