Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards Kawai SK grand samples - strange noise in the bass region.

I can always hear it on the piano (via speakers and headphones, although it is hard to hear via the speakers). When I record on the piano and then play the file on my PC (with the same headphones) I can not hear it anymore. Whatever, I got me an audio cable and will provide a piano recorded sample and a headphone jack recorded sample later today. I hope I can reproduce what I hear and maybe resolve the discussion.

This noise issue, IIRC, has been discussed on this forum before. The noise appears to be emanating from the pre-amp/amp, which means it should be there regardless of the choice of samples. It is possible that the EX sample set, which clearly has higher amplitude upper partials, may be masking the noise to some extent.

I'll do an analysis of your recordings when you post later. It should show the noise.

Here is the soundcloud link to the recordings - I hope quality is OK. At higher volumes I think the noise is easily recognizeable.

Test recordings
(SK Concert Grand with Brillance set to +10)

By the way, thanks to all of you for your helpful replies!

edit @prout: Just saw your post with explaining how I should record - missed that. Can do another set of samples if needed.
edit2: Uploaded another set of examples: Test recordings v2

That makes it sound like something in the amplification, or DAC section of the workflow. The digital sample itself is more or less clean, but when it is converted to analog and amplified for speaker or headphone output, something is either introduced or exacerbated.

I've just tested this as well on a CA67, and first of all, I want to confirm specifically what noise you're referring to? The only noise I can clearly hear, both on your samples and on my own instrument, is a relatively high pitched metallic sound which onsets something like 200ms (will need to wait until I have time to record and do a proper spectrogram, though I'm not convinced it would be terribly visible given its relatively low amplitude and multiple narrow frequency bands) and decays after perhaps 1000ms, and certainly more rapidly than the note. Is this what you're all referring to? Otherwise I'm not entirely sure.

If it is, then I can confirm it's definitely present, but to me it does sound very much like inharmonic overtones, both in the frequency range and the timing. I would also add that I can hear to a smaller extent even in much higher tones (e.g. D#6 has it quite prominently), to a smaller extent in the SK5 instrument, and to my ears at least, very prominently in the EX Concert instrument, which makes me wonder if we're talking about the same sounds? In the EX, it's certainly overtones, but that's also just a more prominent and clearly defined (in terms of frequency) version of what I'm hearing in the SK Concert instrument. Otherwise, could someone please describe the approximate time course of what I'm supposed to be listening for.

I listened to both sets of samples. Here are the graphs from your 2nd set of samples. Again there is no noise of any kind. Everything you see and hear is correctly structurally related to the note played and is a normal part of piano sound that varies from brand to brand, model to model, and especially varies with voicing of the hammers and humidity. I also checked the background noise between each sample segment as you changed from one recording technique to another. The noise levels are all low: -80dB at least and much lower of course with the direct to .wav file.





I listened again to your first set of samples. There is no doubt that the first of the three samples is brilliant, the second muffled, and the third (.wav I think) is, to my ear, the most natural. In no case though, is there any noise. The normal brilliance in the first sample is simply artificially enhanced upper partials.

This is an ambient recording captured with a cheap sony voice recorder. C#2 on SK-EX with + 10 brilliance. Listen to the last 4 notes.

https://soundcloud.com/user-350798041/sk-noise/s-cHlc7

Is this the same noise you're hearing, Hoizmichi?

There are many types of noise produced by a piano, many types of noise produced by electronic audio equipment, and additional noise added to the data stream by digital effects.

Unless you are 50 years old plus, you are probably used to essentially noise free audio reproduction. Even a 16 bit CD can achieve close to -96dBFS. Typical analogue recordings from the 1940 through the 1960s had noise levels at around -60dBFS equivalent.

Playing a chord on a very good acoustic grand piano produces a huge amount of Intermodulation Distortion. This sound is what identifies the instrument as a piano. The IM is caused by Equal Temperament Tuning coupled with the Inharmonicity of the struck strings. This is not noise in the normal sense of the word.

The sound of the hall or room in which the piano lives, also has its own background sound. This is background noise. If you record the sound of the piano, then the recording will have the piano sound, the room noise, and the noise added by the electronics (Johnson–Nyquist noise, shot noise and a few other noise forms as well). If you add any digital effects to the recorded sound, the rounding errors always add noise to the final data stream. The more effects added, the greater the noise. This why 24 bit depth recording is so useful as it gives you a number of bits of rounding error available before you hit CD quality.

Thank you very much for your effort, that most likely means that I am really hearing the upper partials. I thought I knew what they sound like but obviously I was wrong. At least now i know!
Still I feel like the noise is not there on the piano internal recorded WAV, but honestly I know headphones and audio equipment can be different and my hearing (or my brain) even could trick me into hearing things.



Actually the noise at the last four notes more sounds like the damper pedal noise on my piano. Still, the noise I am talking about sounds similar, but it is much quieter. I can hear it in your first two notes for example! It starts right with each note and fades out after one second or so.

At this point I want to say I didn´t want to make a big fuss about that noise, because it is no big deal for me, as it doesn´t concern me while playing. I just found it kind of interesting and wanted to know if more people think or can hear the same. I am not claiming the Kawai pianos are faulty or something like this, I am very happy with the CA67.

This seems very strange : Normally, intermodulation distortion is not made by inharmonicity nor ET, but by non-linearities.

Inharmonicity is caused by a "modified" d'Alembert formulae because of the strings stiffness. The formulaes are still linear, then this can't cause Intermodualtion distorsion.

In order to get some non-linearities we have to hit the strings hard and have important displacements.

You are correct that the piano itself does not produce IM. But, like the tree falling in the forest, does it make a sound if no one is there to hear it?

Remember that IM is simply defined as the aggregate of the sums and differences of all the frequencies produced in a single chord and processed by a non-linear system. The IM we hear is caused by nonlinearities in the ear (and also in speakers). This is why we easily hear the resultant G4 when playing C5 and E5. It does not exist except in our brain. When the sound of a piano is reproduced by an acoustic transducer, it can add real IM to the signal.

edit: One of the neat things about IM is that it produces sub harmonics and harmonics. This means that our ears hear both the inharmonic partials and the harmonic IM products simultaneously. This means that playing an octave on a piano produces, in our ears, nearly coincident partials on top of nearly coincident partials as well as bass reinforcement. What a mess, but what a sound.

Oh, ok. The noise on the 4 last notes are only present in a limited range of velocity, the first notes are my attempt to find the spot. It's as audible through headphones too.
What gets recorded on USB is a very faint noise regardless of the format and is much more similar, if not the same, to what you've pointed out. I can't hear it at a comfortable (to my ears and quite low) volume and certainly not in a recording of regular playing.

First of all - I am by far no expert in analyzing audio files. Maybe what I did is completely useless.

However I took first half of the same tone where I can hear the noise and made an FFT (first pic is via headphone jack, second is intern recorder).
Between 2 kHz and about 7 kHz there are some frequency spikes on the first pic, that either are the noise I hear or some distortion or other quality loss phenomenon.

@ f3r: Wow in your case the phenomenon is very loud then, at least at the last 4 notes. What do the others say to this? Are this partials? In this case I think everyone should be able to hear it.

I can hear it in all the notes in f3r's example too (and to my untrained ears it sounds like the same effect in the first two notes as it does in the last four notes, but quieter. It's really unnatural sounding to me, it sounds like a sizzle (as if someone had left a sheet of paper on the strings when the note was struck).

Definitely not the same as the damper noise as in my HI-XL DP (which only has the EX Concert Grand).

I had a problem with the sound of my CS-11. In this case, the SK lacked clarity and the tone was bass heavy. Because of the overpowering bass I didn't play a lot of notes in the lower register, however there were times where I thought I heard a mild distortion.

This deficiency in sound was present through headphones as well as through the onboard speakers/soundboard, but the sound was better when a file was saved to a flash drive and audio played through another device. Last, the piano also seemed to 'color' audio input the same way - playing the keys as a midi through my iPhone with headphone jack -> audio input -> audio output ->JBL Studio Monitors, I noticed a difference in sound versus bypassing the audio in/out and going straight to the studio monitors.

A technician came yesterday and replaced the amplifier and main board. The result is more clarity and detail with less muddy bass - the lower register sounds beautiful and properly balanced.

The technician didn't hear much of a difference - so much of this is subjective. I am very sensitive to a lot of things, tone being one of them. What I would characterize as a 'huge' difference, to me, is hardly noticeable to some and completely un-noticeable to others.