Music Player Network

Previous Thread
Next Thread
Print Thread
Mono Piano Discussion #2250006 12/02/10 03:37 PM
Joined: Oct 2000
Posts: 8,141
burningbusch Offline OP
MP Hall of Fame Member
OP Offline
MP Hall of Fame Member
Joined: Oct 2000
Posts: 8,141
Maybe this is elementary for everyone here but I've seen comments posted that lead me to believe there's some confusion with regards to mono piano patches.

To my knowledge, most pianos are sampled using some variation of the close microphone technique where two mics are placed inside the piano, one above treble strings and the other above the bass. The mics are spaced no more than a meter apart. The close mic technique is preferred because it presents the player, when listening through headphones or properly positioned near-field monitors, a sound distribution similar to the acoustic instrument (treble largely right, bass largely left). Also it captures more of the hammer strike providing more of a direct, percussive sound. More advanced software-based piano libraries often provide multiple microphone positions from which to choose.

Now what are the challenges of collapsing the stereo image into a mono one? Remember that sound travels at approx. 1ms per foot. Let's say, for sake of simplicity, the two mics are equal distance from middle C. In this case the sound will reach the two mics at the same time and there will be no phase issues. On the other hand, what happens if you are sampling a note that resides directly in line with the treble mic? Here it will have almost zero delay in reaching the treble mic but 2 - 3 ms delay in reaching the bass mic. Can this cause phasing issues? You bet.

Below is a recording of a Rhodes note, which was duplicated and moved 2 ms. As you can see there are major phasing issues happening, in fact the fundamental is completely gone if you'd listen to this in mono.

fig. 1

I used a Rhodes above because it's easier to see what's going on with the simpler waveform. An acoustic piano's waveforms are much more complex. Additionally, the two mics don't pick up the same exact sound so the waveforms are different. One might pick up more of the direct hammer sound while the other is picking up cabinet reflections that will probably have more bass content. Together they create a sound that is full, balanced and has depth/image. The phasing issues are ever changing as they are a mix of the note's frequency and the distance sound has to travel to the microphones. There is no simple solution here. If you collapse to mono the stereo image of a piano sampled using traditional spaced-pair stereo techniques you will hear phasing throughout the keyboard.

There are other more mono-friendly mic'ing techniques that could be deployed such coincident pairs and Mid/Side but I don't know of any examples of these in the sampling world, at least as far as hardware instruments.

Why do these digital pianos sound so good through headphones but ‘kinda suck’ when listening through speakers? Below (fig. 2) is a picture of two sine waves that will produce 100% phase cancelation (read: you won't hear anything) when listened to in mono. What's interesting is that if you pan one far left and the other far right, you will now hear the sine waves. If you listen to the panned versions through headphones, the sine waves will sound perfect. If you pan them back to mono the sound will disappear. Headphones allow one ear to be fed 100% of one waveform while the other gets 100% of the other. This is a pristine listening environment where there are zero phase cancellations. Listening through headphones is very close to hearing exactly what the microphones picked during the recording session. In addition, headphones silence nearly all the ambient noise that can interfere with the sound. So the manufacturers don't intentionally try to make their pianos sound great through headphones, it is simply the best listening environment because it is pristine. Unfortunately, it isn't real world.


fig. 2

Continuing with this thought, if you take two great sounding, top-of-the-line near-field monitors and space them properly, you should have an excellent listening experience. But if you place those same speakers side-by-side they will sound boxy, one-dimensional and probably less full. Though I haven't done the experiment myself, I don't think there's much, if any, difference between collapsing a stereo image down to mono in a mixer vs. placing a pair of speakers side-by-side. When both ears are fed the same amount of the two sources, the result will be mono and phasing cancellation will occur. This is the reason why stereo keyboard amps tend to sound boxy and though are an improvement over mono amps given stereo sources, only very slightly.

So what does this phase cancellation sound like when a stereo sample set is collapsed or summed to mono? Below is an MP3 recorded on an RD300GX using the Superior Grand sound. The first example uses the SuperiorMono patch which is a summed to mono example. Two my ears, it sounds identical to the Superior Piano stereo patch when it is summed to mono using a mixer. The summed mono exampled is followed by what I like to call a split mono. When you think about it, the “stereo” mics used in the sampling process are really just two mono mics spread out. So if you create a split and assign the right half of the stereo wave to the upper split and the left half to the lower, you have a mono sampling of the keyboard but done with two mics not one. This gives a better balance to the sound vs. using one or the other. You of course will need to bring the outputs of the keyboard into separate channels on the mixer and pan these straight up, assuming the keyboard is not able to do this internally.

Summed-to-Mono vs. Split Mono

Please note these examples were recorded with no effects and at a static velocity. They are intentionally sterile so that you can hopefully hear the differences.

People always seem to want to sum-to-mono these stereo sample sets and I don’t completely understand the logic. True, one of the reasons why the stereo piano sounds full and rich is because the two mics (often more are used in studio recordings) pick up the tone as it spreads across the instrument. An acoustic grand piano is a massive instrument. If you position yourself at various points around, above and below the case you’ll hear how different the tone can be. It’s those differences, when combined, which create the complex composite tone we associate with the piano. A single microphone, I’m afraid, is just not capable of capturing the same richness of tone. But, as I believe I’ve shown above, multiple mics will invariably create phase issues and there is no easy way around this. It is the proverbial “rock and a hard place.” Personally, I would rather use a mono source (or split mono) that is free from phase issues vs. the summed-to-mono approach.

So do I belong to the Kanker-school? Overall I would say no as I feel the piano needs to be multi-mic’ed to sound proper but I am completely sympathetic to the issues when performing live. Thankfully, my days of worrying about FOH mix, direct boxes and all the rest are gone. When I play out, which has been less and less, it’s smaller, quiet venues so I can set up a couple of small speakers on poles and not worry about much else. When playing live, I do think it's advantageous to use a simpler representation of the piano vs. a more complex/detail version. You need something that efficiently and effectively gets your musical ideas across to the audience. It only needs to sound enough like a piano so that they (and you) are not distracted by it. Given that, I can certainly understand why someone would make the case that a mono piano fits the bill better than stereo. Maybe physical modeling is the answer. I just don't know.

As a side note, I will say the Yamaha AvantGrand, which I own, is remarkable in its ability to mimic an acoustic piano in a room. It has four channels (tweeter + mid-range on all channels) pointing up and four more speakers pointing to the floor underneath. The top speakers are positioned exactly where the microphones were during the recording process. When played at the proper volume it can sound remarkably like an acoustic piano when you stand a few feet away. So I don’t believe one or two speakers are truly sufficient if you’re looking to re-create the image of a grand piano in a room.

Busch.

KC Island
Re: Mono Piano Discussion [Re: burningbusch] #2250012 12/02/10 04:13 PM
Joined: Aug 2005
Posts: 12,648
K
kanker. Offline
KCFF League Champ '13
10k Club
Offline
KCFF League Champ '13
10k Club
K
Joined: Aug 2005
Posts: 12,648
I would think that micing each note individually, and hopefully in an ideal manner, would produce the best 'raw' piano sound. Add convolution of a piano shell, surround, stereo or mono depending on your taste, and you're probably in the right ballpark. Then the same sample set could probably function effectively in ANY setting.

Then again, I could be wrong.

Thanks for the analysis Busch!


A ROMpler is just a polyphonic turntable.
Re: Mono Piano Discussion [Re: burningbusch] #2250015 12/02/10 04:28 PM
Joined: Jul 2008
Posts: 13,422
J. Dan Offline
Bannination Free Since '08
10k Club
Offline
Bannination Free Since '08
10k Club
Joined: Jul 2008
Posts: 13,422
If someone made a Ribbon mic that were long enough to cover the entire sound board, so that each string was equal distance from a single mic element, do you think it would be possible to capture the full harmonic content in mono? The sound, I think is more complex than that, because you still have reflections within the piano as well as the room that contribute to it's sound.

If the mic positions were known in the example you gave, theoretically, DSP could shift the phase of one signal relative to frequency to correlate to the fundamental of the string and its distance from the other mic. However, the harmonic content would be too complex to apply this same procedure to, so there would still exist comb filtering in the upper harmonics.

Do you think Stereo Room micing would collapse to Mono better than close-micing?


Dan

Acoustic/Electric stringed instruments ranging from 4 to 230 strings, hammered, picked, fingered, slapped, and plucked. Analog and Digital Electronic instruments, reeds, and throat/mouth.
Re: Mono Piano Discussion [Re: burningbusch] #2250018 12/02/10 04:46 PM
Joined: Dec 2002
Posts: 17,576
ITGITC? Offline
Grand Poobah of Posting
10k Club
Offline
Grand Poobah of Posting
10k Club
Joined: Dec 2002
Posts: 17,576

Excellent information! Thank you for taking the time to post it, Busch.

We all want to sound our best live and recorded. It matters less whether it's stereo or mono.

In light of your post, instead of summing to mono, is it a better idea simply to use the left OR the right output from the keyboard or mixer? Other than specifically using mono samples and programs, what are best practices for live as well as recording?

For example, playing with a jazz trio in a coffee shop, is there any reason not to use both outputs and play in stereo if your speakers are spaced far enough apart? Or, instead, should we simply forget that and always either sum to mono or use the left OR the right output?

If I choose to use programs/samples recorded in mono, I'm limiting myself to the small variety of mono sounds offered within my keyboard. Of course, the objective is still to sound as good as possible.

In a stadium setting or on a recording the answer will, of course, be different.

Perhaps there is no all-encompassing answer other than "use your ears to determine how to sound your best". However, we are often in a playing position; not in a listening position along with the audience. It can be difficult to gauge how we sound from the audience's perspective.

This is a complex question with many 'right' answers, depending on the performance environment. I don't think that anyone here enjoys being berated for using stereo when, as you said, in certain circumstances it can sound wonderful in the right environment.

Thanks!

Tom


"Music expresses that which cannot be put into words and that which cannot remain silent.
Re: Mono Piano Discussion [Re: J. Dan] #2250019 12/02/10 04:49 PM
Joined: Oct 2005
Posts: 497
lerber3 Offline
Senior Member
Offline
Senior Member
Joined: Oct 2005
Posts: 497
I would love to hear how a good 'sum to mono' is performed for a piano sound. Nord has a mono button. If I were to implement this feature, here's my guess as to how it is done.

I am assuming that there are 2 mics in fixed positions from which all notes are sampled. I would analyze the stereo signal for each note and sample layer to determine the relative phases of the fundamental frequency for each note. I would pre-process the velocity layers to bring them into phase independently for the left and right channels. I would then either pre-process the samples for each channel to bring them into phase... or store the phase shift along with the samples for each note and do it on the fly before summing the stereo signals into mono.

That's my guess... does anybody know how it's really done?

I type too slow... 80s-LZ suggests the same thing... but I'm still curious how piano samples are processed into mono.

Last edited by lerber3; 12/02/10 04:56 PM. Reason: type too slow
Re: Mono Piano Discussion [Re: lerber3] #2250084 12/02/10 09:06 PM
Joined: Oct 2009
Posts: 13,925
AnotherScott Offline
10k Club
Offline
10k Club
Joined: Oct 2009
Posts: 13,925
On a different but related track... Is it necessarily a given that all stereo sampled pianos create the stereo field from the player's perspective, i.e. left mic over bass strings, right mic over treble strings? That would seem to only be accurate to one person, the player. And that makes sense for a console-style digital piano that is going to be played in the home, or for a headphone mix for the player.

But if a piano is going to me marketed as a "stage piano" for performance, it would seem that, assuming you want stereo at all, you'd want it from the perspective of the left mic near the player end of the piano and the right mic at the opposite end of the piano, to try to emulate how a real grand piano sounds to an audience. Audiences rarely hear a piano from the perspective of standing behind the player. If the point of stereo is to try to simulate the acoustic field of an actual piano, shouldn't the true listener's perspective be taken into account?


Maybe this is the best place for a shameless plug! Our new video at https://www.youtube.com/watch?v=3ZRC3b4p4EI is a 40 minute adaptation of T. S. Eliot's "Prufrock" - check it out!
Re: Mono Piano Discussion [Re: AnotherScott] #2250093 12/02/10 09:42 PM
Joined: Aug 2005
Posts: 12,648
K
kanker. Offline
KCFF League Champ '13
10k Club
Offline
KCFF League Champ '13
10k Club
K
Joined: Aug 2005
Posts: 12,648
Originally Posted By: AnotherScott
On a different but related track... Is it necessarily a given that all stereo sampled pianos create the stereo field from the player's perspective, i.e. left mic over bass strings, right mic over treble strings? That would seem to only be accurate to one person, the player. And that makes sense for a console-style digital piano that is going to be played in the home, or for a headphone mix for the player.

But if a piano is going to me marketed as a "stage piano" for performance, it would seem that, assuming you want stereo at all, you'd want it from the perspective of the left mic near the player end of the piano and the right mic at the opposite end of the piano, to try to emulate how a real grand piano sounds to an audience. Audiences rarely hear a piano from the perspective of standing behind the player. If the point of stereo is to try to simulate the acoustic field of an actual piano, shouldn't the true listener's perspective be taken into account?
It's not taken into account, which is another reason that stereo pianos live are whack wink


A ROMpler is just a polyphonic turntable.
Re: Mono Piano Discussion [Re: AnotherScott] #2250098 12/02/10 09:52 PM
Joined: Aug 2007
Posts: 1,345
Mike Warren Offline
Gold Star Recipient
Platinum Member
Offline
Gold Star Recipient
Platinum Member
Joined: Aug 2007
Posts: 1,345
Originally Posted By: AnotherScott
If the point of stereo is to try to simulate the acoustic field of an actual piano, shouldn't the true listener's perspective be taken into account?


DPs are sold to players, not audiences. Although it looks like most people here do understand what's going on, that is not, IMO, true of the greater percentage of DP buying musicians.

Frankly, I think 95% of audiences couldn't tell, or care.

I used to get comments on how much my Roland HP70 sounded like a real piano. Ug!


DigitalFakeBook Free chord/lyric display software for windows.
Re: Mono Piano Discussion [Re: kanker.] #2250099 12/02/10 09:54 PM
Joined: Dec 2006
Posts: 9,238
Dana. Offline
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Dec 2006
Posts: 9,238
Originally Posted By: kanker.
It's not taken into account, which is another reason that stereo pianos live are whack wink

thu

And thanks to Dr. Busch for dropping knowledge. Surely this can be parlayed into a regular Keyboard column? idea cool

Re: Mono Piano Discussion [Re: Dana.] #2250399 12/04/10 06:06 AM
Joined: Apr 2009
Posts: 1,210
OB Dave Offline
Platinum Member
Offline
Platinum Member
Joined: Apr 2009
Posts: 1,210
This is a topic that is near and dear to my heart.

The solution, I think, is pretty easy - pianos should be sampled using the M-S miking technique, because you can mix an M-S source to mono without any phase weirdness. To my knowledge, very few sound modules take advantage of this.

A few years ago I bought a Receptor and loaded it with Ivory and Komplete, which include Akoustik Piano. Both of these pianos have a continuous control for stereo width, and I think the only way to accomplish this would have been to use the M-S technique. So I'm assuming how both of these gigapianos were sampled. They sound amazing in mono or stereo, but I gig in mono.

Last edited by OB Dave; 12/04/10 06:08 PM.

Moderated by  Dave Bryce, Stephen Fortner 

Powered by UBB.threads™ PHP Forum Software 7.7.3