Is There a Problem Combining Two Pianos?

[GrannyRocks]

I was experimenting with combininb the Warm Grand with the Character Grand, but is there a sound problem created by using two pianos? Dropouts or anything like that? Thanks.

[overover]

Hi Beth,

Some voices are based on the same wave samples (although the voices may sound different). If two such voices are combined, i.e. played at the same time, frequency cancellations (so-called "comb filter effects") can occur, which can, for example, result in a thin "phaser/flanger-like" sound.

However, you can always combine two (or three) voices based on different samples to achieve a fat, rich sound.


Best regards,
Chris

[GrannyRocks]

Hi Beth,

Some voices are based on the same wave samples (although the voices may sound different). If two such voices are combined, i.e. played at the same time, frequency cancellations (so-called "comb filter effects") can occur, which can, for example, result in a thin "phaser/flanger-like" sound.

However, you can always combine two (or three) voices based on different samples to achieve a fat, rich sound.


Best regards,
Chris

Thanks, Chris. Just what I was wondering. So I think the CFX and the Character Grand would be based on different samples. Correct?

[vlbrgt]

Some voices are based on the same wave samples (although the voices may sound different). If two such voices are combined, i.e. played at the same time, frequency cancellations (so-called "comb filter effects") can occur, which can, for example, result in a thin "phaser/flanger-like" sound.

How can we know if two voices are based on the same wave sample ?

Regards
Etienne

[BogdanH]

Hello Etienne,

How can we know if two voices are based on the same wave sample ?

If we would say "let's find piano voices which use the same samples", then that's practically impossible. The reason for that is because resulting sound depends on effects/distortions that have been applied on voice samples.

Theoretically it's quite simple if we make use of interesting effect that Chris mentioned. It goes like this: we record two voices that we wish to compare (at the same velocity). It's quite important that both recorded samples have almost exactly the same start time and similar duration (let's say 1 sec). After that we invert frequency on one of these two samples -that is, we make inverse frequency polarity. Btw. all that can be done in good audio editor.
Now we measure loudness (LUFT) of each sample and if we have done everything properly, both samples should have similar loudness (let's say, each has loudness of about -10LUFT). And finally, we add those two samples -that is, resulting sample is a sum of both samples. Now me measure the loudness of that resulting sample. If resulting loudness is less than -10LUFT (means, result is quieter than each sample), then we can assume that both voices use the same source sample.
In ideal case (if both voices are exactly the same), after adding normal and inverted sound, result it total silence.

What I just described isn't just theory. I used this method for testing when I was creating custom voice for Yamaha CFX piano. Anyway, that process isn't really practical for what we want here.

In reality, the only way to find out which voices use the same sample is, if voice contains specific characteristic (error), which we usually only notice by coincidence. That was the case when I noticed distortion in Saxophone voice (if interested, see this video). There's similar case for some guitar voices where one note (sample actually) is quieter than the rest of them. That is, once we know on what we need to pay attention, it's relative easy to isolate voices that use the same samples.

Bogdan

[overover]

Thanks, Chris. Just what I was wondering. So I think the CFX and the Character Grand would be based on different samples. Correct?

Hi Beth,

Yes, these voices clearly use different samples. The CharacterGrand is a German "Schimmel" piano. So nothing can go wrong when combining it with the Yamaha WarmGrand (regardless of whether C7 WarmGrand or Legacy WarmGrand). 

Basically I would just rely on my ears, i.e. if it suddenly sounds strangely thin and/or phaser/flanger-like, I know that two voices (which, each sound good in their own right) are using the same mulisample, and I'll look for it just another second voice to combine.

A very useful method is to use GM/XG Voices to "thicken" other Voices. These older Voices almost always use different multisamples than the newer Panel Voices. As is well known, on newer models such as Genos you can no longer select the GM/XG Voices directly for the keyboard voice parts because they are not shown by default in the Voice Selection display. However, there is a simple trick to still be able to call up the GM/XG Voices (copying the corresponding voice files to the user drive or to a USB stick). If you need further tips on this, just get in touch here again.


Best regards,
Chris

[andyg]

Since sampled sounds first appeared in synths, organs and keyboards, the golden rule has been:

"Don't use the same sound twice, at the same octave." The instrument can't produce them simultaneously, which leads to thin sounds, phasing etc.

To find out if two voices share the same sample set, just try them together. No need to go into anything complex. If they phase, you've found a set of 'twins'. If they don't, then they come from different sample sets. It's not just pianos, of course, the same applies to woodwinds, percussion, strings (capable of making ghastly sounds!), brass and so on.

If you really want to use two sounds that are twins, you can try this.

1) Pan one left and one right
2) Detune them ever so slightly

This works better for instrumental sounds like brass, saxes etc and, to be honest, it's what you should think about doing whenever you use two trumpets, two saxes or whatever, regardless of whether or not they are twins. Panning is essential. I always tell my students that in a concert, one player doesn't stand on the other's shoulders!

[GrannyRocks]

Hi Beth,

Yes, these voices clearly use different samples. The CharacterGrand is a German "Schimmel" piano. So nothing can go wrong when combining it with the Yamaha WarmGrand (regardless of whether C7 WarmGrand or Legacy WarmGrand). 

Basically I would just rely on my ears, i.e. if it suddenly sounds strangely thin and/or phaser/flanger-like, I know that two voices (which, each sound good in their own right) are using the same mulisample, and I'll look for it just another second voice to combine.

A very useful method is to use GM/XG Voices to "thicken" other Voices. These older Voices almost always use different multisamples than the newer Panel Voices. As is well known, on newer models such as Genos you can no longer select the GM/XG Voices directly for the keyboard voice parts because they are not shown by default in the Voice Selection display. However, there is a simple trick to still be able to call up the GM/XG Voices (copying the corresponding voice files to the user drive or to a USB stick). If you need further tips on this, just get in touch here again.


Best regards,
Chris

Ah. Maybe another time. Grin. Actually when I combined the Character Grand with the CFX Concert or Warm, it seems a little less natural than the Character Grand by itself. Am I making that up, or does it have a more electronic sound?

[GrannyRocks]

Since sampled sounds first appeared in synths, organs and keyboards, the golden rule has been:

"Don't use the same sound twice, at the same octave." The instrument can't produce them simultaneously, which leads to thin sounds, phasing etc.

To find out if two voices share the same sample set, just try them together. No need to go into anything complex. If they phase, you've found a set of 'twins'. If they don't, then they come from different sample sets. It's not just pianos, of course, the same applies to woodwinds, percussion, strings (capable of making ghastly sounds!), brass and so on.

If you really want to use two sounds that are twins, you can try this.

1) Pan one left and one right
2) Detune them ever so slightly

This works better for instrumental sounds like brass, saxes etc and, to be honest, it's what you should think about doing whenever you use two trumpets, two saxes or whatever, regardless of whether or not they are twins. Panning is essential. I always tell my students that in a concert, one player doesn't stand on the other's shoulders!

That's interesting. What if you have a single array speaker? Does the panning make a difference?

[andyg]

To amplify a stereo instrument like Genos 2 (and almost any keyboard) you should have stereo amplification. If you run into mono, you lose all the depth of the stereo image. Many of the sounds are stereo anyway, like the pianos and strings, and these won't sound nearly as effective in mono.

In a home situation, I personally like my speakers fairly close to me, so I position them just a little further apart than the width of the keyboard. That's enough to convey the stereo image. A piano will pan from left to right as you go from bass to treble, just as you would hear it when you're sitting at your Kawai grand. Strings will 'expand', for want of a better term. When you pan a solo sound like a trumpet, flute etc to one side, you can then add a second sound at the other side and probably put the third sound in the centre. One of my fellow teachers used to tell her students "Panning is your friend, get to know it!" I borrowed that and tell my students the same!

None of this can be done in mono so if you are using just one mono amp, then do think seriously about going stereo.

[GrannyRocks]

Thanks for your comment. I do understand that. We use a fantastic line array speaker. It's not stereo but it's great for gigging and it's like stereo, or so they say — haha — because it has a bunch of speakers in the tower and they point in different directions. Are you familiar with them? There's a subwoofer on the floor and a tower. We have the Electro Voice Evolve 30 and the sound is alive. I have studio monitors too, and they are stereo. If we had nothing but money, we would buy a second Electro Voice and that would really be something. Who knows? Thanks so much for writing.

[overover]

Ah. Maybe another time. Grin. Actually when I combined the Character Grand with the CFX Concert or Warm, it seems a little less natural than the Character Grand by itself. Am I making that up, or does it have a more electronic sound?

Hi Beth,

As I said, just rely on your ears. 

If it sounds good when you combine two or three voices (layer), it fits. If it sounds "worse" than every voice played individually, you just try other voices to combine.

So that you hear what I sounded with "thin, phaser/flanger-like" sound, load the same preset voice into two keyboard parts (e.g. Right1 and Right2). It would be best to use a voice with a lot of overtones ("harmonics"), e.g. an Accordion voice . You will notice that the sound becomes thinner/unnatural when both parts are on. In addition, you will notice that the sound changes slightly when you play the same note repeatedly (similar to a phaser/flanger effect). This is due to the fact that the two Voices are not synchronized in signal phase. Therefore, the voices are randomly played by the sound generation with a minimal time offset, which changes slightly with every played note. If you now switch off one of the two parts, you have the usual (good) sound again.

In order to avoid the "phasing" problem, as already mentioned here, you could slightly tune the individual voices up/down. To do this, go to "Menu> Voice Setting > Tune > Tuning". In the case of the accordion example, you could, for example, load the same accordion voice in all three Right parts. For the first part leave the tuning at 0, the second part tune slightly upwards and the third part tune the same amount (or slightly different) down.

For a "fat" accordion sound you could try the following. Use a "Cassotto" or "Clarinet" accordion voice (in normal tuning) in Right1. In Right2 and Right2 you each use a tremolo accordionon voice (e.g. "Steirische") and detune these two parts sightly against each other, as mentioned before.

You can save the result directly in a Registration.

Just try out the above occasionally to find out what exactly I mean with the "phaser/flanger-like" sound mentioned (and how to avoid it).


Best regards,
Chris

[attachment unavailable]

[GrannyRocks]

Hi Chris,

That's wild. In the meantime, I'm trying to use the Character Grand in the right and left hands for a registration, but there are notes that are in the wrong octave. A few in between. I'll try to adjust the split for just that registration, but I'm not sure how. haha. More technical problems. A concert Sunday. Yikes!