Layer velocity vs key velocity question...

[AndreasvanHaren]

I am trying to figure out what the relation is between velocity in the different layers within an instrument, and key velocity.

I know that I can switch with modulation data with my modwheel between the different layers within an instrument patch. I can however still play very soft till loud with key velocity on my keyboard within every layer. So how excactly do I use these 2 in combination to get a good result?

Say for example my 1st violin legato patch has 3 layers.
Does this mean that every layer has a different dynamic/timbre? For example layer 1 could be ppp till mf, layer 2 could be mf till f, layer 3 could be ff till fff (depends of course how the company recorded it). In values every layer will have a range from something like 1-40, 41-80, 81-127.

How now does this combine with key velocity? Are the values in every layer the same as the key velocity values? Or is this a complete different thing?

[GrahamKeitch]

I am trying to figure out what the relation is between velocity in the different layers within an instrument, and key velocity.

I know that I can switch with modulation data with my modwheel between the different layers within an instrument patch. I can however still play very soft till loud with key velocity on my keyboard within every layer. So how excactly do I use these 2 in combination to get a good result?

Say for example my 1st violin legato patch has 3 layers.
Does this mean that every layer has a different dynamic/timbre? For example layer 1 could be ppp till mf, layer 2 could be mf till f, layer 3 could be ff till fff (depends of course how the company recorded it). In values every layer will have a range from something like 1-40, 41-80, 81-127.

How now does this combine with key velocity? Are the values in every layer the same as the key velocity values? Or is this a complete different thing?

Hi Andreas,

I've been trying to get my head around the same topic this past week. If we're talking about sustaining instruments, this is what I think:

In the case of mod wheel controlled progressions through different dynamic layers, I dont think key velocity has any meaning. A good library provides different samples for say pp through to ff and the mod wheel is used to seemlessly cross fade between those samples.

I think the confusion arises from the fact that not all libraries have these multiple samples and I think GPO and hence Finale uses key velocity to simulate the effect. If you look at the options in Finale to control dynamics, you can choose vol, cc1 or cc11 - but each adds key velocity. Key velocity does not change over time and is therefore unsuitable for sustaining instruments. I think it's only used as a baseline or start point in the case of Finale and GPO.

I'm not sure whether you're using Finale but if you do, you will need to delve deeper into Finale's HP to get some of the more advanced libraries working to best effect. In Finale, you can use 'instrument techniques' to select say just mod wheel cc1 with NO reference to key velocity for a specific library or instrument. This works for me and is what has led me to the conclusion that key velocity is not relevant in the case of libraries that cross fade through different dynamic layers.

By way of example, I've been using a high end string library for the past year that uses mod wheel to change the volume and timbre as the strings get louder. Or rather, that's what should have been happening! My global settings in Finale for dynamics were set up for cc11 (expression) plus key velocity. So yes, the strings got louder and quieter but with no change of tone. I changed the dynamic settings for strings to mod wheel (no key vel) and I'm now making full use of the samples with greatly improved results. Key vel has been omitted from the technique and also from the expressions I use for the strings staves as it isn't relevant.

Cross fading is even more important for brass where you might want to cresc from a mellow pp up to a brassy ff.

I hope this helps - although I may not have got all this entirely correct

Regards,

Graham

www.soundclick.com/grahamkeitch

Track 1 Intermezzo uses cc1 to cross fade through layers
Track 2 Romanza doesn't use cc1. I had to change patches to get brighter strings when needed - messy!!

[AndreasvanHaren]

Hi Andreas,

I've been trying to get my head around the same topic this past week. If we're talking about sustaining instruments, this is what I think:

In the case of mod wheel controlled progressions through different dynamic layers, I dont think key velocity has any meaning. A good library provides different samples for say pp through to ff and the mod wheel is used to seemlessly cross fade between those samples.

I think the confusion arises from the fact that not all libraries have these multiple samples and I think GPO and hence Finale uses key velocity to simulate the effect. If you look at the options in Finale to control dynamics, you can choose vol, cc1 or cc11 - but each adds key velocity. Key velocity does not change over time and is therefore unsuitable for sustaining instruments. I think it's only used as a baseline or start point in the case of Finale and GPO.

I'm not sure whether you're using Finale but if you do, you will need to delve deeper into Finale's HP to get some of the more advanced libraries working to best effect. In Finale, you can use 'instrument techniques' to select say just mod wheel cc1 with NO reference to key velocity for a specific library or instrument. This works for me and is what has led me to the conclusion that key velocity is not relevant in the case of libraries that cross fade through different dynamic layers.

By way of example, I've been using a high end string library for the past year that uses mod wheel to change the volume and timbre as the strings get louder. Or rather, that's what should have been happening! My global settings in Finale for dynamics were set up for cc11 (expression) plus key velocity. So yes, the strings got louder and quieter but with no change of tone. I changed the settings to mod wheel (no key vel) for dynamics and I'm now making full use of the samples with greatly improved results. Key vel has been omitted from the technique and also from the expressions as it isn't relevant. It's even more important for brass where you might want to cresc from a mellow pp up to a brassy ff.

I hope this helps - although I may not have got all this entirely correct

Regards,

Graham

www.soundclick.com/grahamkeitch

Hi Graham, thanks for your info.

I am using different libraries at the time, GPO but also recently the Sonivox Complete Orchestral Collection. I make my recording in Logic 9 studio and use Sibelius for notation.

Using a string legato patch from Sonivox, I do get a timbre change when holding down a key on my keyboard and turning the modwheel up and down. SO there is a XF between layers going on that also changes the color, not only the volume. According to their website, this legato patch does have 4 layers.

I try to find a good work flow though. At the moment I recorded the strings of one of my orchestral pieces (playing them on my keyboard), and changed the key velocities values according to a schedule I made, diving the range 1-127 from pp till ff

The balance seems to work with these values, although I don't know for sure if these are the values that the recording engineers used.

Now however, I want to do the modulation part in the automation part what would affect the different layers as I understand it correctly. Both the key velocity and layer velocity in this library can change the timbre so do I double work? Maybe I should only use the modulation change on long notes that can't be changed with key velocity?

I noticed that for example with the pizz strings, there are 2 layers, the 1st going from 1-64 and gives a normal pizz sounds, the second layer being triggered starting at 65 and gives a snapp pizz sound. So here it's importand to use the correct value for modulation.

Still withing each layer, I can get different volume attacks with key velocity...

What would be a good work flow in using key and layer velocity?

[GrahamKeitch]

Hi Graham, thanks for your info.

I am using different libraries at the time, GPO but also recently the Sonivox Complete Orchestral Collection. I make my recording in Logic 9 studio and use Sibelius for notation.

Using a string legato patch from Sonivox, I do get a timbre change when holding down a key on my keyboard and turning the modwheel up and down. SO there is a XF between layers going on that also changes the color, not only the volume. According to their website, this legato patch does have 4 layers.

I try to find a good work flow though. At the moment I recorded the strings of one of my orchestral pieces (playing them on my keyboard), and changed the key velocities values according to a schedule I made, diving the range 1-127 from pp till ff

The balance seems to work with these values, although I don't know for sure if these are the values that the recording engineers used.

Now however, I want to do the modulation part in the automation part what would affect the different layers as I understand it correctly. Both the key velocity and layer velocity in this library can change the timbre so do I double work? Maybe I should only use the modulation change on long notes that can't be changed with key velocity?

I noticed that for example with the pizz strings, there are 2 layers, the 1st going from 1-64 and gives a normal pizz sounds, the second layer being triggered starting at 65 and gives a snapp pizz sound. So here it's importand to use the correct value for modulation.

Still withing each layer, I can get different volume attacks with key velocity...

What would be a good work flow in using key and layer velocity?

Hi Andreas,

Taking the last few points first, I should have mentioned that mod wheel isn't just used for x-fade. It can be used for vibrato or as you note above, to trigger a different method of pizz. It does depend on the library or instruments within a library. I'm told some of the old Vienna libs used it to change articulations through a wide range of choices from arco to pizz which might have been quite useful. Better than keyswitching!

To answer one of your other points, yes - mod wheel is for sustaining instruments. For non-sustaining, eg percussion, harp and pizz strings, key velocity is needed.

Key vel can sometimes be used with mod wheel controlled instruments to add attack - but this is something else again. The string lib I use has an option to overlay staccato patches to give extra kick and this could well be an instance of adding key velocity into the mix - but in this case, it's not really anything to do with the x-fade process.

I'm not sure about the workflow question because I haven't encountered a need to manage key vel and layer control at the same time. For me it's either mod wheel (for sustaining x fade instruments) or key vel (for percussive instruments).

G

[rbowser-]

Hi - interesting discussion.

I'm rather sure, I've always understood this to be true, that in GPO and the Garritan Libraries, layers of samples come into play through Mod Wheel, and Velocity controls the envelope--mainly the Attack.

Turn your Mod Wheel up all the way so you have the loudest available volume for an instrument, such as a flute soloist. Now hit a note at varying velocities. The volume stays the same, since CC1 is controlling that, but at high velocities, you'll get a sharp attack, lower velocities will get you slower, softer attacks.

This is great, because it's possible to have both a soft attack yet high volume - rather than having a slow attack always attached to low volume, which is admittedly more the usual case - but the programming makes these instruments more versatile because of the blend of controllers used.

Randy

[GrahamKeitch]

Hi - interesting discussion.

I'm rather sure, I've always understood this to be true, that in GPO and the Garritan Libraries, layers of samples come into play through Mod Wheel, and Velocity controls the envelope--mainly the Attack.

Turn your Mod Wheel up all the way so you have the loudest available volume for an instrument, such as a flute soloist. Now hit a note at varying velocities. The volume stays the same, since CC1 is controlling that, but at high velocities, you'll get a sharp attack, lower velocities will get you slower, softer attacks.

This is great, because it's possible to have both a soft attack yet high volume - rather than having a slow attack always attached to low volume, which is admittedly more the usual case - but the programming makes these instruments more versatile because of the blend of controllers used.

Randy

Hello Randy, it's been a while since we last 'spoke'. Hope all is well with you.

Your explanation of cc1 and key vel as used in GPO makes perfect sense and explains some of the GPO optimised settings I'm encountering with Finale. I'm sure this will help enlighten Andreas too.

Kind regards,

Graham

[rbowser-]

Hello Randy, it's been a while since we last 'spoke'. Hope all is well with you.

Your explanation of cc1 and key vel as used in GPO makes perfect sense and explains some of the GPO optimised settings I'm encountering with Finale. I'm sure this will help enlighten Andreas too.

Kind regards,

Graham

Hi, Graham - It's very good to see you! And I'm glad what I posted is helpful. Best wishes on your musical endeavors.

Randy

[AndreasvanHaren]

What about this, something else I am experimenting with:
Not using key velocity for sustain instruments like legato strings, at all. Only modulation (modwheel) velocity. I noticed that the legato lines this way are much more fluid, no sudden attacks.

The problem here is that when recording the part on my midi keyboard, there is always key velocity registered. Logic can't filter out key velocity as far as I could find in the settings, and the string patch I am using doesn't come in a version that doesn't register key velocity, so what I did was selecting the recorded part and set all the key velocity to the same value, in this case I used 50.

[AndreasvanHaren]

Hi Andreas,

In the case of mod wheel controlled progressions through different dynamic layers, I dont think key velocity has any meaning. A good library provides different samples for say pp through to ff and the mod wheel is used to seemlessly cross fade between those samples.

By way of example, I've been using a high end string library for the past year that uses mod wheel to change the volume and timbre as the strings get louder. Or rather, that's what should have been happening! My global settings in Finale for dynamics were set up for cc11 (expression) plus key velocity. So yes, the strings got louder and quieter but with no change of tone. I changed the dynamic settings for strings to mod wheel (no key vel) and I'm now making full use of the samples with greatly improved results. Key vel has been omitted from the technique and also from the expressions I use for the strings staves as it isn't relevant.

Cross fading is even more important for brass where you might want to cresc from a mellow pp up to a brassy ff.

I hope this helps - although I may not have got all this entirely correct

Regards,

Graham

Hi Graham, I realize that my previous post is exactly what you explained to me! How did you prevent key velocity from being recorded? Or otherwise, if recorded, how did you handle the recorded key velocity? Like I said in my last post, I set all the values to a fixed one, for now 50.

[rbowser-]

What about this, something else I am experimenting with:
Not using key velocity for sustain instruments like legato strings, at all. Only modulation (modwheel) velocity. I noticed that the legato lines this way are much more fluid, no sudden attacks.

...

Hi, Andreas

I understand what the motivation for your experiments are - you want to get the best sound possible. Good goal, of course!

--Let me try to explain why what you're suggesting here would greatly reduce the expressiveness of the instruments. It may sound like an idea to try, but to have no velocities, or all low ones, is to ignore and not use a good portion of the great programming in GPO which makes it more natural sounding.

In my last post, I explained that velocity controls the shape of the attack envelope in GPO's sustaining instruments. The attack is the most critical part of a sound - it's in the first split second of a sound when our brains recognize what an instrument, largely because of what the attack sounds like.

You need full range of velocity values, 1-127, in order to get the most out of the instruments.

BUT what you're unhappy with is still hearing too much attack on legato string sections, as a perfect example.

TWO THINGS in regards to that:

1) You need to use a sustain pedal, CC64, for legato passages. The full value of 127 for the first CC64 event which is played or inserted right after the first note in a passage has started. A value of zero for the second 6654 event which is played or inserted after the last note has been played in the passage.

That is EXTREMELY important because CC64 truncates the beginning of notes! - It cuts off the first part of the attack--It was designed specifically to make legato passages sounding more natural for strings, winds and brass.

Much of my editing time using Garritan instruments has always been in carefully bracketing all my legato passages by hand in the Piano Roll View of my software, Sonar, making sure I'm making full use of that clever bit of programming.

2) You're right that velocity values should be low after the first note of a legato passage. You said you noticed that played notes have a variety of high velocity values - naturally, because you played them.

Keyboards are capable of sending out velocities in different curves, including having all the velocities be the same - AVOID that unless you want to instantly robotize your music. Your experiment of having all the string values at 50 won't work.

In your recording software's Piano Roll View, it's extremely easy to reduce the velocity values of a passage. After making sure that the first note of a passage has the right value, then sweep through with your pencil tool and chop off the tops of all the other velocities, or use a plug-in to reduce their values by a percentage.

Andreas, I'm very accomplished at using GPO, and have been recording with it for 5 years. You need to take this input seriously. I recognize your desire to improve your playback, and to get the most out of your software. I'm writing this response first thing this morning because I Really want to help you understand why to do things like flatten all velocity values will only drain your music of life - it will do just the opposite of what you're expecting.

Randy