You can set the a range of notes that will sound a .wav that is loaded to the MONTAGE. This is called it’s KeyBank. A KeyBank contains data from an audio file (.wav) — it is assigned to a Key that will sound it at its original Pitch. A KeyBank is defined by a Note Range (horizontally) across the MIDI Keyboard, and a Velocity Range (veritically) 1-127.

A Waveform is a collection of samples. As many as 256 samples can occupy a single Waveform. Each AWM2 Element references a Waveform. Eight Elements can occupy a Part.

If you have the Range set to just C1~C1, then only that Key will trigger the audio, and it will play at its original Pitch. If you were to increase the Key Range to C1~C2 then the Keys in that Range will create a chromatic scale, each higher Note is 1/12 of an octave higher in pitch. This is accomplished by speeding up the playback of the audio to create an Equal Temperament scale the default)... when you reach C2, one octave higher the pitch will be precisely an octave higher.
It plays back twice as fast... which means it lasts half the time.

In general, the more pitches you sample from the original instrument the better it will sound (to a point). One sample will not survive being stretched too far... this is why multiple samples are done across the full range of the instrument.

In the AWM2 engine you can “stretch” an imported audio wave + two octaves going up and four octaves going down... the phenomena that occurs when the audio is stretched is called “chipmunking” or “munchkinization”... as any recording is sped up the pitch goes up. How they create the chipmunk voices for cartoons...

If you record your voice speaking, and then play it back. Your first response is, you don’t recognize your own voice (this is normal). But all your friends standing there will verify that, “yes, that’s you!” But as soon as you take it up a semitone... everyone laughs, because it’s no longer you! Why it’s funny is another discussion, but we all start to laugh. So much of how our ears and brain recognize sound is based on correct pitch and harmonic content. What happens is there are some things in your Voice that have changed Pitch that normally do not. It’s not just you talking faster at a higher pitch, it’s some how no longer even you. This is the physics of sound.

This is because the fixed frequency components in your voice, that all your friends use to recognize your voice (no matter how faster, or at what pitch you are speaking) those things have been altered. Sampling one pitch and “stretching” it across a range of keys will not necessarily fool the ear/brain. It is why multi-sampling is done.

Editing the .wav to make it a Waveform
When you load a .wav to a Key, you would start by pressing [CATEGORY]
Select “Init” > “Init Normal (AWM2)”
Press [PART SELECT 1]
Press [EDIT]
Select “Elem1” in the screen
Tap “Osc/Tune” > touch the box “New Waveform”
Here you can load an Audio File.
Set the target with the KEYBOARD SELECT option — this will select the note that will sound the wav at its original pitch

Once loaded you touch the Elem1 box that says “Edit Waveform”
It is here you can set the Velocity Range and the Note Range — these determine where and how it will respond

Thank you for your prompt response Bad Mister, I will follow your advice to see how it goes, it is very much appreciated. I have 2 other questions and hope you or someone can help.

Obviously I would like to sample many pitches but the problem I am having is that when I try to record the sound played from my mp3 in SampleRobot, the software does not seem to get the signal from the mp3, not sure what the problem was. In the setup I chose "Audio In Device" as "Microsoft Sound Mapper - Input" from SampleRobot so that was why I had to record the sound with one note and convert to wave form before I imported into SampleRobot. If sampleRobot can recognise the sound from mp3 then it would be great, so if you know how then please let me know.

Also as I mentioned in my original post I had to use the option "Copy starting at key 36 (c1) in no special order)" to import the sound, none other options would allow me to import, which means regardless of what key that single note in it will always start at C1 so no way for me to record a note, for example, in G# and import in at position G#, it will always play at C1 with the pitch G#. I hope you understand what I meant.

Also if I record a wave with 3 or 4 notes then when I play on Montage it plays back those 3-5 notes in the sample key C1 which is even worst.

Thanks in advance for your help.

Kind regards

Martin

Also as I mentioned in my original post I had to use the option "Copy starting at key 36 (c1) in no special order)" to import the sound, none other options would allow me to import, which means regardless of what key that single note in it will always start at C1 so no way for me to record a note, for example, in G# and import in at position G#, it will always play at C1 with the pitch G#. I hope you understand what I meant.

No, I’m sorry, you’ve lost me. Even the word “copy” in this context is confusing... sorry I don’t understand.

If you are having issues with Sample Robot and mp3, I’m sure you can contact them about how to or even if that works. But once you have a .wav File, I can help you load that to a specific Key, and how to map it. See above.

Hi Bad Mister, I have attached here the screenshot from SampleRobot of the options I had to choose when trying to import a wave file. I didn't make it up. Thanks again.

Martin

The "copy" option lets you move the sample to some key other than the one designated as root inside the sample itself. Samples usually contain meta-data for various things - loop points, root note, etc. They do not have to have these, but they generally do. It's conceivable that you want to import a bunch of samples at a time and all have the same root note of C3. If Samplerobot assumed the meta-data for root note was king - then it would import all the samples to the same key - which is not what you want. If you got lucky and all of your sample files have the root note set up differently - C3, C#3, D3, etc... then you lucked out and can use that method. If not, you may need a way to change the root note easily. The easiest thing to do would be to simply rename the filename. Add the root note in the filename. This prevents you from having to do anything fancy with sample editors. That's what the "fineNAME" options are about. Then there's the option to just place them on the next key starting with some key. This way - you wouldn't be sure which sample is processed first. They may not end up in an order that makes sense. But that's an option - the one you're using now.

1) Copy to file's root key - This will take a sample from an imported file and place the note on the keyboard using the root note assigned by the sample file's (wav, for example) meta data. I would think if there is no meta data for root note - then it would pick one for you (and it would always be the same, not pick a random place).

2) Copy to fileNAME's root key - Assuming the filename contains a MIDI note value - I'm guessing in decimal - they don't say, the note will be placed at this MIDI note location. You have to edit the .WAV (or other sample format) fileNAME to contain a note. I'm sure if there is no note in the filename (like "my sample.wav" ) - Samplerobot will assume some default - just like "1) " I believe an example of a sample with the root key number in the filename would be 041.wav would be a root note (in the filename) of F1. "F1" is the MIDI note designation. 000.wav would be MIDI root note of C-2. 127.wav would be a MIDI root note of G8. Probably SampleRobot supports decimal numbers - that's a standard convention. Maybe it will handle hex if using 0x00 through 0x7F maybe 00h through 7Fh. Maybe not. Cannot be "10" hex since "10" in hex is 16 in decimal. Hex has to be noted in a way Samplerobot can distinguish - and not sure how that's parsed. Maybe Samplerobot handles MIDI note names - like C3, D#5. Not sure exactly - you'll have to try different things and see what works. Most sample software lets you put arbitrary text before the root key in the filename. Like MyGreatSample_041.wav You'll have to experiment to see what it accepts because it's not spelled out for you.

3) Copy to file's root key (if available), else to filename's root key - This will place the sample on the root location pointed to by the meta-data information (inside the sample file) UNLESS that key is already mapped. If some other sample has mapped that key - then the filename will be used to map the key. That's my interpretation at least.

4) Copy starting at key 28 (e0) in no special order - Root note as defined by the file's meta data is ignored. The filename information is ignored. Sample Robot will just start at the defined note (decimal 28 in this case) and assign samples from the list of imported files to the next open slot 28, 29, 30 ... etc. And order is not enforced by anything special. Maybe filename sort order? Not sure how Samplerobot imports and processes. I think this option is there if you just want to land the samples anywhere on your keyboard - and your keyboard has an easy facility to re-order. Some keyboards may have this - Montage doesn't fit this category.

... as a side note - the Montage Operation Manual has a typo. It calls one of the options: "Copy to filename's root key (if available), else to filename's root key". It has two options in an "else" clause where both options are the same. Logically, it doesn't compute. Just a typo. Version 4 of Samplerobot documentation (not for Montage) AND the screenshot you provided both have "file" (without "name" ) to the left of the else and "filename" to the right of the else. Two different things. "File" means meta data and "filename" means text in the filename.

Circling back to "copy" in Samplerobot's world. "Copy" means "stick your sample at this location". Not a file operation or cloning or anything. It's confusing and probably some of the confusion is due to translation. The "copy" section is determining the method to figure out where to stick a sample within the keybank.

Thanks Jason for your explanation, it was very helpful and slowly I understood a bit. I followed your instruction and renamed the file G because G is the note in the wav file, when imported it still failed. Please see below the screen shot. Thanks

Martin

Apologies Jason, I have renamed to G3 and it seemed to work, just G alone is not good enough. But to be honest, this is just a temporary solution by playing the note of the instrument from MP3, recorded and converted to a wav file which in a way it lost a lot of the actual feel and sound of the instrument. I think Yamaha should have done is releasing the far east sample pack for people from Vietnam, China...

Trust me, working from scratch is not always easy especially for people like myself who is not good at sampling, I have spent many hours in the past trying to sample some of the sound I wanted but still struggled to get any close to.

BTW, there seems no forum for SampleRobot, I even emailed them but haven't got a reply yet.

Thanks again Jason.

Martin

The question is: why are you attempting to use Sample Robot for this task? That is what I don’t understand. If you have samples of the instruments, why do you need a Sample Robot? Just load the .wav data directly to the instrument and construct your Waveform in the MONTAGE.

Hi Bad Mister, I did try your suggestion but here is the problem:

As you predicted, because there is only one note C3 in this case the sound quality was really bad and got worst after one octave either direction and also the note cut shorter the higher I play and at one point all the keys sound the same.

So if I can record different pitches and put all into SampleRobot and I hope that would give me a better quality as you suggested. This is only a hope so I will try to experience this and see how it goes...

Thanks

Martin

"G" does not mean anything in any dimension. All MIDI notes have a number after a letter - possibly a minus - and possibly a "#" sign, maybe a "b". "G" by itself is not a hex digit - so it fails that. It's not a decimal number. Looks like the MIDI note names (C-2 through G8) are supported through your experimentation.

If you wanted to use decimal numbers, here's a chart: http://www.somascape.org/midi/basic/notes.html

... digging this up from the grave - there may be another thread after this one which also had the rhetorical question of why one would use Sample Robot to import wav files - when possibly doing this directly on Montage would be the right tool for the job.

What Montage does not have is a way to detect the best loop points and touch up the meta data in order to ensure your sample file has loop points for Montage to consume. It's possible that you would grab .wav files which do not have the loop points - or generate your own "raw" wav files - also without loop points. And then want to be able to have this sound in the decay1/decay2 phase loop "forever".

Montage is well tasked for importing one-shot audio samples and manipulating those (no loops).

It's also well suited for handling audio files with loop points already embedded. That works out fine importing straight to Montage and bypassing SampleRobot.

However, if you have audio samples that need loop points - SampleRobot is a fine tool for this task. It has provisions for determining where the loop points should be set and then exports those into the final save file for import to Montage.

Arguably, SampleRobot is "better" than tools used in the past for setting loop points of raw samples that need them.

So I believe there's a window of opportunity of using SampleRobot on audio files - and this is one reason why it would be "bad" if, hypothetically, SampleRobot removed this feature in order to force the hand of the user to use Montage's native audio file import features. I know no one is saying to remove this feature - but the suggestion was nearly to never use this feature which seems to be putting it on the chopping block - at least figuratively.