mySoftware [Updates]
newProducts [YOK]
Recording a Performance as Audio to the XF Sequencer
We will start our discussion of real time recording of a Motif XF PERFORMANCE as an audio sample (to SDRAM) with a discussion of the multiple arpeggiators. A Performance can be a combination of as many as four VOICES (each Voice is placed in a ‘PART’). Each PART can have it own arpeggiator and can access 5 different arpeggio types selectable via the [SF1]-[SF5] buttons. Over the past 10 years the arpeggiator in the Motif-series has evolved and in each new version something new was introduced. In order to get a clearer picture of why we are even talking about recording a PERFORMANCE as audio, it will be necessary to talk about the arpeggios.
There is this issue you may run into - we’ll call it: The arpeggio/sequencer conundrum
As you explore the XF you will discover that there are many different types of arpeggio. You may know there are 7,881 different arpeggio types, but there is more to them then just pre-recorded phrases. They are flexible, interactive, dynamic and can be quite complex. You know that an XF VOICE can have as many as 8 multi-samples (called ELEMENTS). Each Element can be thought of as an “oscillator” - it could be a complete instrument or it could be a specific building block used to construct a complete instrument.
You assign an arpeggiator to the entire VOICE (all Elements of the VOICE typically will respond). This has been true in all 10 years of the Motif series. The 256 arpeggios of the original Motif have evolved into something quite different by the time you get to the XF. The original Motif was a 4-Element VOICE architecture - when an arpeggio was assigned to a VOICE, the entire Voice responded. The current XF has 8 Elements and while it is still true that the entire VOICE shares the selected arpeggio, special arpeggio patterns were developed so that you can have some Elements play normally while other Elements respond to the arpeggiator. Arpeggios are created using note-on events and/or controller data. The ‘note-on’ arpeggios, quite naturally, include specific note-on velocity information, in addition to the note number and note duration. This velocity, note number, and gate information is used to target specific Elements within a Voice such that if you play a pad sound with Elements that have a VELOCITY Range of 1-112… and you can have a group of percussive sounding Elements that are Velocity Limited to a range of 113-127. You can sustain a pad sound with the keyboard, while certain Elements “dance” to the arpeggio’s rhythm. This type of compelling VOICE was not possible in the 2001 original Motif; it would have required combining several Voices into a Performance.
The Motif-series evolved. It is important to note that you can tailor each arpeggio type to an amazing degree with the ARP PLAY FX. Arpeggio Playback Effects allow you to adjust the feel and attitude of the pattern of notes being generated by the arpeggiator. Swing adjustments include not just timing (where the note-on events occur) but the duration (gate time) of notes, and velocity (how loud they sound). It is possible to increase or decrease the ‘swung note’, if you will. This can dramatically adjust how an arpeggio plays.
This level of complexity at the VOICE level is one of the very cool, cool things about the Motif XS/XF_ MOX synth engine. If you explore the Musical Effect VOICE category in the basic sound library, you will find these multi-dimensional VOICES - they are very much like “mini-PERFORMANCES” because they sound like combinations of completely different sounds. “Vanilla”, “Paris at Night”, “On My Way to UK”, “Everlasting Glory”, “Festival of Harmony” are a few favorites, that typify this unique type of VOICE. Please see the referenced (linked) articles for more in depth details.
Here it will be important to understand what the arpeggio KEY MODE parameter means and how it works. The arpeggiators are located physically between the keybed of the XF and the tone generator of the XF. This signal flow is important to understand. The key presses travel to the arpeggio block before going to the sequencer. This allows you to choose which event gets recorded - the key press that triggers the arppegio or the note, or notes, as output by the arpeggiator.
The Motif XF’s internal sequencer can record either you playing the keys (normal record, called the “direct” notes) or it can record the output of the arpeggiator (the “sorted” notes). It cannot record both simultaneously. And this is fine for most normal arpeggio situations: take, for example, a guitar strumming arpeggio pattern or a bass line or even a drum groove… when you press a key to ‘trigger’ the start of the arpeggio pattern, you’re not really interested in recording the actual note or notes that you used. Say you play a Cminor7 chord and hold it for 6 beats. You are actually more interested in having the sequencer record the “sorted notes” (the ones being generated by the arpeggiator); not the notes you used to trigger the arp pattern. In the case of the strummed guitar, you definitely would like to discard that dotted whole note, Cminor7 chord, being held throughout - and just keep the XF arp’s strummed guitar output.
Creative thinking: Ever notice when you are playing one of those guitar arps, that even though you are holding down a chord, the notes under your fingers are strumming away? How come you do not hear the held notes? How come you can hear the very keys you are holding, lifting up and being retriggered? Ever think about how the sequencer will handle that situation?
Because MIDI is not sound what you hear is not what is recorded - MIDI is a series of coded messages that are written to the sequencer, the tone generator will interpret these messages and recreate the sound. The ‘conundrum’ we mentioned is that, to the sequencer the held notes appear as NOTE-ON events, the sorted notes of the arpeggio pattern appear as NOTE-ON events. It is possible to have a KEY receive, simultaneously, a message that says, “Play: key “C3” at measure 001, beat 01, clock 000, with a velocity of 126, and duration of 16 beats, 000 clocks”. And at the exact same time receive a message that says: “Play: key “C3” at measure 001, beat 01, clock 000, with a velocity of 37, and duration of 00 beats, 212 clocks”
This is illogical. This is why you cannot record an arpeggio that requires both “direct” and “sorted” note-on events, without some chaos. This is why you want to use the Integrated Sampler’s RESAMPLE function to record your Performance play when this type of combination arpeggio type is used.
When recording normal bass line arps, and even drum grooves… you actually want to keep the direct notes hidden. So recording arps as MIDI data works well for all arpeggios with the KEY MODE set “sort”. It also applies to the arps with KEY MODE set to “thru” ("Thru" is a version of the ‘sort’ method - where the order in which the keys are pressed changes the order in which notes are added to the arpeggio pattern). The issue arises when the KEY MODE is “sort+direct” or “thru+direct”.
The “direct” KEY MODE was developed to allow the notes that you play to be recorded - and was necessary for the “CONTROLLLER” category arpeggio types. Arpeggios were expanded beyond just generating notes… a whole category of controller arpeggio patterns were introduced where pan position, volume (expression), filter and other controller data is stored in the arp pattern. These patterns are not note data at all, they are recordings of rhythmic controller movements. The control change messages are typically synchronized to the Tempo, so you can have gated effects that stutter step the sound in a rhythmically sync’d manner. The note-on messages generated by the keys that you press are allowed to go directly to the tone engine. The arp data be it panning (cc010) or filter (cc071/cc074) or volume change (cc011) is applied to the sound the tone generator is playing. So in this case (the case of “CTRL” cateogry arps) you must set the KEY MODE to “direct” or you will hear nothing. After all you cannot hear a Pan change message by itself, nor can you hear a Volume change message by itself, nor can you hear Filter changes by themselves, you can only hear their influence on a synth sound.
In the second generation Motif (the ES, circa 2003), the “sort+direct” and “thru+direct” arpeggio KEY MODES were introduced - allowing for a more complex VOICE response to arpeggio data. The original Motif (2001) had the typical arpeggio types: up, down, down/up, and some rhythmic seq and controller stuff. In the ES you could have much more complex arps where even velocity output could be precisely controlled to go and get a particular articulation or effect (Velocity Zone arpeggios were introduced in the ES). The Velocity Range of some Elements were set to precisely respond to particular arp pattern data, (as in the MegaVoice arps and voices, and certain Musical Effects category Voices) such that some Elements could act as pads (and play normally and be held by the keys you were pressing) while others would only respond when a specific velocity that was being sent by the arp pattern. This created a unique sound: pad+arp… where you actually could play and hear the pad Elements normally, while the “hands of the arpeggio” played the percussive sound, rhythmically. And all this within a single VOICE! There is a category of arpeggio types called “Z.Pad” that exploit this very powerful architectural improvement - and the programmers dug in. Z.Pad means velocity zoned arpeggio data which is designed so that the velocity is specific enough to play a specific multi-sample within the assigned sound.
These very complex Voices added an extra dimension to Motif XS/XF pads… You can read and learn more about this topic here: Introduction to VELOCITY ZONE Arpeggios.
The internal sequencer can still only record one type of note data or the other. Either the directly played trigger notes or the notes as output by the arpeggiator (sorted or thru’d).
Solutions include:
_ Recording the musical performance as audio (this can be to the Motif XF’s own sequencer or to your favorite DAW, if you have the capability to record audio externally). That is what we will cover here in this article.
_ Recording the musical performance as MIDI by recording the sorted Elements separate from the pad Elements. (This method requires deconstructing the VOICES that are using “sort+direct” KEY MODE type and recording them together and those that are simply using the “sort” mode and recording them separately). This requires a bit of programming and has a degree of difficulty. It will be covered separately.
_ If you are recording to an external sequencer you maybe able to record just the trigger notes and play those back to the XF via MIDI. Instead of recording the “sorted” note output, simply record the trigger notes and use the external sequencer to take your place ‘playing’ the arp patterns (this would require that you leave the ARPEGGIO ON/OFF button ON , because you will be triggering the ARPs, via MIDI, from an external device).
By the time the Motif-series reached the third generation (Motif XS circa 2007) arpeggio data had evolved yet again to allow recording of the PERFORMANCE directly to Sequencer (PERFORMANCE RECORD) in real time. This innovation greatly enhanced workflow - allowing composers to “jump start” their compositions by playing multiple Arp patterns simultaneously and recording them to the sequencer in real time. It allowed you to play from the keyboard on a single channel, yet output data to four separate Tracks and it automatically remapped the channel and PART data for you.
Again, these even more complex functions required a change in the way the sequencer and the KEY MODES worked.
No longer does the note data in the Motif XS/XF’s internal sequencer trigger the arpeggiator. The arpeggiators can only be triggered by actual key presses on its keyboard or by note data arriving by external MIDI… be that via the 5-pin MIDI jacks, via USB-MIDI or FW-MIDI (i.e., not from note data that is in the internal sequencer).
Let’s take a look at a unique new addition to the story: A Performance that features 4 PARTS, with two using the “sort” KEY MODE and the other two using the “sort+direct” KEY MODE.
Recall the following PERFORMANCE USER 1:007(A07) “Purple Stoneâ€
The four PARTS are under control of the arpeggiators:
PARTS 1 and 3 are set to KEY MODE = “sort” (and need to be recorded without their trigger notes)
PARTS 2 and 4 are set to KEY MODE = “sort+direct” (and cannot be recorded to the sequencer easily)
You can find this as follows:
Press [EDIT]
Press a Track button [1]-[4] to select a PART for editing
Press [F2] ARP MAIN
Here you can see the main arpeggio parameters for this particular PART. At [F3] ARP OTHER, you can adjust how the arpeggio behaves.
This is the conundrum: How to capture both sorted and direct data?
If you want to use a PERFORMANCE that combines multiple KEY MODES to the internal sequencer - I recommend that you record it as an AUDIO data. The XF comes with 128MB of SDRAM to which you can record your playing performance in perfect synchronization with the sequencer tempo.
Here’s how:
RECORDING a “live” PERFORMANCE to the Internal Sequencer as Audio
Call up USER 1: 007(A07) Purple Stone
Theory of Operation:
1) In a blank SONG or PATTERN mode program we will COPY the PERFORMANCE so that it occupies the first 4 PARTS of the MIXING setup - you can play it just as you did in PERFORMANCE mode… all four PARTS will be assigned to MIDI channel 1; and you will be able to play them all when you select Track 1.
2) In the INTEGRATED SAMPLER we’ll setup to “resample” a live keyboard performance directly as audio.
3) The Integrated Sampling Sequencer will punch us out, then it will create a new USER SAMPLE VOICE, assign it to PART 1, and it will create a Note-On event to precisely trigger the playback of the audio sample
4) We will then reclaim PARTS 2, 3, 4 and return their MIDI Receive Channels for normal use.
The end result will be an accurate recording of your playing the Performance - it will use just one note-on event to trigger the entire thing (as opposed to the hundreds it would take if recorded as MIDI data). It will use just one PART and the data will occupy just one Track.
Why it is recommended to record as audio: The XF introduced several new PERFORMANCES that utilize multiple KEY MODE arpeggios. Additionally, they use the [AF1] and [AF2] buttons controlling the XA CONTROL parameter (Expanded Articulation Control) to create dramatic changes in the program. As you perform you can hit the [AF1] and [AF2] buttons to bring in different timbres/rhythms and even change the sound of a partiuclar rhythmic component. Moving the MOD WHEEL will also have dramatic affect on the feel of the resulting output.
Step-by-step
Call up an empty SONG or PATTERN
Press [MIXING]
Press [JOB]
Press [F3] COPY
Press [SF3] PERF
Select “USER1: 007(A01) Purple Stone” and mark the attributes you wish to COPY from the original PERFORMANCE to the current MIXING setup
Press [ENTER] to execute
Doing so you have now assigned the four PARTS to the first four PARTS of the MIX - all assigned to Receive Channel 1 - so you can play them together from the XF keybed. In the screen shot below we have elected to view “4 PARTS” simultaneously so you can easily see all four PARTS are initially set to MIDI channel 1 - you can toggle the view between “1 PART” and “4 PART” via the [SF5] button.
On the main SONG or PATTERN screen you will see the Drum Kit VOICE (R&B Kit 3) assigned to TRACK 1, while TRACKS 2, 3 and 4 appear blank - this is because the PARTS 2, 3 and 4 are assigned to be triggered together by MIDI Receive Channel 1. Tracks 2, 3 and 4 no longer have a PART assigned to them.
Select TRACK [1] - you can now play the PERFORMANCE same as you did in PERFORMANCE mode.
Because we will be recording a real-time performance to the sampler, we need to setup a metronome for SAMPLING COUNT-IN. When sampling audio you only want the COUNT-IN to alert you when to begin playing - then the metronome must STOP. After the count-in we do not want to record it clicking away - we are recording audio - the Performances arpeggio will give us the tempo. The XF has a special Sampling Count-in specifically for this reason.
Press [SEQ SETUP]
Press [F1] CLICK
Set the parameter SAMPLING COUNT-IN to the number of measures that is comfortable for you, 1-8 measures.
Press [INTEGRATED SAMPLING]
Select a blank WAVEFORM number
Press [F6] REC>
Setup the SAMPLE SOURCE and DESTINATION:
Record Type = “sample+note”
Input Source = “resample”
Mono/Stereo = “stereo”
Frequency = “44.1kHz”
Key = “C3”
Track = “01”
By using “sample+note” - the ISS will create a USER SAMPLE VOICE, assign it to the mix and create a MIDI note-on event to trigger the playback of that audio.
By using “resample” we are able to record audio directly from all available inputs to the XF sampler (including playing the keyboard directly, as we will be doing).
Press [F6] STANDBY
On the Standby screen you can set the TRIGGER MODE, etc., and you can review your current setup. Make sure you always check the current MEASURE location of the sequencer (upper right corner) - use the TRANSPORT buttons to return to the top: 001
Trigger mode = “meas” (measure)
Set a punch in and punch out region - this is recommended because starting and stopping the recording on cue by automation gives you one less thing to worry about. Say you want to record 16 measures of the PERFORMANCE data… set the MEASURE = 001-017
Always set the the punch out point to the top of the NEXT measure (one more than the number of complete measures you want to record). If you are targeting a PATTERN, it will not let you go farther than the current LENGTH setting of the PATTERN.
The “RECORD MONITOR” and “RECORD GAIN” parameters can be set on this screen as well.
RECORD MONITOR is subjective and does not affect record levels, only listening volume
RECORD GAIN is objective and does affect record levels, experment by playing your Performance and increase the gain if necessary. +6dB or +12 dB boosts are available; -6dB and -12dB cuts are available. The meter (lower right corner) will give you a stereo reading of the level for Left/Right channels… in general, increase so that the loudest peaks reach nearly to the top of the scale.
Press [F5] START
STANDBY will change to “WAITING...”
Press the PLAY button on the XF transport to begin the COUNT-IN
Begin to play at the appropriate time according to the number of COUNT-IN measures you set.
When the Punch out measure is reached - recording will stop automatically.
You can press the [SF1] AUDITION button to hear your performance. If acceptable press [ENTER] to fix. If not, Press [EXIT] and do it over. Making sure you return the sequencer to measure 001.
If you have a keeper… press [ENTER]
The ISS (Integrated Sampling Sequencer) will create the USER SAMPLE VOICE (SMPL 001), and the NOTE-ON EVENT (default velocity is 100, you can increase or decrease this VELOCITY value to change the playback VOLUME of your audio clip). The duration (or GATE TIME) of the NOTE-ON EVENT will be the exact number of beats to HOLD the audio sample for the number of measures you created with the PUNCH IN/OUT points.
Return to SONG/PATTERN mode (whichever you have been using)
Before you PLAYBACK your sequence, you will want to return PARTS 2, 3 and 4 to RECEIVE on MIDI Channels 2, 3 and 4 respectively - or else you will hear your audio sample playing back 4 times (quite unnecessary since we have created an audio clip).
Press [MIXING]
Press [EDIT]
Press [Track 2] to view PART 2 parameters
Press [F1] VOICE
Press [SF1] VOICE
Set the MIDI RECEIVE CHANNEL = 2
press [TRACK 3]
Set the MIDI RECEIVE CHANNEL = 3
press [TRACK 4]
Set the MIDI RECEIVE CHANNEL = 4
Now you can playback your audio clip. You are also free to use Tracks 2, 3 and 4 with PARTS 2, 3 and 4.
Final Notes:
RESAMPLE as your INPUT SOURCE allows all available inputs to be recorded by the Integrated Sampling Sequencer… this includes the A/D INPUT, and the FW input, in addition to your real time play on the keyboard.
Hope that helps.
Permalink