]> Interessengemeinschaft Software Engineering e. V.

Max-Planck-Str. 39 74081 Heilbronn Germany schoenebeck at software minus engineering dot org

LinuxSampler Developers LSCP The LinuxSampler Control Protocol (LSCP) is an application-level protocol primarily intended for local and remote controlling the LinuxSampler backend application, which is a sophisticated server-like console application essentially playing back audio samples and manipulating the samples in real time to certain extent.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in . This protocol is always case-sensitive if not explicitly claimed the opposite. In examples, "C:" and "S:" indicate lines sent by the client (front-end) and server (LinuxSampler) respectively. Lines in examples must be interpreted as every line being CRLF terminated (carriage return character followed by line feed character as defined in the ASCII standard ), thus the following example: C: "some line"    "another line" must actually be interpreted as client sending the following message: "some line<CR><LF>another line<CR><LF>" where <CR> symbolizes the carriage return character and <LF> the line feed character as defined in the ASCII standard. Due to technical reasons, messages can arbitrary be fragmented, means the following example: S: "abcd" could also happen to be sent in three messages like in the following sequence scenario: server sending message "a" followed by a delay (pause) with arbitrary duration followed by server sending message "bcd<CR>" again followed by a delay (pause) with arbitrary duration followed by server sending the message "<LF>" where again <CR> and <LF> symbolize the carriage return and line feed characters respectively.

LSCP will certainly be extended and enhanced by-and-by. Each official release of the LSCP specification will be tagged with a unique version tuple. The version tuple consists at least of a major and minor version number like: "1.2" In this example the major version number would be "1" and the minor version number would be "2". Note that the version tuple might also have more than two elements. The major version number defines a group of backward compatible versions. That means a frontend is compatible to the connected sampler if and only if the LSCP versions to which each of the two parties complies to, match both of the following rules: Compatibility: The frontend's LSCP major version and the sampler's LSCP major version are exactly equal. The frontend's LSCP minor version is less or equal than the sampler's LSCP minor version. Compatibility can only be claimed if both rules are true. The frontend can use the "GET SERVER INFO" command to get the version of the LSCP specification the sampler complies with.

LinuxSampler is a so called software sampler application capable to playback audio samples from a computer's Random Access Memory (RAM) as well as directly streaming it from disk. LinuxSampler is designed to be modular. It provides several so called "sampler engines" where each engine is specialized for a certain purpose. LinuxSampler has virtual channels which will be referred in this document as "sampler channels". The channels are in such way virtual as they can be connected to an arbitrary MIDI input method and arbitrary MIDI channel (e.g. sampler channel 17 could be connected to an ALSA sequencer device 64:0 and listening to MIDI channel 1 there). Each sampler channel will be associated with an instance of one of the available sampler engines (e.g. GigEngine, DLSEngine). The audio output of each sampler channel can be routed to an arbitrary audio output method (ALSA / JACK) and an arbitrary audio output channel there.

Main focus of this protocol is to provide a way to configure a running LinuxSampler instance and to retrieve information about it. The focus of this protocol is not to provide a way to control synthesis parameters or even to trigger or release notes. Or in other words; the focus are those functionalities which are not covered by MIDI or which may at most be handled via MIDI System Exclusive Messages.

There are two distinct methods of communication between a running instance of LinuxSampler and one or more control applications, so called "front-ends": a simple request/response communication method used by the clients to give commands to the server as well as to inquire about server's status and a subscribe/notify communication method used by the client to subscribe to and receive notifications of certain events as they happen on the server. The latter needs more effort to be implemented in the front-end application. The two communication methods will be described next.

This simple communication method is based on TCP. The front-end application establishes a TCP connection to the LinuxSampler instance on a certain host system. Then the front-end application will send certain ASCII based commands as defined in this document (every command line must be CRLF terminated - see "Conventions used in this document" at the beginning of this document) and the LinuxSampler application will response after a certain process time with an appropriate ASCII based answer, also as defined in this document. So this TCP communication is simply based on query and answer paradigm. That way LinuxSampler is only able to answer on queries from front-ends, but not able to automatically send messages to the client if it's not asked to. The fronted should not reconnect to LinuxSampler for every single command, instead it should keep the connection established and simply resend message(s) for subsequent commands. To keep information in the front-end up-to-date the front-end has to periodically send new requests to get the current information from the LinuxSampler instance. This is often referred to as "polling". While polling is simple to implement and may be OK to use in some cases, there may be disadvantages to polling such as network traffic overhead and information being out of date. It is possible for a client or several clients to open more than one connection to the server at the same time. It is also possible to send more than one request to the server at the same time but if those requests are sent over the same connection server MUST execute them sequentially. Upon executing a request server will produce a result set and send it to the client. Each and every request made by the client MUST result in a result set being sent back to the client. No other data other than a result set may be sent by a server to a client. No result set may be sent to a client without the client sending request to the server first. On any particular connection, result sets MUST be sent in their entirety without being interrupted by other result sets. If several requests got queued up at the server they MUST be processed in the order they were received and result sets MUST be sent back in the same order.

Result set could be one of the following types: Normal Warning Error Warning and Error result sets MUST be single line and have the following format: "WRN:<warning-code>:<warning-message>" "ERR:<error-code>:<error-message>" Where <warning-code> and <error-code> are numeric unique identifiers of the warning or error and <warning-message> and <error-message> are human readable descriptions of the warning or error respectively. Examples: C: "LOAD INSTRUMENT '/home/me/Boesendorfer24bit.gig" 0 0 S: "WRN:32:This is a 24 bit patch which is not supported natively yet." C: "GET AUDIO_OUTPUT_DRIVER_PARAMETER INFO ALSA EAR" S: "ERR:3456:Audio output driver 'ALSA' does not have a parameter 'EAR'." C: "GET AUDIO_OUTPUT_DEVICE INFO 123456" S: "ERR:9:There is no audio output device with index 123456." Normal result sets could be: Empty Single line Multi-line Empty result set is issued when the server only needed to acknowledge the fact that the request was received and it was processed successfully and no additional information is available. This result set has the following format: "OK" Example: C: "SET AUDIO_OUTPUT_DEVICE_PARAMETER 0 CHANNELS=4" S: "OK" Single line result sets are command specific. One example of a single line result set is an empty line. Multi-line result sets are command specific and may include one or more lines of information. They MUST always end with the following line: "." Example: C: "GET AUDIO_OUTPUT_DEVICE INFO 0" S: "DRIVER: ALSA"    "CHANNELS: 2"    "SAMPLERATE: 44100"    "ACTIVE: true"    "FRAGMENTS: 2"    "FRAGMENTSIZE: 128"    "CARD: '0,0'"    "." In addition to above mentioned formats, warnings and empty result sets MAY be indexed. In this case, they have the following formats respectively: "WRN[<index>]:<warning-code>:<warning-message>" "OK[<index>]" where <index> is command specific and is used to indicate channel number that the result set was related to or other integer value. Each line of the result set MUST end with <CRLF>. Examples: C: "ADD CHANNEL" S: "OK[12]" C: "CREATE AUDIO_OUTPUT_DEVICE ALSA SAMPLERATE=96000" S: "WRN[0]:32:Sample rate not supported, using 44100 instead."

This more sophisticated communication method is actually only an extension of the simple request/response communication method. The front-end still uses a TCP connection and sends the same commands on the TCP connection. Two extra commands are SUBSCRIBE and UNSUBSCRIBE commands that allow a client to tell the server that it is interested in receiving notifications about certain events as they happen on the server. The SUBSCRIBE command has the following syntax: SUBSCRIBE <event-id> where <event-id> will be replaced by the respective event that client wants to subscribe to. Upon receiving such request, server SHOULD respond with OK and start sending EVENT notifications when a given even has occurred to the front-end when an event has occurred. It MAY be possible certain events may be sent before OK response during real time nature of their generation. Event messages have the following format: NOTIFY:<event-id>:<custom-event-data> where <event-id> uniquely identifies the event that has occurred and <custom-event-data> is event specific. Several rules must be followed by the server when generating events: Events MUST NOT be sent to any client who has not issued an appropriate SUBSCRIBE command. Events MUST only be sent using the same connection that was used to subscribe to them. When response is being sent to the client, event MUST be inserted in the stream before or after the response, but NOT in the middle. Same is true about the response. It should never be inserted in the middle of the event message as well as any other response. If the client is not interested in a particular event anymore it MAY issue UNSUBSCRIBE command using the following syntax: UNSUBSCRIBE <event-id> where <event-id> will be replace by the respective event that client is no longer interested in receiving. For a list of supported events see . Example: the fill states of disk stream buffers have changed on sampler channel 4 and the LinuxSampler instance will react by sending the following message to all clients who subscribed to this event: NOTIFY:CHANNEL_BUFFER_FILL:4 [35]62%,[33]80%,[37]98% Which means there are currently three active streams on sampler channel 4, where the stream with ID "35" is filled by 62%, stream with ID 33 is filled by 80% and stream with ID 37 is filled by 98%. Clients may choose to open more than one connection to the server and use some connections to receive notifications while using other connections to issue commands to the back-end. This is entirely legal and up to the implementation. This does not change the protocol in any way and no special restrictions exist on the server to allow or disallow this or to track what connections belong to what front-ends. Server will listen on a single port, accept multiple connections and support protocol described in this specification in it's entirety on this single port on each connection that it accepted. Due to the fact that TCP is used for this communication, dead peers will be detected automatically by the OS TCP stack. While it may take a while to detect dead peers if no traffic is being sent from server to client (TCP keep-alive timer is set to 2 hours on many OSes) it will not be an issue here as when notifications are sent by the server, dead client will be detected quickly. When connection is closed for any reason server MUST forget all subscriptions that were made on this connection. If client reconnects it MUST resubscribe to all events that it wants to receive.

This chapter will describe the available control commands that can be sent on the TCP connection in detail. Some certain commands (e.g. "GET CHANNEL INFO" or "GET ENGINE INFO") lead to multiple-line responses. In this case LinuxSampler signals the end of the response by a "." (single dot) line.

White lines, that is lines which only contain space and tabulator characters, and lines that start with a "#" character are ignored, thus it's possible for example to group commands and to place comments in a LSCP script file.

Instances of drivers in LinuxSampler are called devices. You can use multiple audio devices simultaneously, e.g. to output the sound of one sampler channel using the ALSA audio output driver, and on another sampler channel you might want to use the JACK audio output driver. For particular audio output systems it's also possible to create several devices of the same audio output driver, e.g. two separate ALSA audio output devices for using two different sound cards at the same time. This chapter describes all commands to configure LinuxSampler's audio output devices and their parameters. Instead of defining commands and parameters for each driver individually, all possible parameters, their meanings and possible values have to be obtained at runtime. This makes the protocol a bit abstract, but has the advantage, that front-ends can be written independently of what drivers are currently implemented and what parameters these drivers are actually offering. This means front-ends can even handle drivers which are implemented somewhere in future without modifying the front-end at all. Note: examples in this chapter showing particular parameters of drivers are not meant as specification of the drivers' parameters. Driver implementations in LinuxSampler might have complete different parameter names and meanings than shown in these examples or might change in future, so these examples are only meant for showing how to retrieve what parameters drivers are offering, how to retrieve their possible values, etc.

Use the following command to get the number of audio output drivers currently available for the LinuxSampler instance: GET AVAILABLE_AUDIO_OUTPUT_DRIVERS Possible Answers: LinuxSampler will answer by sending the number of audio output drivers. Example: C: "GET AVAILABLE_AUDIO_OUTPUT_DRIVERS" S: "2"

Use the following command to list all audio output drivers currently available for the LinuxSampler instance: LIST AVAILABLE_AUDIO_OUTPUT_DRIVERS Possible Answers: LinuxSampler will answer by sending comma separated character strings, each symbolizing an audio output driver. Example: C: "LIST AVAILABLE_AUDIO_OUTPUT_DRIVERS" S: "ALSA,JACK"

Use the following command to get detailed information about a specific audio output driver: GET AUDIO_OUTPUT_DRIVER INFO <audio-output-driver> Where <audio-output-driver> is the name of the audio output driver, returned by the "LIST AVAILABLE_AUDIO_OUTPUT_DRIVERS" command. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. At the moment the following information categories are defined: DESCRIPTION - character string describing the audio output driver VERSION - character string reflecting the driver's version PARAMETERS - comma separated list of all parameters available for the given audio output driver, at least parameters 'channels', 'samplerate' and 'active' are offered by all audio output drivers The mentioned fields above don't have to be in particular order. Example: C: "GET AUDIO_OUTPUT_DRIVER INFO ALSA" S: "DESCRIPTION: Advanced Linux Sound Architecture"    "VERSION: 1.0"    "PARAMETERS: DRIVER,CHANNELS,SAMPLERATE,ACTIVE,FRAGMENTS, FRAGMENTSIZE,CARD"    "."

Use the following command to get detailed information about a specific audio output driver parameter: GET AUDIO_OUTPUT_DRIVER_PARAMETER INFO <audio> <prm> [<deplist>] Where <audio> is the name of the audio output driver as returned by the "LIST AVAILABLE_AUDIO_OUTPUT_DRIVERS" command, <prm> a specific parameter name for which information should be obtained (as returned by the "GET AUDIO_OUTPUT_DRIVER INFO" command) and <deplist> is an optional list of parameters on which the sought parameter <prm> depends on, <deplist> is a list of key-value pairs in form of "key1=val1 key2=val2 ...", where character string values are encapsulated into apostrophes ('). Arguments given with <deplist> which are not dependency parameters of <prm> will be ignored, means the front-end application can simply put all parameters into <deplist> with the values already selected by the user. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. There are information which is always returned, independently of the given driver parameter and there are optional information which is only shown dependently to given driver parameter. At the moment the following information categories are defined: TYPE - either "BOOL" for boolean value(s) or "INT" for integer value(s) or "FLOAT" for dotted number(s) or "STRING" for character string(s) (always returned, no matter which driver parameter) DESCRIPTION - arbitrary text describing the purpose of the parameter (always returned, no matter which driver parameter) MANDATORY - either true or false, defines if this parameter must be given when the device is to be created with the 'CREATE AUDIO_OUTPUT_DEVICE' command (always returned, no matter which driver parameter) FIX - either true or false, if false then this parameter can be changed at any time, once the device is created by the 'CREATE AUDIO_OUTPUT_DEVICE' command (always returned, no matter which driver parameter) MULTIPLICITY - either true or false, defines if this parameter allows only one value or a list of values, where true means multiple values and false only a single value allowed (always returned, no matter which driver parameter) DEPENDS - comma separated list of parameters this parameter depends on, means the values for fields 'DEFAULT', 'RANGE_MIN', 'RANGE_MAX' and 'POSSIBILITIES' might depend on these listed parameters, for example assuming that an audio driver (like the ALSA driver) offers parameters 'card' and 'samplerate' then parameter 'samplerate' would depend on 'card' because the possible values for 'samplerate' depends on the sound card which can be chosen by the 'card' parameter (optionally returned, dependent to driver parameter) DEFAULT - reflects the default value for this parameter which is used when the device is created and not explicitly given with the 'CREATE AUDIO_OUTPUT_DEVICE' command, in case of MULTIPLCITY=true, this is a comma separated list, that's why character strings are encapsulated into apostrophes (') (optionally returned, dependent to driver parameter) RANGE_MIN - defines lower limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, this parameter is often used in conjunction with RANGE_MAX, but may also appear without (optionally returned, dependent to driver parameter) RANGE_MAX - defines upper limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, this parameter is often used in conjunction with RANGE_MIN, but may also appear without (optionally returned, dependent to driver parameter) POSSIBILITIES - comma separated list of possible values for this parameter, character strings are encapsulated into apostrophes (optionally returned, dependent to driver parameter) The mentioned fields above don't have to be in particular order. Examples: C: "GET AUDIO_OUTPUT_DRIVER_PARAMETER INFO ALSA CARD" S: "DESCRIPTION: sound card to be used"    "TYPE: STRING"    "MANDATORY: false"    "FIX: true"    "MULTIPLICITY: false"    "DEFAULT: '0,0'"    "POSSIBILITIES: '0,0','1,0','2,0'"    "." C: "GET AUDIO_OUTPUT_DRIVER_PARAMETER INFO ALSA SAMPLERATE" S: "DESCRIPTION: output sample rate in Hz"    "TYPE: INT"    "MANDATORY: false"    "FIX: false"    "MULTIPLICITY: false"    "DEPENDS: card"    "DEFAULT: 44100"    "." C: "GET AUDIO_OUTPUT_DRIVER_PARAMETER INFO ALSA SAMPLERATE CARD='0,0'" S: "DESCRIPTION: output sample rate in Hz"    "TYPE: INT"    "MANDATORY: false"    "FIX: false"    "MULTIPLICITY: false"    "DEPENDS: card"    "DEFAULT: 44100"    "RANGE_MIN: 22050"    "RANGE_MAX: 96000"    "."

Use the following command to create a new audio output device for the desired audio output system: CREATE AUDIO_OUTPUT_DEVICE <audio-output-driver> [<param-list>] Where <audio-output-driver> should be replaced by the desired audio output system as returned by the "LIST AVAILABLE_AUDIO_OUTPUT_DRIVERS" command and <param-list> by an optional list of driver specific parameters in form of "key1=val1 key2=val2 ...", where character string values should be encapsulated into apostrophes ('). Note that there might be drivers which require parameter(s) to be given with this command. Use the previously described commands in this chapter to get this information. Possible Answers: "OK[<device-id>]" - in case the device was successfully created, where <device-id> is the numerical ID of the new device "WRN[<device-id>]:<warning-code>:<warning-message>" - in case the device was created successfully, where <device-id> is the numerical ID of the new device, but there are noteworthy issue(s) related (e.g. sound card doesn't support given hardware parameters and the driver is using fall-back values), providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Examples: C: "CREATE AUDIO_OUTPUT_DEVICE ALSA" S: "OK[0]" C: "CREATE AUDIO_OUTPUT_DEVICE ALSA CARD='2,0' SAMPLERATE=96000" S: "OK[1]"

Use the following command to destroy a created output device: DESTROY AUDIO_OUTPUT_DEVICE <device-id> Where <device-id> should be replaced by the numerical ID of the audio output device as given by the "CREATE AUDIO_OUTPUT_DEVICE" or "LIST AUDIO_OUTPUT_DEVICES" command. Possible Answers: "OK" - in case the device was successfully destroyed "WRN:<warning-code>:<warning-message>" - in case the device was destroyed successfully, but there are noteworthy issue(s) related (e.g. an audio over ethernet driver was unloaded but the other host might not be informed about this situation), providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example: C: "DESTROY AUDIO_OUTPUT_DEVICE 0" S: "OK"

Use the following command to count all created audio output devices: GET AUDIO_OUTPUT_DEVICES Possible Answers: LinuxSampler will answer by sending the current number of all audio output devices. Example: C: "GET AUDIO_OUTPUT_DEVICES" S: "4"

Use the following command to list all created audio output devices: LIST AUDIO_OUTPUT_DEVICES Possible Answers: LinuxSampler will answer by sending a comma separated list with the numerical IDs of all audio output devices. Example: C: "LIST AUDIO_OUTPUT_DEVICES" S: "0,1,4,5"

Use the following command to get current settings of a specific, created audio output device: GET AUDIO_OUTPUT_DEVICE INFO <device-id> Where <device-id> should be replaced by numerical ID of the audio output device as e.g. returned by the "LIST AUDIO_OUTPUT_DEVICES" command. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. As some parameters might allow multiple values, character strings are encapsulated into apostrophes ('). At the moment the following information categories are defined (independently of device): DRIVER - identifier of the used audio output driver, as also returned by the "LIST AVAILABLE_AUDIO_OUTPUT_DRIVERS" command CHANNELS - amount of audio output channels this device currently offers SAMPLERATE - playback sample rate the device uses ACTIVE - either true or false, if false then the audio device is inactive and doesn't output any sound, nor do the sampler channels connected to this audio device render any audio The mentioned fields above don't have to be in particular order. The fields above are only those fields which are returned by all audio output devices. Every audio output driver might have its own, additional driver specific parameters (see ) which are also returned by this command. Example: C: "GET AUDIO_OUTPUT_DEVICE INFO 0" S: "DRIVER: ALSA"    "CHANNELS: 2"    "SAMPLERATE: 44100"    "ACTIVE: true"    "FRAGMENTS: 2"    "FRAGMENTSIZE: 128"    "CARD: '0,0'"    "."

Use the following command to alter a specific setting of a created audio output device: SET AUDIO_OUTPUT_DEVICE_PARAMETER <device-id> <key>=<value> Where <device-id> should be replaced by the numerical ID of the audio output device as given by the "CREATE AUDIO_OUTPUT_DEVICE" or "LIST AUDIO_OUTPUT_DEVICES" command, <key> by the name of the parameter to change and <value> by the new value for this parameter. Possible Answers: "OK" - in case setting was successfully changed "WRN:<warning-code>:<warning-message>" - in case setting was changed successfully, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example: C: "SET AUDIO_OUTPUT_DEVICE_PARAMETER 0 FRAGMENTSIZE=128" S: "OK"

Use the following command to get information about an audio channel: GET AUDIO_OUTPUT_CHANNEL INFO <device-id> <audio-chan> Where <device-id> is the numerical ID of the audio output device as given by the "CREATE AUDIO_OUTPUT_DEVICE" or "LIST AUDIO_OUTPUT_DEVICES" command and <audio-chan> the audio channel number. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. At the moment the following information categories are defined: NAME - arbitrary character string naming the channel, which doesn't have to be unique (always returned by all audio channels) IS_MIX_CHANNEL - either true or false, a mix-channel is not a real, independent audio channel, but a virtual channel which is mixed to another real channel, this mechanism is needed for sampler engines which need more audio channels than the used audio system might be able to offer (always returned by all audio channels) MIX_CHANNEL_DESTINATION - numerical ID (positive integer including 0) which reflects the real audio channel (of the same audio output device) this mix channel refers to, means where the audio signal actually will be routed / added to (only returned in case the audio channel is mix channel) The mentioned fields above don't have to be in particular order. The fields above are only those fields which are generally returned for the described cases by all audio channels regardless of the audio driver. Every audio channel might have its own, additional driver and channel specific parameters. Examples: C: "GET AUDIO_OUTPUT_CHANNEL INFO 0 0" S: "NAME: studio monitor left"    "IS_MIX_CHANNEL: false"    "." C: "GET AUDIO_OUTPUT_CHANNEL INFO 0 1" S: "NAME: studio monitor right"    "IS_MIX_CHANNEL: false"    "." C: "GET AUDIO_OUTPUT_CHANNEL INFO 0 2" S: "NAME: studio monitor left"    "IS_MIX_CHANNEL: true"    "MIX_CHANNEL_DESTINATION: 1"    "." C: "GET AUDIO_OUTPUT_CHANNEL INFO 1 0" S: "NAME: 'ardour (left)'"    "IS_MIX_CHANNEL: false"    "JACK_BINDINGS: 'ardour:0'"    "."

Use the following command to get detailed information about specific audio channel parameter: GET AUDIO_OUTPUT_CHANNEL_PARAMETER INFO <dev-id> <chan> <param> Where <dev-id> is the numerical ID of the audio output device as returned by the "CREATE AUDIO_OUTPUT_DEVICE" or "LIST AUDIO_OUTPUT_DEVICES" command, <chan> the audio channel number and <param> a specific channel parameter name for which information should be obtained (as returned by the "GET AUDIO_OUTPUT_CHANNEL INFO" command). Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. There are information which is always returned, independently of the given channel parameter and there is optional information which is only shown dependently to the given audio channel. At the moment the following information categories are defined: TYPE - either "BOOL" for boolean value(s) or "INT" for integer value(s) or "FLOAT" for dotted number(s) or "STRING" for character string(s) (always returned) DESCRIPTION - arbitrary text describing the purpose of the parameter (always returned) FIX - either true or false, if true then this parameter is read only, thus cannot be altered (always returned) MULTIPLICITY - either true or false, defines if this parameter allows only one value or a list of values, where true means multiple values and false only a single value allowed (always returned) RANGE_MIN - defines lower limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, usually used in conjunction with 'RANGE_MAX', but may also appear without (optionally returned, dependent to driver and channel parameter) RANGE_MAX - defines upper limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, usually used in conjunction with 'RANGE_MIN', but may also appear without (optionally returned, dependent to driver and channel parameter) POSSIBILITIES - comma separated list of possible values for this parameter, character strings are encapsulated into apostrophes (optionally returned, dependent to driver and channel parameter) The mentioned fields above don't have to be in particular order. Example: C: "GET AUDIO_OUTPUT_CHANNEL_PARAMETER INFO 1 0 JACK_BINDINGS" S: "DESCRIPTION: bindings to other JACK clients"    "TYPE: STRING"    "FIX: false"    "MULTIPLICITY: true"    "POSSIBILITIES: 'PCM:0','PCM:1','ardour:0','ardour:1'"    "."

Use the following command to alter a specific setting of an audio output channel: SET AUDIO_OUTPUT_CHANNEL_PARAMETER <dev-id> <chn> <key>=<value> Where <dev-id> should be replaced by the numerical ID of the audio output device as returned by the "CREATE AUDIO_OUTPUT_DEVICE" or "LIST AUDIO_OUTPUT_DEVICES" command, <chn> by the audio channel number, <key> by the name of the parameter to change and <value> by the new value for this parameter. Possible Answers: "OK" - in case setting was successfully changed "WRN:<warning-code>:<warning-message>" - in case setting was changed successfully, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example: C: "SET AUDIO_OUTPUT_CHANNEL PARAMETER 0 0 JACK_BINDINGS='PCM:0'" S: "OK" C: "SET AUDIO_OUTPUT_CHANNEL PARAMETER 0 0 NAME='monitor left'" S: "OK"

Instances of drivers in LinuxSampler are called devices. You can use multiple MIDI devices simultaneously, e.g. to use MIDI over ethernet as MIDI input on one sampler channel and ALSA as MIDI input on another sampler channel. For particular MIDI input systems it's also possible to create several devices of the same MIDI input type. This chapter describes all commands to configure LinuxSampler's MIDI input devices and their parameters. Instead of defining commands and parameters for each driver individually, all possible parameters, their meanings and possible values have to be obtained at runtime. This makes the protocol a bit abstract, but has the advantage, that front-ends can be written independently of what drivers are currently implemented and what parameters these drivers are actually offering. This means front-ends can even handle drivers which are implemented somewhere in future without modifying the front-end at all. Commands for configuring MIDI input devices are pretty much the same as the commands for configuring audio output drivers, already described in the last chapter. Note: examples in this chapter showing particular parameters of drivers are not meant as specification of the drivers' parameters. Driver implementations in LinuxSampler might have complete different parameter names and meanings than shown in these examples or might change in future, so these examples are only meant for showing how to retrieve what parameters drivers are offering, how to retrieve their possible values, etc.

Use the following command to get the number of MIDI input drivers currently available for the LinuxSampler instance: GET AVAILABLE_MIDI_INPUT_DRIVERS Possible Answers: LinuxSampler will answer by sending the number of available MIDI input drivers. Example: C: "GET AVAILABLE_MIDI_INPUT_DRIVERS" S: "2"

Use the following command to list all MIDI input drivers currently available for the LinuxSampler instance: LIST AVAILABLE_MIDI_INPUT_DRIVERS Possible Answers: LinuxSampler will answer by sending comma separated character strings, each symbolizing a MIDI input driver. Example: C: "LIST AVAILABLE_MIDI_INPUT_DRIVERS" S: "ALSA,JACK"

Use the following command to get detailed information about a specific MIDI input driver: GET MIDI_INPUT_DRIVER INFO <midi-input-driver> Where <midi-input-driver> is the name of the MIDI input driver as returned by the "LIST AVAILABLE_MIDI_INPUT_DRIVERS" command. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. At the moment the following information categories are defined: DESCRIPTION - arbitrary description text about the MIDI input driver VERSION - arbitrary character string regarding the driver's version PARAMETERS - comma separated list of all parameters available for the given MIDI input driver The mentioned fields above don't have to be in particular order. Example: C: "GET MIDI_INPUT_DRIVER INFO ALSA" S: "DESCRIPTION: Advanced Linux Sound Architecture"    "VERSION: 1.0"    "PARAMETERS: DRIVER,ACTIVE"    "."

Use the following command to get detailed information about a specific parameter of a specific MIDI input driver: GET MIDI_INPUT_DRIVER_PARAMETER INFO <midit> <param> [<deplist>] Where <midit> is the name of the MIDI input driver as returned by the "LIST AVAILABLE_MIDI_INPUT_DRIVERS" command, <param> a specific parameter name for which information should be obtained (as returned by the "GET MIDI_INPUT_DRIVER INFO" command) and <deplist> is an optional list of parameters on which the sought parameter <param> depends on, <deplist> is a key-value pair list in form of "key1=val1 key2=val2 ...", where character string values are encapsulated into apostrophes ('). Arguments given with <deplist> which are not dependency parameters of <param> will be ignored, means the front-end application can simply put all parameters in <deplist> with the values selected by the user. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. There is information which is always returned, independent of the given driver parameter and there is optional information which is only shown dependent to given driver parameter. At the moment the following information categories are defined: TYPE - either "BOOL" for boolean value(s) or "INT" for integer value(s) or "FLOAT" for dotted number(s) or "STRING" for character string(s) (always returned, no matter which driver parameter) DESCRIPTION - arbitrary text describing the purpose of the parameter (always returned, no matter which driver parameter) MANDATORY - either true or false, defines if this parameter must be given when the device is to be created with the 'CREATE MIDI_INPUT_DEVICE' command (always returned, no matter which driver parameter) FIX - either true or false, if false then this parameter can be changed at any time, once the device is created by the 'CREATE MIDI_INPUT_DEVICE' command (always returned, no matter which driver parameter) MULTIPLICITY - either true or false, defines if this parameter allows only one value or a list of values, where true means multiple values and false only a single value allowed (always returned, no matter which driver parameter) DEPENDS - comma separated list of parameters this parameter depends on, means the values for fields 'DEFAULT', 'RANGE_MIN', 'RANGE_MAX' and 'POSSIBILITIES' might depend on these listed parameters, for example assuming that an audio driver (like the ALSA driver) offers parameters 'card' and 'samplerate' then parameter 'samplerate' would depend on 'card' because the possible values for 'samplerate' depends on the sound card which can be chosen by the 'card' parameter (optionally returned, dependent to driver parameter) DEFAULT - reflects the default value for this parameter which is used when the device is created and not explicitly given with the 'CREATE MIDI_INPUT_DEVICE' command, in case of MULTIPLCITY=true, this is a comma separated list, that's why character strings are encapsulated into apostrophes (') (optionally returned, dependent to driver parameter) RANGE_MIN - defines lower limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, this parameter is often used in conjunction with RANGE_MAX, but may also appear without (optionally returned, dependent to driver parameter) RANGE_MAX - defines upper limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, this parameter is often used in conjunction with RANGE_MIN, but may also appear without (optionally returned, dependent to driver parameter) POSSIBILITIES - comma separated list of possible values for this parameter, character strings are encapsulated into apostrophes (optionally returned, dependent to driver parameter) The mentioned fields above don't have to be in particular order. Example: C: "GET MIDI_INPUT_DRIVER_PARAMETER INFO ALSA ACTIVE" S: "DESCRIPTION: Whether device is enabled"    "TYPE: BOOL"    "MANDATORY: false"    "FIX: false"    "MULTIPLICITY: false"    "DEFAULT: true"    "."

Use the following command to create a new MIDI input device for the desired MIDI input system: CREATE MIDI_INPUT_DEVICE <midi-input-driver> [<param-list>] Where <midi-input-driver> should be replaced by the desired MIDI input system as returned by the "LIST AVAILABLE_MIDI_INPUT_DRIVERS" command and <param-list> by an optional list of driver specific parameters in form of "key1=val1 key2=val2 ...", where character string values should be encapsulated into apostrophes ('). Note that there might be drivers which require parameter(s) to be given with this command. Use the previously described commands in this chapter to get that information. Possible Answers: "OK[<device-id>]" - in case the device was successfully created, where <device-id> is the numerical ID of the new device "WRN[<device-id>]:<warning-code>:<warning-message>" - in case the driver was loaded successfully, where <device-id> is the numerical ID of the new device, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example: C: "CREATE MIDI_INPUT_DEVICE ALSA" S: "OK[0]"

Use the following command to destroy a created MIDI input device: DESTROY MIDI_INPUT_DEVICE <device-id> Where <device-id> should be replaced by the device's numerical ID as returned by the "CREATE MIDI_INPUT_DEVICE" or "LIST MIDI_INPUT_DEVICES" command. Possible Answers: "OK" - in case the device was successfully destroyed "WRN:<warning-code>:<warning-message>" - in case the device was destroyed, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example: C: "DESTROY MIDI_INPUT_DEVICE 0" S: "OK"

Use the following command to count all created MIDI input devices: GET MIDI_INPUT_DEVICES Possible Answers: LinuxSampler will answer by sending the current number of all MIDI input devices. Example: C: "GET MIDI_INPUT_DEVICES" S: "3"

Use the following command to list all created MIDI input devices: LIST MIDI_INPUT_DEVICES Possible Answers: LinuxSampler will answer by sending a comma separated list with the numerical Ids of all created MIDI input devices. Examples: C: "LIST MIDI_INPUT_DEVICES" S: "0,1,2" C: "LIST MIDI_INPUT_DEVICES" S: "1,3"

Use the following command to get current settings of a specific, created MIDI input device: GET MIDI_INPUT_DEVICE INFO <device-id> Where <device-id> is the numerical ID of the MIDI input device as returned by the "CREATE MIDI_INPUT_DEVICE" or "LIST MIDI_INPUT_DEVICES" command. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. As some parameters might allow multiple values, character strings are encapsulated into apostrophes ('). At the moment the following information categories are defined (independent of driver): DRIVER - identifier of the used MIDI input driver, as e.g. returned by the "LIST AVAILABLE_MIDI_INPUT_DRIVERS" command ACTIVE - either true or false, if false then the MIDI device is inactive and doesn't listen to any incoming MIDI events and thus doesn't forward them to connected sampler channels The mentioned fields above don't have to be in particular order. The fields above are only those fields which are returned by all MIDI input devices. Every MIDI input driver might have its own, additional driver specific parameters (see "GET MIDI_INPUT_DRIVER INFO" command) which are also returned by this command. Example: C: "GET MIDI_INPUT_DEVICE INFO 0" S: "DRIVER: ALSA"    "ACTIVE: true"    "."

Use the following command to alter a specific setting of a created MIDI input device: SET MIDI_INPUT_DEVICE_PARAMETER <device-id> <key>=<value> Where <device-id> should be replaced by the numerical ID of the MIDI input device as returned by the "CREATE MIDI_INPUT_DEVICE" or "LIST MIDI_INPUT_DEVICES" command, <key> by the name of the parameter to change and <value> by the new value for this parameter. Possible Answers: "OK" - in case setting was successfully changed "WRN:<warning-code>:<warning-message>" - in case setting was changed successfully, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example: C: "SET MIDI_INPUT_DEVICE_PARAMETER 0 ACTIVE=false" S: "OK"

Use the following command to get information about a MIDI port: GET MIDI_INPUT_PORT INFO <device-id> <midi-port> Where <device-id> is the numerical ID of the MIDI input device as returned by the "CREATE MIDI_INPUT_DEVICE" or "LIST MIDI_INPUT_DEVICES" command and <midi-port> the MIDI input port number. Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. At the moment the following information categories are defined: NAME - arbitrary character string naming the port The field above is only the one which is returned by all MIDI ports regardless of the MIDI driver and port. Every MIDI port might have its own, additional driver and port specific parameters. Example: C: "GET MIDI_INPUT_PORT INFO 0 0" S: "NAME: 'Masterkeyboard'"    "ALSA_SEQ_BINDINGS: '64:0'"    "."

Use the following command to get detailed information about specific MIDI port parameter: GET MIDI_INPUT_PORT_PARAMETER INFO <dev-id> <port> <param> Where <dev-id> is the numerical ID of the MIDI input device as returned by the "CREATE MIDI_INPUT_DEVICE" or "LIST MIDI_INPUT_DEVICES" command, <port> the MIDI port number and <param> a specific port parameter name for which information should be obtained (as returned by the "GET MIDI_INPUT_PORT INFO" command). Possible Answers: LinuxSampler will answer by sending a <CRLF> separated list. Each answer line begins with the information category name followed by a colon and then a space character <SP> and finally the info character string to that info category. There is information which is always returned, independently of the given channel parameter and there is optional information which are only shown dependently to the given MIDI port. At the moment the following information categories are defined: TYPE - either "BOOL" for boolean value(s) or "INT" for integer value(s) or "FLOAT" for dotted number(s) or "STRING" for character string(s) (always returned) DESCRIPTION - arbitrary text describing the purpose of the parameter (always returned) FIX - either true or false, if true then this parameter is read only, thus cannot be altered (always returned) MULTIPLICITY - either true or false, defines if this parameter allows only one value or a list of values, where true means multiple values and false only a single value allowed (always returned) RANGE_MIN - defines lower limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, this parameter is usually used in conjunction with 'RANGE_MAX' but may also appear without (optionally returned, dependent to driver and port parameter) RANGE_MAX - defines upper limit of the allowed value range for this parameter, can be an integer value as well as a dotted number, this parameter is usually used in conjunction with 'RANGE_MIN' but may also appear without (optionally returned, dependent to driver and port parameter) POSSIBILITIES - comma separated list of possible values for this parameter, character strings are encapsulated into apostrophes (optionally returned, dependent to device and port parameter) The mentioned fields above don't have to be in particular order. Example: C: "GET MIDI_INPUT_PORT_PARAMETER INFO 0 0 ALSA_SEQ_BINDINGS" S: "DESCRIPTION: bindings to other ALSA sequencer clients"    "TYPE: STRING"    "FIX: false"    "MULTIPLICITY: true"    "POSSIBILITIES: '64:0','68:0','68:1'"    "."

Use the following command to alter a specific setting of a MIDI input port: SET MIDI_INPUT_PORT_PARAMETER <device-id> <port> <key>=<value> Where <device-id> should be replaced by the numerical ID of the MIDI device as returned by the "CREATE MIDI_INPUT_DEVICE" or "LIST MIDI_INPUT_DEVICES" command, <port> by the MIDI port number, <key> by the name of the parameter to change and <value> by the new value for this parameter (encapsulated into apostrophes) or NONE (not encapsulated into apostrophes) for specifying no value for parameters allowing a list of values. Possible Answers: "OK" - in case setting was successfully changed "WRN:<warning-code>:<warning-message>" - in case setting was changed successfully, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example: C: "SET MIDI_INPUT_PORT_PARAMETER 0 0 ALSA_SEQ_BINDINGS='20:0'" S: "OK" C: "SET MIDI_INPUT_PORT_PARAMETER 0 0 ALSA_SEQ_BINDINGS=NONE" S: "OK"

The following commands describe how to add and remove sampler channels, associate a sampler channel with a sampler engine, load instruments and connect sampler channels to MIDI and audio devices.

An instrument file can be loaded and assigned to a sampler channel by one of the following commands: LOAD INSTRUMENT [NON_MODAL] '<filename>' <instr-index> <sampler-channel> Where <filename> is the name of the instrument file on the LinuxSampler instance's host system, <instr-index> the index of the instrument in the instrument file and <sampler-channel> is the number of the sampler channel the instrument should be assigned to. Each sampler channel can only have one instrument. Notice: since LSCP 1.2 the <filename> argument supports escape characters for special characters (see chapter "Character Set and Escape Sequences" for details) and accordingly backslash characters in the filename MUST now be escaped as well! The difference between regular and NON_MODAL versions of the command is that the regular command returns OK only after the instrument has been fully loaded and the channel is ready to be used while NON_MODAL version returns immediately and a background process is launched to load the instrument on the channel. The GET CHANNEL INFO command can be used to obtain loading progress from INSTRUMENT_STATUS field. LOAD command will perform sanity checks such as making sure that the file could be read and it is of a proper format and SHOULD return ERR and SHOULD not launch the background process should any errors be detected at that point. Possible Answers: "OK" - in case the instrument was successfully loaded "WRN:<warning-code>:<warning-message>" - in case the instrument was loaded successfully, but there are noteworthy issue(s) related (e.g. Engine doesn't support one or more patch parameters provided by the loaded instrument file), providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example:

A sampler engine type can be associated to a specific sampler channel by the following command: LOAD ENGINE <engine-name> <sampler-channel> Where <engine-name> is an engine name as obtained by the "LIST AVAILABLE_ENGINES" command and <sampler-channel> the sampler channel as returned by the "ADD CHANNEL" or "LIST CHANNELS" command where the engine type should be assigned to. This command should be issued after adding a new sampler channel and before any other control commands on the new sampler channel. It can also be used to change the engine type of a sampler channel. This command has (currently) no way to define or force if a new engine instance should be created and assigned to the given sampler channel or if an already existing instance of that engine type, shared with other sampler channels, should be used. Possible Answers: "OK" - in case the engine was successfully deployed "WRN:<warning-code>:<warning-message>" - in case the engine was deployed successfully, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example:

The number of sampler channels can change on runtime. To get the current amount of sampler channels, the front-end can send the following command: GET CHANNELS Possible Answers: LinuxSampler will answer by returning the current number of sampler channels. Example: C: "GET CHANNELS" S: "12"

The number of sampler channels can change on runtime. To get the current list of sampler channels, the front-end can send the following command: LIST CHANNELS Possible Answers: LinuxSampler will answer by returning a comma separated list with all sampler channels numerical IDs. Example: C: "LIST CHANNELS" S: "0,1,2,3,4,5,6,9,10,11,15,20"

A new sampler channel can be added to the end of the sampler channel list by sending the following command: ADD CHANNEL This will increment the sampler channel count by one and the new sampler channel will be appended to the end of the sampler channel list. The front-end should send the respective, related commands right after to e.g. load an engine, load an instrument and setting input, output method and eventually other commands to initialize the new channel. The front-end should use the sampler channel returned by the answer of this command to perform the previously recommended commands, to avoid race conditions e.g. with other front-ends that might also have sent an "ADD CHANNEL" command. Possible Answers: "OK[<sampler-channel>]" - in case a new sampler channel could be added, where <sampler-channel> reflects the channel number of the new created sampler channel which should be used to set up the sampler channel by sending subsequent initialization commands "WRN:<warning-code>:<warning-message>" - in case a new channel was added successfully, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example:

A sampler channel can be removed by sending the following command: REMOVE CHANNEL <sampler-channel> Where <sampler-channel> should be replaced by the number of the sampler channel as given by the "ADD CHANNEL" or "LIST CHANNELS" command. The channel numbers of all subsequent sampler channels remain the same. Possible Answers: "OK" - in case the given sampler channel could be removed "WRN:<warning-code>:<warning-message>" - in case the given channel was removed, but there are noteworthy issue(s) related, providing an appropriate warning code and warning message "ERR:<error-code>:<error-message>" - in case it failed, providing an appropriate error code and error message Example:

The front-end can ask for the number of available engines by sending the following command: GET AVAILABLE_ENGINES Possible Answers: LinuxSampler will answer by sending the number o