Skip to content

Latest commit

 

History

History
118 lines (106 loc) · 6.42 KB

README.md

File metadata and controls

118 lines (106 loc) · 6.42 KB
Raw

ACRCloud iOS SDK

Overview

ACRCloud provides services such as Music Recognition, Broadcast Monitoring, Custom Audio Recognition, Copyright Compliance & Data Deduplication, Live Channel Detection, and Offline Recognition etc.

Requirements

Follow one of the tutorials to create a project and get your host, access_key, and access_secret.

Identify Music or TV with iOS SDK

This demo shows how to identify music ( songs ) or detect live TV channels by recorded sound with ACRCloud iOS SDK. Contact us if you have any questions or special requirements about the SDK: support@acrcloud.com

Preparation

  • The newest ACRCloud iOS SDK which contains both ObjectC and Swift demo projects.
  • If you want to recognize music, you need an Audio Recognition project. ( See How to Recognize Music )
  • If you want to detect tv channels, you need a Live Channel Detection project. ( See How to Detect Live TV Channels )
  • Save the information of “host”, “access_key”, “access_secret” of your project.
  • Make sure you have Xcode installed.

Test the demo

  • Download the ACRCloud iOS SDK package and unzip it.
  • Open either ACRCloudDemo or ACRCloudDemo_Swift
  • Update accessKey, host, and accessSecret in ViewController with the information of your project.
  • Run the demo project to test recognizing contents in the buckets of your project.

How to use the iOS SDK

1, Init recognition configure

    _config = [[ACRCloudConfig alloc] init];
    _config.accessKey = @"your project access key";
    _config.accessSecret = @"your project secret key";
    _config.host = @"your project host";
    _config.protocol = @"https";
    //if you want to identify your offline db, set the recMode to "rec_mode_local"
    _config.recMode = rec_mode_remote;
    if (_config.recMode == rec_mode_local || _config.recMode == rec_mode_both)
        _config.homedir = [[[NSBundle mainBundle] resourcePath] stringByAppendingPathComponent:@"acrcloud_local_db"];
        
    //if you use One-time Recording Session Recognition, please set these block.
    __weak typeof(self) weakSelf = self;
    _config.stateBlock = ^(NSString *state) {
        [weakSelf handleState:state];
    };
    _config.volumeBlock = ^(float volume) {
        //do some animations with volume
        [weakSelf handleVolume:volume];
    };
    //result callback. when the sdk gets a result or error, it will call this block.
    _config.resultBlock = ^(NSString *result, ACRCloudResultType resType) {
        [weakSelf handleResult:result resultType:resType];
    };

recMode is depending on the type of your project,

  • rec_mode_remote is for Audio & Video Recognition, Live Channel Detection, Hybrid Recognition, it’s online recognition
  • rec_mode_local is for Offline Recognition, please put the offline database ( such as “acrcloud_local_db” ) into your app project’s workspace.

2, Create a recognition client

  _client = [[ACRCloudRecognition alloc] initWithConfig:_config];

3, Recognition

One-time Recording Session Recognition

Click the “Start” Button and the App Demo will begin to record and recognize. When it detects a result, the "config.resultBlock" will be called, the app demo will stop and display the result. In the progress of recording and detecting, you can stop (click the “Stop” Button”) this recognition at any time.

-(void)startRecordRec;
-(void)stopRecordRec;

Open Prerecording Recognition

Open prerecording will make the recognition much faster. If you want open this feature, call "-(void)startPreRecord:(NSInteger)recordTime" before "startRecordRec" The parameter recordTime is the prerecording time. The recommended value is 3000-4000

  [_client startPreRecord:3000]

Low-Level Function

File/PCM/Fingerprint Recognition

if you recognize audio data or get the audio fingerprint, the audio format should be RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16-bit, mono 8000 Hz, you also can use the resample function to convert the audio data to what we need.

-(NSString*)recognize:(char*)buffer len:(int)len;
-(NSString*)recognize:(NSData*)pcm_data;
-(NSString*)recognize_fp:(NSData*)fingerprint;
+(NSData*)get_fingerprint:(NSData*)pcm
+(NSData*)get_fingerprint:(char *)pcm len:(int)len

Humming Recognition

if you recognize humming data, the audio format should be RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16 bit, mono 8000 Hz, you also can use the resample function to convert the audio data to what we need.

//Get humming fingerprint.
+(NSData*)get_hum_fingerprint:(NSData*)pcm;
//recognize the humming fingerprint
-(NSString*)recognize_hum_fp:(NSData*)fingerprint;

Resample Function

Convert your audio format to RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16-bit, mono 8000 Hz

+(NSData*) resample:(char*)pcm
                len:(unsigned)len
         sampleRate:(unsigned)sampleRate
           nChannel:(short)nChannel;
+(NSData*) resample_bit32:(char*)pcm
                      len:(unsigned)bytes
               sampleRate:(unsigned)sampleRate
                 nChannel:(short)nChannel
                  isFloat:(bool)isFloat;