### Align Audio Files with Detailed Timing Analysis

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

Use the -matchalign option to get a detailed match of timing offset and clock skew. For precise measurements, consider adjusting -hoptime and -timesize, and use -oversamp 1 to avoid deliberate file offsetting.

```bash
audfprint -dbase fpdbase -match query.mp3 -density 15 -matchalign 1
```

```bash
audfprint -dbase fpdbase -match query.mp3 -density 15 -matchalign 1 -alignoutdir alignout
```

--------------------------------

### Interpreting Query Results

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

Example output from a query operation. The 'matching-count' indicates the number of common aligned fingerprints; a higher number suggests a better match. 'match-time' reports the delay between the start of the reference item and the start of the aligned query.

```text
Hash table read from fpdbase (13 tracks, 847 hashes)
query.mp3 (5.8 s) analyzed to 80 hashes
query.mp3 1 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3 2 0.070
matched 1 tracks (5.763 secs, 80 hashes, 13.8817 hashes/sec) in 0.4 sec = 0.077 x RT
done
Hash table read from fpdbase (13 tracks, 847 hashes)
query.mp3 (5.8 s) analyzed to 208 hashes
query.mp3 1 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3 8 0.070
matched 1 tracks (5.763 secs, 208 hashes, 36.0924 hashes/sec) in 0.2 sec = 0.038 x RT
done
```

--------------------------------

### Create Audio Fingerprint Database

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

Use this command to initialize a new fingerprint database and add audio files from a list or a single URL. Options like -cleardbase and -matchonaddthresh can be specified during initialization.

```matlab
audfprint -dbase fpdbase -cleardbase 1 -addlist reflist.txt
```

```matlab
audfprint -dbase fpdbase -matchonaddthresh 5 -add http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/01-Nine_Lives.mp3
```

--------------------------------

### Query Audio Database

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

Use this command to match a query soundfile against an existing database. The output provides details on the top matches, including file paths, matching counts, and match times.

```bash
audfprint -dbase fpdbase -match query.mp3
```

```bash
audfprint -dbase fpdbase -match query.mp3 -density 15
```

--------------------------------

### Audio Fingerprint Database Creation Output

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

This output shows the process of adding audio files to the database, including the number of hashes generated and the time taken. It also indicates when tracks are skipped because they already exist in the database.

```text
Target density = 7 hashes/sec
01-Apr-2014 23:09:05 Adding #1 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/01-Nine_Lives.mp3...10.0 s, 58 hashes
01-Apr-2014 23:09:06 Adding #2 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/02-Falling_In_Love.mp3...10.0 s, 75 hashes
01-Apr-2014 23:09:06 Adding #3 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/03-Hole_In_My_Soul.mp3...10.0 s, 69 hashes
01-Apr-2014 23:09:06 Adding #4 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/04-Taste_Of_India.mp3...10.0 s, 77 hashes
01-Apr-2014 23:09:06 Adding #5 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3...10.0 s, 48 hashes
01-Apr-2014 23:09:07 Adding #6 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/06-Something_s_Gotta_Give.mp3...10.0 s, 44 hashes
01-Apr-2014 23:09:07 Adding #7 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/07-Ain_t_That_A_Bitch.mp3...10.0 s, 73 hashes
01-Apr-2014 23:09:07 Adding #8 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/08-The_Farm.mp3...10.0 s, 60 hashes
01-Apr-2014 23:09:07 Adding #9 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/09-Crash.mp3...10.0 s, 114 hashes
01-Apr-2014 23:09:07 Adding #10 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/10-Kiss_Your_Past_Good-bye.mp3...10.0 s, 57 hashes
01-Apr-2014 23:09:07 Adding #11 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/11-Pink.mp3...10.0 s, 57 hashes
01-Apr-2014 23:09:08 Adding #12 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/12-Attitude_Adjustment.mp3...10.0 s, 74 hashes
01-Apr-2014 23:09:08 Adding #13 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/13-Fallen_Angels.mp3...10.0 s, 41 hashes
added 13 tracks (130 secs, 847 hashes, 6.5154 hashes/sec) in 3.0 sec = 0.023 x RT
Hash table saved to fpdbase (13 tracks, 847 hashes)
done
Hash table read from fpdbase (13 tracks, 847 hashes)
Target density = 7 hashes/sec
01-Apr-2014 23:09:12 Adding #1 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/01-Nine_Lives.mp3... 
*** skipping http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/01-Nine_Lives.mp3 - matched to http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/01-Nine_Lives.mp3
0.0 s, 0 hashes
added 1 tracks (0 secs, 0 hashes, NaN hashes/sec) in 0.2 sec = Inf x RT
done
```

--------------------------------

### Audfprint Command Line Usage

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

This section outlines the general usage and available command-line options for the audfprint utility.

```APIDOC
## Audfprint Command Line Options

All parameters to `audfprint` are specified in the command line via "-optionname value" pairs. The full set of options is:

### General Usage
```
audfprint -help
```

### Database Options
- **-dbase <file>**
  - Description: The reference database file.
- **-cleardbase 0/1**
  - Description: Create a new database with options...
  - **-nhashbits <num>** (Optional, default: 20)
    - Description: log_2 of hash table size.
  - **-maxnentries <num>** (Optional, default: 100)
    - Description: maximum number of entries per bin.
  - **-timesize <num>** (Optional, default: 16384)
    - Description: Maximum value of abs time index.
  - **-hoptime <time>** (Optional, default: 0.02322)
    - Description: Hop between time windows.
  - **-targetsr <rate>** (Optional, default: 11025)
    - Description: Resample to this Sample Rate (SR).
  - **-jenkins 0/1** (Optional, default: 0)
    - Description: use jenkins hash on hashes.

### Audio Processing Options
- **-density <num>** (Optional, default: 7.0)
  - Description: Target hashes per second.
- **-add <file ...>**
  - Description: Sound file(s) to add to database.
- **-addlist <file>**
  - Description: List of audio files to add to database.
- **-adddir <dir>**
  - Description: Watch this directory and add any files.
- **-addskip <count>** (Optional)
  - Description: Skip this many initial files in addlist.
- **-addcheckpoint <count>** (Optional)
  - Description: Save database every <count> tracks.
- **-matchonaddthresh <thr>** (Optional, default: 0)
  - Description: Don't add files if match >= thr.
- **-remove <name ...>**
  - Description: Delete named track(s) from dbase.
- **-removelist <file>**
  - Description: Delete tracks named in file from dbase.
- **-match <file ...>**
  - Description: Audio file(s) to match.
- **-matchlist <file>**
  - Description: List of audio files to match against database.
- **-matchdir <dir>**
  - Description: Watch this directory and match any files.
- **-skip <time>**
  - Description: Drop time from start of each sound.
- **-maxdur <time>** (Optional, default: 0=all)
  - Description: Truncate soundfiles at this duration.
- **-oversamp <num>** (Optional)
  - Description: Oversampling factor for queries (0=special).
- **-userawcounts 0/1** (Optional)
  - Description: Count hits without applying synchrony filter.

### Matching Parameters
- **-matchmaxret <num>** (Optional, default: 5)
  - Description: Maximum number of matches to report for each query.
- **-matchmincount <num>** (Optional, default: 0)
  - Description: Minimum count of common hashes to report.
- **-matchminprop <num>** (Optional, default: 0.1)
  - Description: Minimum proportion of max hash count to report.
- **-maxtofilter <num>** (Optional, default: 100)
  - Description: Maximum tracks retained per hash hit.
- **-matchwidth <bins>** (Optional, default: 1)
  - Description: Width of "modal time skew" search.
- **-matchalign 0/1** (Optional, default: 0)
  - Description: Calculate time skew alignment for match.
- **-warpmax <factor>** (Optional)
  - Description: Repeat filtering with warp factors to this.
- **-warpstep <step>** (Optional)
  - Description: Step size for repeating with warps.

### Output Options
- **-alignoutdir <dir>**
  - Description: Write aligned versions of queries here.
- **-alignoutext <ext>** (Optional, default: .mp3)
  - Description: Default extension for aligned outputs.
- **-list <regexp>** (Optional, default: . for all)
  - Description: List matching files in the database.
- **-quiet 0/1** (Optional)
  - Description: Suppress status messages.
- **-out <file>** (Optional)
  - Description: File to write matches out to (stdout).
- **-outdir <dir>**
  - Description: Write match reports to this directory.

### Error Messages
- **No dbase specified!**
  - Description: Indicates that a database file was not provided when required.

```

--------------------------------

### Output of Audio Alignment Analysis

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

This output shows the results of an audio alignment analysis, including hash comparisons, best match timings, and the equivalent sox command for resampling and aligning the audio. It also indicates the processing time relative to real-time.

```text
Hash table read from fpdbase (13 tracks, 847 hashes)
query.mp3 (5.8 s) analyzed to 208 hashes
48;208 hashes, 6 in common, 6 within 0.064 s of median for warp -0.02
Best match for time T sec in tk 1 (ref) is -0.070+(1+0.000000)*T in tk 2 (qry)
sox <query.wav> <alignout.wav> speed 1.000000 delay 0.06966 0.06966 trim 0 13.2663
query.mp3 1 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3 8 0.070
matched 1 tracks (5.763 secs, 208 hashes, 36.0924 hashes/sec) in 1.1 sec = 0.184 x RT
done
Hash table read from fpdbase (13 tracks, 847 hashes)
query.mp3 (5.8 s) analyzed to 208 hashes
48;208 hashes, 6 in common, 6 within 0.064 s of median for warp -0.02
Best match for time T sec in tk 1 (ref) is -0.070+(1+0.000000)*T in tk 2 (qry)
sox <query.wav> <alignout.wav> speed 1.000000 delay 0.06966 0.06966 trim 0 13.2663
Warning: popenw not available, writing temporary file
Wrote skewed audio to alignout/http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3 from query.mp3 with offset=-0.070s skew=1.000000 maxdur=13.0s
Equivalent sox command:
sox query.mp3 alignout/http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3 speed 1.000000 delay 0.06966 trim 0 13
Warped query written to alignout/http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3
query.mp3 1 http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3 8 0.070
matched 1 tracks (5.763 secs, 208 hashes, 36.0924 hashes/sec) in 1.8 sec = 0.317 x RT
done
```

--------------------------------

### Output After Removing Item from Database

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

This output shows the database being updated after an item is removed, followed by a query that now finds no hits, indicating the item was successfully removed.

```text
Hash table read from fpdbase (13 tracks, 847 hashes)
Hash table saved to fpdbase (12 tracks, 799 hashes)
done
Hash table read from fpdbase (12 tracks, 799 hashes)
query.mp3 (5.8 s) analyzed to 80 hashes
*** NO HITS FOUND ***
matched 1 tracks (5.763 secs, 80 hashes, 13.8817 hashes/sec) in 0.2 sec = 0.034 x RT
done
```

--------------------------------

### Remove Audio Items from Database

Source: http://www.ee.columbia.edu/~dpwe/resources/matlab/audfprint

Use the -remove option to delete a specific track's hashes from the database. This is useful for cleaning up or correcting the database.

```bash
audfprint -dbase fpdbase -remove http://labrosa.ee.columbia.edu/~dpwe/tmp/Nine_Lives/05-Full_Circle.mp3
```

```bash
audfprint -dbase fpdbase -match query.mp3
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.