Wave Corrector

Convert Vinyl Records and Tapes to Digital Audio

Ganymede Test & Measurement
HomeVersionsSamplesOpinionsDownloadFAQ'sSupportForumsFeedbackTutorialUser ManualPart 1Part 2Part 3Part 4Part 5Part 6Part 7LinksAudio TopicsGanymede


User Manual Part 3.  Processing Audio Files



3.1 File Types

Wave Corrector supports both compressed and uncompressed audio formats. Compression reduces the disc space required to store waveform data and is necessary if large volumes of audio are to be permanently stored on hard disk. There are two types of compression that can be used. Lossless compression permits a bit-perfect copy of the original wave file be to recovered. Hence, lossless compression is preferred by users who want to retain the maximum possible fidelity in their recordings. Lossy compression, on the other hand, uses psycho-acoustic processing to produce smaller files than can be achieved with lossless compression. Lossy compression will degrade your audio files to some extent but the degradation is very little if a sensible compression ratio is applied.

Wave Corrector will load files in the uncompressed Microsoft .wav format and the losslessly compressed Monkey's Audio .ape format. Lossy compression formats are not supported for loading.

Files loaded into Wave Corrector must also conform to the following requirements:

   1. Bit depth: 16 or 24 bits
   2. Sampling Rate: 44.1, 48, 88.2 or 48kHz
   3. 1 or 2 channels (ie mono or stereo)


For saving, Wave Corrector uses the above .wav and .ape and also the Ogg Vorbis .ogg format.

Ogg Vorbis is an open source audio compression system similar to MP3 but without the licensing restrictions imposed by the MP3 patent. Independent tests have also indicated that Ogg Vorbis produces better audio quality than the common MP3 encoders Ogg Vorbis is open source and is free of ‘digital rights management’ issues which can compromise library security if, for example, you need to back up your library onto a second computer. For these reasons we recommend it over other encoding systems.

Wave Corrector also supports the use of external command line encoders such as LAME.EXE for the production of MP3 files. This support give you the flexibility to save in any format for which a command line encoder exists. Note, this option allows you to use the command line versions of the ogg and ape encoders. These give you a wider range of encoding options than are present in Wave Corrector's built-in encoders. See §3.4 Using an External Encoder for more information.


File Type

Advantages

Disadvantages

Wave (.wav)

Fast load and save.

Best possible audio quality.

Compatible with most audio software.

Inefficient use of disc space.

Does not support textual tagging.

Ape (.ape)

Supports tagging.

Best possible audio quality.

Efficient use of disc space.

Slower saving and loading.

Limited compatibility with other programs.

Ogg Vorbis (.ogg)

Supports tagging.

Very efficient use of disc space.

Slower saving.

Limited compatibility with other programs.

Some loss of audio quality.

.ape and .ogg formats are supported by PC audio applications such as Foobar2000, Winamp, and Media Jukebox. Ganymede Test & Measurement recommends these applications for playing your music files.

3.2 Sample Rate and Bit Depth

Wave Corrector supports sample rates between 44.1 kHz and 96 kHz and word lengths of 16 and 24 bit. The following list shows how the formats are used in different audio applications.

All these formats produce very high quality results, much better than the compressed audio available from Internet radio or Satellite or Terrestrial Digital broadcasting.

Although 16-bit and 44.1 or 48 kHz sampling gives excellent quality results, there are advantages to using higher bit and sample rates when transferring audio from the analogue domain. This is because the extra quality gives more headroom when digitally processing your recordings. Also, the higher sample rate allows Wave Corrector's click detection to be more discriminating. For these reasons, it is recommended to make your original recordings in a higher sample rate format if your soundcard supports this.

Sample Rate Conversions

Many computer audio programs and operating systems are able to convert between different sample rates. However, the quality of these conversions is often very poor. Therefore, for good quality, you should keep conversions to a minimum. If you intend making Audio CD's, you can record at 88.2kHz sampling rate and use Wave Corrector's built-in Down-Sample option to convert to 44.1 kHz. Because this is a simple 2:1 conversion, the quality of the recording will not be compromised.

If you need to do more complicated nont-integral conversions, we recommend you use the open source program SSRC which integrates with Wave Corrector via the 'External Encoder' interface. See §3.4 Using an External Encoder.

3.3 File Tagging in Wave Corrector

File tagging is the name given to the process of adding text fields to audio files. The fields typically contain information such as the title, artist, genre etc. Tags are used principally by audio library programs to sort and catalogue the items in the library.

Because Wave Corrector usually works with album files consisting of several tracks, it provides the option to apply 'tags' to the individual tracks of a file being processed. These tags can be saved in the output files produced by the program and can also be used to generate output file names using a customisable rule.

File tags cannot be saved in .wav files. This is because there is no standardised way to do this and file tags on wave files are misinterpreted by many external programs.

The following tags are supported by Wave Corrector:

  1. Track Title

  2. Track Artist

  3. Track Number

  4. Album Title

  5. Genre




3.4 Using an External Encoder

The use of external command line encoders expands the flexibility of Wave Corrector beyond its built-in ogg and ape encoders. External encoders allow you to save in the commonly used FLAC and MP3 formats, and several other encoders can also be used if desired.

Encoders can be downloaded from various sites on the Internet. Download links for all the encoders directly supported in Wave Corrector can be found on the Wave Corrector website at
Encoder Download Links

Encoders that the program recognises are displayed in the Encoder setup window. See External Encoder Setup.

Each time you start Wave Corrector, it automatically scans for encoders, searching the path specified at the top of the setup window.

To add an encoder, place it in somewhere in the auto-scan path. Alternatively, you can use the Add button in the setup window to add an encoder in any arbitrary location. Wave Corrector automatically sets a default command line and file extension.

The following list shows the automatically recognised encoders and their associated command lines:

See the §3.5, Command Line Help for an explanation of the command line parameters.

When you use an external encoder in Wave Corrector, you specify which encoder to use from a drop down list in the Save window (also in the Batch window).



3.5 Command Line Help


Although Wave Corrector will automatically generate a command line for most encoders, you may wish to customise this to meet your own requirements. For example, you may wish to try out different compression settings.


All command line programs come with documentation describing the commands they recognise and their required format. If you cannot find the documentation, you can open the windows command prompt (under accessories on the Start menu) and navigate to the directory where your encoder is located. Then type in name of the encoder followed by the command "-h"; for example, "lame.exe -h". By convention, this will display a help screen for the program.


The creation of a command line is best illustrated with an example.


From the LAME documentation, the usage is as follows:

lame [options] <infile> [outfile]


The square brackets '[ ]' mean that the item contained in them is optional.


A simple option is '-b' to set a bit rate; so for example, '-b 160' sets a bit rate of 160 kbit/sec.


Wave Corrector translates the command that you enter using 'tokens' for file names and textual tags. When Wave Corrector encounters a token, it replaces it with the corresponding value associated with the file. Tokens always start with the % character. The following tokens are recognised by Wave Corrector.



Note, the LAME encoder also allows you to include tag information, eg it uses --tt to add a title tag.

Therefore, for example, to add a title tag, you would use: --tt %T

A complete command line example is therefore:

-b 160 -tt %T %IN %OUT

The Compression Options dialogue is used to enter your command line.


Figure 7 Compression Settings Dialogue



3.6 Setting the Auto-Scan Options

Wave Corrector automatically scans the source file when it is first loaded or immediately after it is recorded if you select the 'Scan' option.

Note, if you find the results are not satisfactory, you can re-scan for clicks or tracks, correct the volume, or apply a range of digital filters via the Waveform menu. However, the bulk of the work is done for you automatically by Wave Corrector.

Auto-scan is performed according to a range of options that are available to the user. The options available are summarised below:

Source Type

The source type of 'Vinyl' or 'Tape' can be selected. 'Vinyl' enables click detection and correction; 'tape' enables the hiss filter.

Note if you wish to apply both click correction and the hiss filter, you should select ''vinyl', and after the file has been auto-scanned, manually apply the hiss filter via the Waveform menu (or use the toolbar button).

Click Threshold

This setting determines the sensitivity of the click detector. The click threshold can be set between 0 and 5 where '0?'is 'Off' ie no clicks will be detected. The default value is 3.

Hiss Filter

This setting determines the amount of hiss reduction applied by the hiss filter. Four settings are available between 0dB(off) and 20dB. The default value is 8dB.

Enable Track Splitting

This option causes the program to place markers at the start and end of each track it detects.

Gapless Track Boundaries

This setting is provided to enable the production of 'Disc-At-Once' recordings on your CDr.

In gapless mode, the end of a track is always made exactly adjacent to the start of the next track. It is also positioned on a 588-sample boundary. This enables the CDr burner to produce a seamless track change with no gap between tracks. You can only do this if your burning software supports 'Disc-At-Once'.

In normal (non-gapless) mode, you can cut out unwanted gaps between tracks. Remember that the CDr burner will normally insert its own 2-second silence between tracks.

3.7 Loading Files into Wave Corrector

Wave Corrector only processes files of the appropriate type. See §3.1 File Types for the loading restrictions.

To open a file for processing, select the File - Open File command. If a file is already being processed, you will be prompted whether to save or discard any outstanding corrections.

Note, Wave Corrector also provides the File - Restore Session command . This command is used to restore a complete list of click corrections and track boundaries together with filter and volume settings.

3.8 Batch Processing Wave Files

Sometimes, if you have a number of files to process, it is more convenient to process them all together as a batch job. With batch processing, you give the program a list of files to process and a sequence of tasks to be performed on each file. Then, the program will work its way through the list without any further intervention by the user.

To set up a batch job, select the File - Batch Open command

Any combination of the following tasks can be included in a batch job

  • Click removal

  • Digital filtering

  • Normalising

  • Channel balancing

  • Trimming silence

  • Converting format


You cannot do track splitting as part of a batch job as the process is designed to work with single 'song' files. However, it is possible to trim the silence from the start and end of each file.

Note, batch processing is also useful even if you are only working with a single file. In this case, you can use Batch Open to open the single file and you can leave the program to perform the sequence you require.

3.9 Recording Wave Files

Wave Corrector also provides the facility to record your own wave files direct from the Line-In port of your sound card.

Wave Corrector provides the facility to record your own wave files direct from the Line-In port of your soundcard.

Wave files are recorded in 16 or 24 bit stereo at a sample rate of 44.1, 48, 88.2 or 96kHz. If you are going to burn CDr's use a sampling rate of 44.1 kHz or 88.2 kHz. The the default sampling rate is 44.1kHz.

To start recording, select the File - Record New Wave File command. If a wave file is already being processed, you will be asked whether to save or discard the outstanding corrections. You will be prompted to supply a file name for your new recording and then the Record New File Dialogue will appear.

The 'Record New File' dialogue provides level monitoring and recorder controls with record timer and start/stop on signal options.

Wave Corrector uses the Windows 'Volume Control' applet to select the recording source and to set the recording level. To launch the applet click on the 'Set Volume' button in the Record Dialogue. Check the recording source you wish to use (usually 'line in'; and adjust the recording level using the associated slider.

When setting the recording level you should take account of the performance of your sound card. See §3.9 Recording Tips for more advice on this subject.

Once your file has been recorded, you can immediately scan it into Wave Corrector using the 'Scan Now' button in the Record Dialogue.

Important: If you are recording a tape or cassette and intend to use the Hiss Filter, please ensure that you include a short lead-in or lead-out period. Wave Corrector produces a 'noise profile' by analysing the noise in the lead-in or lead-out period.

3.10 Recording Tips

Recording Levels

To obtain the maximum benefit from Wave Corrector, recording levels should be as high as possible whilst ensuring that the signal does not distort. Some sound cards are unable to utilise the full 16-bit dynamic range of the system and start to clip before the maximum signal level is reached. For this reason it is sensible to make some test recordings before committing your recordings to CDr.

If you find there is audible distortion from your sound card when recording at maximum level, you should back off the recording level until the distortion reduces to an acceptable level. If the overall amplitude is now too low, you can use Wave Corrector's 'Normalise' function to correct the level before burning your CDr.

Left-Right Channel Reversal Problems

You should be aware that some CDr burning software does not correctly allocate the left and right channels to the CDr. This seems to be because the wave file format allows different size header blocks but that the CDr burning software does not take account of this when identifying the first sample to burn. Therefore, depending upon which wave recording program you use, the resulting CDr may or may not have the channels reversed. For example, using Adaptec Easy CD Creator 3.5b, the channels are reversed if you use Goldwave to make your original recording; but are correct if you use Wave Studio.

This is another reason to carry out test recordings before committing your record collection to CDr! If you find that the CDr channels are reversed, then the simplest way overcome the problem is to reverse the left/right leads between your record player deck and your sound card. This will, in effect, give you a double reversal that will solve the problem. Remember however to re-perform the check whenever you change either your CDr burning software or your wave file recording software.

To ensure that it does not compound this problem, Wave Corrector strictly maintains the header block size of all its output files to be the same as the header block size of the corresponding input file. This means that once you have performed the above check, you can use Wave Corrector without worrying about its effect on channel reversal.

3.11 Saving your Work

Wave Corrector supports various file types for saving your work.

If you want to break off from an editing session and return to it at a later time, you can save a session file. Session files contain a list of click corrections and track boundaries values together with filter and volume settings. The session file enables Wave Corrector to restore the editing session when required. Use the File - Save Session command to create a session file. Note, session files do not include any audio data. When you restore a session, the audio data is obtained form the original uncorrected audio file. Therefore never delete the original uncorrected audio file until all editing work is complete.

When editing is complete, you can save your work in one of three different built-in audio formats. You can also use an external encoding program to save to a file type not directly supported by Wave Corrector, eg MP3. See §3.1 File Types for help with deciding which file type to use. See also §3.3 Using an External Encoder.

If you are embarking on a complicated editing session, you can select File - Auto-Save command. This causes a session file to be saved once every five minutes. Also, with auto-save in operation, a session file will automatically be saved whenever you save an audio file (see below).

If you have finished making adjustments to click corrections, track boundaries etc., then you should save your results in one of the audio formats described above. Choose a format that can be read by your CDr burning software or by the other audio applications that you use.

Use the File - Save command if you want to continue editing after the save, or use the File - Save & Close command to automatically close the input file after the save operation is complete.

You will be prompted to specify how to name the output files and where you to store them. If you have assigned tags to your individual tracks (title, artist, genre etc) then you can generate file names based on these tags. Alternatively, you can use the default naming convention which is to take the original file name and add the suffix 'cor' followed by a pair of numerals equal to the track number. See §3.11 File Naming for guidance on the file naming features of Wave Corrector.

If any of the corrected file names already exist on your computer, you will be prompted whether to overwrite them.



3.12 File Naming

Wave Corrector provides a very versatile mechanism for naming your output files. File names are generated using a File Naming Rule that you either specify explicitly or select from a drop down list. The file naming rule consists of free text and tokens. The tokens are replaced by tags that you have entered using the Track Properties command.

The following tokens are used for rule generation:

TITLE, ARTIST, GENRE, ALBUM, INPUT, #

Whenever a token is encountered in a file naming rule, it is replaced by the corresponding tag from the track's properties.

An example will illustrate the application of a file naming rule. Suppose you have a track with the following properties:

Input File: LK4574a.wav
Genre: Popular
Album Title: Magnificent Pianos
Track Title: Love Letters
Track Artist: Ronnie Aldrich
Track Number: 4

The following table shows the output names generated by various naming rules.

File Naming Rule

Output File Name

INPUTcor#

LK4574acor04

ALBUM#

Magnificent Pianos04

ALBUM\# - TITLE

Magnificent Pianos\04 - Love Letters

GENRE\ALBUM\# - TITLE

Popular\Magnificent Pianos\04 - Love Letters



Note, you can include a nest of subfolders in a rule by including backslash ('\') characters. If a folder doesn't exist, it will be created.


3.13 Correcting the Volume

Wave Corrector provides tools for correcting both the overall volume and also the relative volume of the left and right channels (channel balance).

To correct the overall volume, use the Waveform - Normalise command. This command searches for the maximum sample value in your file and displays this value relative to the maximum possible value with the 16-bit CD audio standard. The value is expressed in decibels.

Use the slider control to select the new volume level you require. If you want the maximum possible volume, set the slider to '0dB'. The default setting is '-1dB' which corresponds to the volume of typical commercially produced CD's.

If you want to see or audition the section of the file where the maximum sample occurs, use the Waveform - Goto Max command. This centres the display on the maximum sample but does not invoke the normalise dialogue.

You can adjust the left - right channel balance using the Waveform - Channel Balance command. The command shows the average imbalance between the channels as a decibel value. If you choose auto-balance, a correction is applied to bring the average channel amplitude into balance. Alternatively, you can manually select a re-balancing value to apply using the slider.

You should be aware that some stereo recordings are deliberately unbalanced. Therefore, you should use the auto-balance option with caution.



3.14 Digital Filtering

Wave Corrector provides five different digital filters that can be used to correct deficiencies in either vinyl or tape/cassette source recordings. The filters can be applied singly or in any combination. The following filters are available.



Filtering is used for two principle reasons:


Of the filters available in Wave Corrector, the Hiss, Hum and Rumble filters are solely for noise removal and the Treble Control filter is solely for correcting the frequency response. The Graphic Equaliser, being user-defined is more versatile and can be used for both purposes.

The Hiss and Hum reduction filters use a special technique known as 'Noise Subtraction'. This method uses a 'noise profile' that represents the unwanted noise present in the file The noise profile is used to create a mask that is 'subtracted' from the entire file. Noise subtraction is very effective at removing unwanted noise whilst preserving the full range of frequencies in the underlying music. However, for the technique to be successful, there are some important precautions to be observed. See §3.14 Noise Profiles, for extra guidance.

The rumble filter is a simple high pass filter that removes the sub-audio frequencies below about 30Hz. Rumble is generated by some tone-arm/pickup cartridge combinations. When it is excessive, rumble can cause damage to loudspeakers so it is important to filter it out it you suspect its presence.

The hum filter removes the power line frequency (50/60Hz) and its harmonics. Like the hiss filter, the hum filter uses noise subtraction to help avoid removing too much of the musical content. However, you should only use the hum filter as a last resort. If there is hum present in your recordings, you should make every effort to isolate and remove the cause - usually incorrect grounding of one or more components in your recording set-up. Using the hum filter 'after the event' will inevitably lead to some degradation of music quality.

The treble control filter provides up to 12dB of treble lift or treble cut. You can experiment with this filter to obtain the optimum tonal balance from your vinyl, tape or cassette recordings. Cassette recordings, in particular, often need to be tonally corrected due to a mis-match between the Dolby-B pre-emphasis and de-emphasis characteristics. If you find that your cassettes sound too dull or too bright, you can correct them using this control. You may also find that you obtain good results by recording your cassettes with the Dolby turned off and using Wave Corrector's hiss reduction and treble control combination to correct the sound. Although this gives a theoretically false result, it can sometimes be more pleasing with some types or source material.

The graphic equaliser enables you to 'draw' a required frequency response with the mouse and have the filter applied to your wave file. See §3.16 Graphic Equaliser for more information.

3.15 Noise Profiles

As already mentioned, the Hiss and Hum reduction filters use a technique known as 'Noise Subtraction'. This technique takes a 'snapshot' of the unwanted noise that is to be removed. This snapshot is known as a 'noise profile' and it is used to generate a noise mask that is 'subtracted' from the entire file.

For the technique to be successful, it is very important that the noise profile consists of typical noise from the file. It must be taken from a quiet passage where there is noise present but no music. If there is any music at all present, then program will over-correct the noise and unpleasant artefacts are likely to be the result.

It is also important to remember that the noise must be constant throughout the file. If the noise varies throughout the file then it won't be possible to locate a typical noise profile. This will result in some sections being over-corrected and others under-corrected.

The technique is really only successful for removing tape hiss, hum or interfering tones. However, in the case of interfering tones, these might be more effectively removed with a notch filter. See §3.16 on the Graphic Equaliser/Filter.

Wave Corrector automatically generates a noise profile when you load a wave file. To do this it searches the first and last 30 seconds of the file, and if that is unsuccessful, it looks for a quiet section within the file (for example at a track change). To aid the program in its task, you should always ensure that your recordings include a few seconds of lead-in or lead-out. In this way you will help the program to locate a suitable profile. There are certain types of recording however, that it is not possible to auto-profile. An example is a live concert that starts with applause. Applause has very noise-like characteristics and it is likely to be mistaken by the program for actual noise. As, in the example above, this will result in the program over-correcting the noise and causing artefacts.

If the program cannot find a suitable profile, it issues a warning that the profile may be unreliable. If you see this warning, or if you are at all suspicious that your recording may not be suitable for auto-profiling, then you should check the validity of the auto-profile and manually create a new one if necessary.

To check the auto-profile, use the Waveform - Edit Noise Profile command. This will centre the main display on the current profile and enable you to audition it..

If the auto-profile is unsatisfactory, you can move its position to somewhere more suitable.


3.16 Processing Shellac Records

Shellac records (eg 78's) are more difficult to process than vinyl because there is a greater amount of surface noise. This type of noise is particularly difficult to remove because it is too continuous for the click to detector to be fully effective but too spiky for the hiss filter to be fully effective

The best technique is therefore to run the click filter for several passes to remove most of the impulsive noise and then to run the hiss filter at its gentlest 8dB setting to mop up some of the remaining noise.

Large diameter shellac records will have a noise profile that varies as the record plays. This is because the stylus velocity increases as it travels towards the centre of the record. This causes the noise spectrum to skew linearly with time. To accommodate this effect, Wave Corrector allows you to specify two noise profiles, one at the start of the record and another at the end. The program calculates a moving average of these two profiles to provide an optimum match of the surface noise. Use the Waveform - Edit Noise Profile command to set a dual profile.

When recording shellac records, if possible run your turntable at a lower speed (eg 45rpm) and use Wave Corrector's speed conversion feature to compensate for this. This allows you to make a recording with a lower stylus velocity and hence lower distortion.


3.17 The Graphic Equaliser

    As well as its regular filters, Wave Corrector also provides a more versatile graphic equaliser/filter. This allows more advanced users to define their own filter to cope with special situations.

The frequency response of the graphic filter is defined by 256 equally spaced points on a frequency axis spanning 0 to 22kHz. Each of these points can be individually set by the user.

The most common uses for this filter will be to equalise the frequency response of an old recording; and/or to notch out single frequency tone interference. An example of the latter is the 19kHz pilot tone in FM broadcasts. The graphic filter allows you to set up at deep notch at this single frequency while passing all other frequencies unaltered.

Figure 8 Example Notch Filter Centred on 8kHz

The filter is very easy to adjust. You can add or move points with the left mouse button and remove them with the right mouse button. In this way you can quickly create a filter of your choice.


HOME   TOP PREV   NEXT