SplitAWav

Easy to use solution for vinyl recording

Download SplitAWav (for Window 98 and above)

Version 2.1: December 2004:karaoke option working, various fixups

Version 1.9, July25 2004: document missing pictures, security feature didn't work on NT/XP/2000

Please note you will need version 1.9 or greater in order to get a registered version later, though you can do this upgrade at any time.

Heavily Revised July 10th 2004. [Click removal, sound card compatibility, other features]

New feature: Live, added 14th July.

Important note:

To protect my investment of time and effort, you may only download and Splitawav once.  The downloaded program has a hard-wired date set, which will be updated from time to time, so that generally you will have a couple of months of usage. Once that deadline is reached, you can only download new versions, if any are released, but if the version is the same, the fresh download will not work - otherwise anyone could simply download afresh every time the expiry was reached. Please support this inexpensive piece of shareware. 

Features

I wrote SplitAWav (if you like, a sister to AddAWav) because I was trying to preserve some of my old records onto CD's, by sampling them in.  I wanted an easy and automatic way of making each track its own wav file, so that cutting a CD would not just produce one long track.  I later decided it would be nice to be able to sample directly and split on the fly, and then also do normalisation (as an option), and remove clicks and pops.  I also didn’t want to have to mess around a lot every time I had a new record to sample.

I looked around and tried a program already out there, but it failed to work correctly in two very simple examples, so I wrote my own.

There's really two main areas to Splitawav, which can be used separately in different ways, or combined.  The first of these is splitting of the tracks.  Now, on a clean digital source this is an easy thing to do, because the tracks will end with nice zeroes in the sample data, but on a record, there is noise between the tracks.  Records may have quiet sections, and it's important to try to tell the track breaks apart from mere silence on the recording.  Splitawav detects track breaks, and creates separate wav files, which can then be used to create CD's with tracks that match, in most cases, the original record's layout.  The second function of Splitawav is that of a click remover.  This is quite effective and on all but the near-highest setting works without detriment to the original recording.  The click (and pop) remover can be used in a stand-alone fashion - live - not recording at all, but merely listening to the music, and reducing the amount of clicks and pops heard.

Using the program

Important note

Splitawav is designed only for stereo 16 bit, 44100 samples per second audio.  This format is the standard rate for CDs.  Use this format only, until you have made your desired wav files. The results of not using this format may be incorrect output files (wrong speed etc)

Understanding the Dialog Box of SplitAWav

There are three sections - an area at the top for general settings, the "split wave" section, and the "record" section. The middle "split wave" section is for splitting wave files, and for declicking wave files. Click the appropriate button to achieve this.  Note that declicking a wave file destroys the original file, replacing it with the declicked file.  You can also press the LIVE button to listen to audio and have the clicks removed.  This depends on the slider bar.

The lower section allows you to set a filename (and directory) either by typing it directly or choosing a file by pressing the base filename button.  Then use the "record and split" button.  To apply spike removal, use the option at the top, and the same applies for normalization (making each track as high a volume level as possible) and silence removal. Note that silence removal is not always recommended, and neither is normalization, which for tiny tracks will enormously amplify the noise in your recording, so you will have to use these with caution - they are useful sometimes.

Notes

The gap value is expressed in seconds - rule of thumbs is - use 2 for pop/rock music, 4 for classical music.  I've tried a dozen or so LPs and this worked sensibly for all of them.  When changing this value, be aware that the granularity of the value is a quarter of a second.  In other words, if you put in 1.6, the value 1.5 will be used - because 1/4 second is the amount of data handled at a time.  A couple of "concept" albums don't always find the expected track breaks, because the music simply runs on, or because the music is out of the ordinary.  What I'd say is that although there are times that track breaks aren't completely correct, they are usually sane - either extra tracks where a section of music has component parts, or merged tracks where the breaks are minimal.

As well as the gap value, there is an overall level value (the value of which is a kind of accumulation value over 1/4 second) set at the top of SplitAWav's dialog box.  This value defaults to 10800 - which works well on two audio cards at least (one internal).  You can use the calibrate setting to work out your own best value for this.  You can also study the log file c:\templog.txt when "write debug file" is set, the values shown against the times are the values detected, higher values indicating a likely track break.

 

Splitawav is designed to make it easy to sample in your old records by automatically detecting track breaks. It is not much good for cassettes, or other weird media (unless you've previously recorded . It also doesn’t do anything about hiss or hum, at least in this version, and severe crackling may be beyond any hope (but seems to be improved markedly). There are other tools that do this out there.  It does have a click and pop remover, can remove introductory silence (noisy silence, let’s say) and can normalise tracks to make the volume level at maximum.

Setup

The program is easy to use - download and install it, then choose the wav file. Alternatively, wire up your record player, and record-and-split.  You should adjust the levels (using volume control).  It’s beyond the scope of this document to describe exactly how to do this, but you may need some kind of a preamp between your record deck and the sound card.  You can use the level monitor, shown as a percentage, in Splitawav to get the amplification on line-in correct, when you press the record and split button.  You can use the LIVE UNSPIKE option to monitor levels too - the green bars show input levels - but if those bars go red, then your input level is too high.  Yellow is a warning that your levels are at 90% or above.  The two level monitors are different - the colour bars are more or less live reflections of the input levels, but after and clicks and pops are removed.  The value shown as a percentage (Peak level) logs the highest level found, so is easier to trap the highest level.

To use Record and Split

When you are sure your levels are right, type a filename in the box next to "base filename", or click "base filename" and find an existing file.  You don't need to put ".wav" on the end.  The base filename will be appended with __01 for the first file, __02 for the second, etc. This format makes it easy to select and sort filenames later.  While recording, you can hit the "stop recording" to finish off.

For best results I recommend starting the record player first, and then hit the record and split button, otherwise you mess up the normalise option settings, due to the loud click of the stylus landing.

Live Unspike

Now, one important thing - when you are using the "LIVE UNSPIKE" option (to hear your records with the click remover running) you must make sure your mixer settings are correct, or you will hear either a terrible echo effect, or ever increasing noise.  Your mixer settings are usually obtained by clicking the volume control item, and if that isn't showing, select it in the control panel under multimedia.  Now, some mixers have two sections, one is quite well hidden.  Note that on my internal sound card, which I only use for testing, the recording section chooses "selected items" while the playback section choose "muted" items. It's important to realise this!  Settings for the recording need to be set so the line-in only is selected.  If you have something like "stereo mix" "mono mix", CD, "wave" - untick these.  In the playback section of the mixer, this shows which bits are muted - ensure you mute line-in, or else you will hear an echo, delayed by a second.

The mixer setting for splitawav, outside of the "Live Unspike" option are simply that you choose your input source in the normal way, and you don't need to mute line-in - so you can hear what you are doing.

Of course if you want to use the karaoke function - or even remove clicks that have found their way onto CD, then choose CD as your input device.

 

Click and Pop Removal

Click and pop removal is done after the recording is complete - during this time, you will still hear the clicks and pops present on the record.  If you just want to hear your records without clicks, use the "LIVE UNCLICK" option instead (see earlier).

As an example of the performance of my click remover, here are the waveforms of a piece of music, before and after.  You can download and listen to the difference too.

Waveforms after and before.

A large view of one minute's of music (the example that can be downloaded) showing the whole waveform, with the automatically cleaned up signal above the same recording.  No special treatment was given to give these results.

And then a closeup of one fixed section - note how the "good" section of waveform is completely undisturbed, but the nasty click has been removed, and can't be heard.

 

Example File 1: recorded without click removal (files are about 500K)

Example File 2: recorded with click removal.  An almost magical difference. 

 

Karaoke Mode

 Karaoke Mode manipulates the live playback coming from whatever recording source is selected, and so can work on a CD.  It can be combined with scratch removal too, just by ticking the appropriate box.

 In this mode, the stereo signals are subtracted leaving just the difference. This results in a strange effect which tends to move the main foreground vocals to a more diminished background, leaving in the instrumental parts. The resulting effect is mono.  This option will simply produce silence on a mono signal!  The volume level is deliberately reduced during this mode, to avoid too much distortion, but note it is in the nature of this option that the resulting sound may not be high fidelity.

 Different music will give different results.  A good example of almost perfect vocal removal is on the Van Morrison track - "Brown Eyed Girl".  On classical music, the effects are sometimes not even noticeable.

  The slider below the Karaoke option can adjust the effect, so emphasis is given to the left or the right - this can help eliminate vocals when they are not dead-centre.

 The output result is not intended to be recorded - it is not possible to apply karaoke effect to a wav file.

 

Gaps

You can choose to remove the gap at the start of tracks - this should be used with caution. Once again it worked in most cases, but occasionally might trim too much.  It's good for certain types of music, no good where the gap detected is just a lull in the music, since this option actually removes time from the recording.  Finally, I added a normalize option so that the individual tracks don't go too quiet.  

The debug option can be ignored: it provides information on the audio, and can be used to judge why breaks aren't made. The file created, with one line per quarter second is stored in c:\templog.txt.  Also, a file called PointsN is created for each file "unspiked", which shows which samples were altered as sample ranges. These files can get rather large!  The Points file contains values in terms of sample numbers.  The sample numbers are incorrect after the first 50 Megs of data (Intention of use is for debugging)

Spike Settings

This is quite an important slider bar - if you are using the unspike option.  As you adjust this, a tooltip warns you of the meaning of the setting.  Generally, the more to the right you go, the more spikes are removed. Naturally there is a price to pay - it is possible that what is detected as a spike is part of the music. Although I've allowed a ginormous amount of noise to be removed by dragging this slider to the far right, this would only be recommended in an extreme case, for example an ancient record with speech on it.  The examples that you can download (MP3 files) are about as far to the right as you'd want to go - setting 17 - this recording is an ancient Petrouchka performance and is well scratched.  For a normal record with one or two scratches, I'd recommend a setting around 14, though you can go lower if you want.  If you go much lower than 10, you might find only the most extreme scratches are detected, but nevertheless, I've kept the option there.  Try 13, and work your way up.  The bar labelled "progress or clicks" shows any despiking that happens, but only on the "LIVE UNSPIKE" option.  You can only change this bar during the "Live Unspike" playbar, or when not recording.

Other things:

The calibrate option can be ticked (set any old filename first).  The program still does its stuff with calibrate on, but also sets the all-important noise limit value (its exact meaning is proprietary, but in general a higher value is a detected track break).  To calibrate, stick the stylus down 20 seconds or so before a track break, and start recording with the record and split button. Stop after 20 seconds or so into the new track - you can run this as many times as you like.  You can just leave it running, but then the level might drop a little too much.  I found 10800 was about the right figure, but anywhere between 9000 and 12000 will be sensible.  When calibrating, you will see two figures in the lower part of the dialog box.  The first value will change while the music is playing normally, and the second should only change when there's a suspicion of a track break - you should see this go to high values when the music fades out, and at the same time the calibrate value is set to a kind of average between the two values below. Make sure you complete calibration if you start it - otherwise there won't be a valid number in this noise limit box.

The base filename has a number appended to it for each track made, so if you start up with "jazzblues.wav", you will get jazzblues__1.wav etc.

To help set the input levels, I provide a peak level as a percentage.  I don't do fancy graphics for this, because I want to try to make this program as solid as possible, and not waste processor time.  Naturally, you should try to keep this as high as possible without clipping (it will say "CLIP" to warn you).  While setting up, you can push the Zero button to re-measure the levels. Note that scratches will cause clipping when you have the levels high, but this is generally OK, because those scratches will be eliminated with the unspike option.  Although you can zero the value, be aware that this will cause problems if you are normalising a track.  You should normally abandon any recording if you press the zero button unless you are sure that you'll get a correct peak value before the recording ends.

Choosing a gap length

This can be tricky.  I found two seconds worked in an awful lot of cases, and the times it didn't were because the music was patchy, or didn't actually have gaps when I listened. Some modern music has tracks that simply run into each other, some or all of the time.  For really dodgy cases, use the "manual splits" option (you have to push the button to cause a track break).  The track break value is in lumps of quarter of a second.  Anything less than half a second is pretty meaningless, since you would get track breaks all the time, but you could for instance use a value of 2.5.  For very quiet classical music, try longer gaps, such as 4 seconds, and turn off "trim leading silence" so that nothing is lost.

My email address is geoff@geffers.FISHu-net.com - spam protection please remove the "FISH" from the email address.

Speed

Splitawav is not fantastically fast when operating on wav files.  It should run fine on any modern computer above 800Mhz, with any modern sound card, and with Windows 98 and above. There are no guarantees however, so you should only buy this product if the shareware version meets your expectations.

Disclaimer: No warranty is given with this program.  It is designed and written in good faith, but I will not be held liable for any loss, or damage caused by its usage, in whatever form this might take.

Costs:

This program is shareware and has a limitation of usage built in of around 60-90 days. If you find it useful you are encouraged to use the PayPal system to donate money, and this keeps my overheads down.  In return you will get a version registered to yourself, without time limits.  Programs are individually compiled and will work on the intended machine only – they can be traced to the original registration. Minimum price £15 (UK) $25 (rest of world)  You can pay more than this if you like!  Use the button at the end of this page.

Quick FAQ

1. My record wasn't split at a track break.  I've found that on many records, especially pop/rock, the record producer has decided at certain times not to really put in a break at all - when the program fails, 9 times out of 10 when you listen closely, you will find that the tracks run into each other, sometimes with just a quiet fade out/in.  You could reduce the gap, try 1.5 if 2 doesn't break.  I wouldn't recommend anything below 1.00.

2. My record made too many track breaks.  If you are making a CD up, you can use "disc at once", and then it won't really matter if you have additional tracks.  Some music seems to be more prone to this than others, slow rambling classical music, for instance.  You can try upping the gap interval - perhaps even to five seconds.

3. Using the LIVE UNCLICK creates a distorted sound.   On the top two settings for click removal, you are guaranteed to get a bad result, but sometimes it's preferable to very loud clicks and pops and crackles.  You shouldn't get a noticeable distortion on any levels below the top 4 settings, and hardly any on the two immediately below the top 2.

4. How does the Unclick facility compare with other solutions? I've tried a few, no names, no pack drill, and they all came out worse, running against my test sample shown here, in some cases the clicks were only removed in the silent parts, and in others distortion resulted from removing the clicks.

5.  How does the split facility compare with other solutions?  I've tried a couple, and they behaved abysmally, on the simplest possible record.

6. Can I buy Splitawav without Paypal. Yes, please contact me for an address to send the cheque to. The program will be sent by email, and is individually compiled for each user.  I cannot accept credit cards, except through Paypal.

7. Can I copy Splitawav for my friends. I'd rather you didn't. It's a low price product - some declickers are selling for hundreds of dollars.  In any case, the program is individually compiled for customers, so if one turns up on a warez site, I can trace it back to its original owner - the registration is almost imperceptibly hidden.  The shareware version can of course be distributed.

8. Help, I hear an echo when I use the Live Unclick option! It is likely you have not muted Line In, on the OUTPUT settings on your mixer.  This means the line-in signal is being directly amplified, and merged with what SplitAWav is doing.

9. My sounds ended up too fast/too slow.  It sounds like you are not using stereo 44.1Khz samples, which is necessary with  this program.

10. I am not happy with SplitAWav That's why the program is shareware - if you don't like what it does, feel free to tell me, but also, don't buy or use it if you don't want to.

11. Why has the program locked up?  When you've done a record, and hit the "stop recording" button. The program will quite intensively do the normalize/unspike options. In fact that's another thing - when using Splitawav's record facility only use the computer lightly - it seems quite resiliant, but I wouldn't advise doing anything other than word processing, email, or browsing while it's in progress.  Certainly avoid anything that might lock the machine, such as a DOS program, or control-alt-delete.  When this last stage is going on, your hard drive will become very busy.  The process takes about 60 seconds for a side of a record on my machine, so just let it do its stuff!

12. Who do I contact? My email address is geoff@geffers.SPAAAMu-net.com  (remove the SPAAAM)

13. How to I register? You need the unique registration code, which you can copy and paste from the dialog box - it's down the bottom. Use the Paypal button below to make the payment. I will need your code to make you your own registered program.   If you change your computer, or install a new operating system, please email for a new code (the new computer would show a different license code in the bottom of the window).  The program reverts to shareware version when the registration code does not match the machine.  This is done to protect the program against copying.  Your registered program will not work on other machines - feel free to ask for codes for each machine you own, within reason!

If you are too poor to afford to buy this program - I'd rather you asked me for a free copy than looked for, or engineered a Warez version.

(remove the line break - this is done to prevent spam)

Links:

Music shopping and Music resources website: clearance items and the best deals on CD's.