For these many long years The Git has been using Goldwave to both record and edit sound files on his PC. Goldwave is very easy to use compared to other software of its ilk that The Git has tried out, is inexpensive to purchase and has never, ever failed to perform flawlessly over the decade and a half or more he has used it. Goldwave’s creator promised to never charge for upgrades and also allows the end user to have it installed on more than one machine providing only one copy is in active use at a time.
The only fly in the ointment occurred when using Goldwave running on Windows 7. The Git records vinyl records to the media PC in the main living area, then transfers the raw sound files to his main working PC for editing. There’s a “brief” description of the process at the end of this post. Playing the raw file in Goldwave produced a bizarre parody of the original sound; it was at a much higher pitch and sounded as though it had been heavily modified on a Fairlight CMI! Playing the same file using Foobar 2000 produced the sounds expected. Fortunately, Goldwave’s excellent FAQ led to the offending setting in Windows 7 that was the cause of the problem and all was well with the world — well, almost.
When using Windows XP, you can constantly display the recording level controls and adjust the levels in real time. This is not possible in Windows 7. The recording level setting is nested several dialog boxes deep: Control Panel, Hardware and Sound, Manage audio devices, Select the Recording tab tab, Select the Line-in device, Click the Properties button, Select the Line-in tab and you can finally adjust the recording level. It seems that Microsoft has determined to stymie its users’ use of Windows for sound recording!
Fortunately, there’s a neat shareware tool called Power Mixer that you can run either as an application, or a service. It provides an uncluttered window with your playback and recording level controls all neatly displayed for immediate use. It’s free for a 14 day trial and $US18 to purchase, though look at their discount offers before buying. As well as solving The Git’s problem, it allows you to save your settings to be invoked at the press of a hot key. Very useful when you have multiple sound sources that have widely differing sound levels. It’s arguable that Microsoft’s implementation of the mixer in Windows XP was more than good enough for general use and should have been continued in Windows Vista and 7. The Gittish POV is that paying $US18 to save endless farting about with deliberately crippled controls is a bargain, but then even in retirement, The Git’s time is worth significantly more than $US18/hr.
Digitising Vinyl Records
(originally posted on the old blog on 17 April 2007)
The Git has a rather fine collection of vinyl records, mainly from the 1970s, and he has been converting them to digital format. Why would he bother doing that? Many of these recordings have not been released on CD, or weren’t when The Git transcribed them. Quite often, when one of these old recordings is released on CD, it has been remastered and all too often remastered horribly. Some, like Electronique Guerilla’s first album, are quite expensive to purchase on CD. Here is how the transcription process works from a very Gittish point of view.
First, you need a turntable. The Git’s old Jim Sugden turntable died some time ago, so it was replaced by the serviceable and far from expensive modern equivalent from Pro-Ject. It came with a carbon fibre tone arm and Ortofon OM-5 cartridge, though The Git had the supplier replace the latter with his Audio Technica moving coil cartridge. This may have been a mistake. When it came time to replace the worn stylus, the cartridge needed to be removed and sent away for the stylus to be replaced. Since there was a backlog of disks to be transcribed, The Git researched which reasonable quality cartridge to purchase as an inexpensive replacement. It turned out that the Ortofon was much more than good enough and remarkably inexpensive.
Next you need to get the signal from the turntable into your computer. Vinyl records were made with a limited bass signal and a high level of higher frequencies. Restoring the signal to a flat frequency response (RIAA compensation) can be performed by your recording software, but it’s not generally recommended. The Git purchased a Pro-Ject pre-amplifier that can accept either moving magnet, or moving coil input. There are cheaper and higher quality versions and The Git chose the former. The results are more than good enough for his old ears. Unlike the ears of some of his friends, they have not been severely damaged by Neil Young concerts so the chances are good you don’t need to shell out the extra dollars. It’s your call.
Few people appreciate that moving a stylus around in a wiggly vinyl groove generates some remarkably high forces. It softens the vinyl sufficiently for each play to create permanent distortion in the form of increased high frequency content even at very low tracking forces. Any particle of dust caught between the record and the stylus creates major permanent damage at that point and results in a click, or pop. So, cleaning the record before transcription is very important. The Git’s budget doesn’t stretch to a vacuum cleaning system. First, he rinses the record thoroughly before very gently scrubbing the record with a very soft old toothbrush dipped in dilute non-ionic detergent. The idea here isn’t to reach into the grooves with the bristles, but to have them flicking about on the high points of the vinyl, hopefully generating turbulent suds to swish any gritty particles out of the grooves. Since the results are extremely gratifying sonically, The Git assumes that what he is doing is much better than doing nothing, or merely brushing the record with a carbon fibre brush. Some pundits will be horrified, but The Git expects that he won’t be playing these records again any time soon. Finally, the record is thoroughly rinsed in clean water. You may need to use distilled, or deionised water. The Git’s water supply leaves no visible residue.
Very stubborn, waxy residues on records require the addition of a small amount of methylated spirit to the detergent solution. It’s likely that this will also remove some of the vinyl plasticiser that keeps the vinyl flexible, so it’s probably a good idea not to use this unless strictly necessary.
The Git has a couple of document racks to hold the records while any excess water evaporates. The bulk of the water is removed by dabbing with a clean cotton handkerchief that has seen better days.
The Git was hoping to use his Mac to do some recording, but the software accompanying the iMic USB sound adaptor, Final Vinyl, produced weird results. When opening the resultant WAV file on his PC for editing, he discovered that the sound was mono, rather than stereo and half-speed. As well, the recording level controls were far too small and fiddly to use. The Git has been using Goldwave for sound recording on his PC for several years now. The supplier has been true to his word and never charged for upgrades. The Git has had no problem creating and modifying sound files using Goldwave, though he has on occasion evaluated other software.
If you are old enough to have recorded copies of vinyl records to audio cassette, or reel-to-reel tape, you will remember having to keep the recording level high enough to keep tape noise to a minimum, yet not to saturate the tape. An occasional excursion of signal level into the non-linear area of tape response might pass unnoticed, but overloading a digital recording creates awful sounds that can be ignored by no-one. There is no significant noise floor you are trying to keep the recorded signal above in a digital recording, so there’s no pressing need to keep the signal peaks as high as possible. A recording that’s a bit too “quiet” can have its volume increased after it’s recorded.
The recording leaves you with a 30-50 minute long sound file that needs some treatment before it’s ready for commitment to CD, or other format. First, the beginning of the file needs to be snipped where the stylus was placed in the run-in groove. A cue is then placed at the beginning of the file with the name of the first track. The name can be anything you like, but if you are going to be storing the files to be read by a computer, the song title is a good idea. If the results are to be stored as an audio CD, it can be any arbitrary string as the software can automatically number the tracks in sequence. For example, The Git labelled the cue point of the first track of John Coltrane’s A Love Supreme ”Acknowledgement” (without the quotes). Each section is demarked by labelled cue points, the middle portion when the record was turned over snipped as well as the end of the sound file. When all this is done, the software is instructed to split the file at the cue points into separately labelled files, one for each track. Using the previous example, the first track of John Coltrane’s A Love Supreme became “01 Acknowledgement.wav”.
Having created the separate wave files, they could then be used by Nero Burning ROM (for example) to directly create an audio CD. Chances are though that the sound needs some titivating first. It’s worth bearing in mind that you have by now invested more time than the time it took to play the record. You are a little over a third to halfway to the end of the process. Before manipulating the sound files, you should consider storing them in their current form. If you later decide to re-titivate the files, you don’t want to go all the way back to the beginning of the process. Windows WAV files contain a considerable amount of redundant information, so they can be compressed to save on storage space. Instead of PKZIP, or your other favourite file compressor, you can use either Monkey’s Audio, or FLAC so that you can play the resultant sound file with suitable plug-ins. The Git will have more to say about audio compression later in this piece.
The main distortions in the recording that we can do something about are:
- pops and clicks
- tape hiss
Pops and clicks are caused by major groove damage. Fortunately, the waveform of pops and clicks are distinctive and the software is capable of analysing the recording and either removing them completely, or diminishing them to a level where they are no longer a distraction. You decide the responsiveness of the software to the pops and clicks, either accepting the default, or changing the sensitivity. The more vigorous your declicking and depopping, the more you affect the overall music quality. The Git usually goes with the default setting in Goldwave as it is not over-zealous in its effect. Any major clicks, he removes manually before the automatic process.
To remove a click manually requires zooming in on the offending portion of the signal. A segment of the signal immediately adjacent to the click is copied to the clipboard, then pasted over the click. Obtaining a satisfactory result requires a little practice and occasionally, a final tweak of the signal. The signal trace can be moved by pointing at it with the mouse cursor and holding down the left mouse button and dragging. Practice makes perfect here.
The Git only decrackles an occasional recording, such as Little Stevie Wonder’s The 12 Year Old Genius. Yes, The Git is that old Decrackling has even more of an effect on the wanted music than declicking/depopping. Removing tape hiss more damaging still to the underlying sound. However, it is up to you how much cleaning up of your recordings is needed.
The Finished Product
Once you have your finished files, its time to play them. Of course you can play them as-is, but WAV files are somewhat bulkier than they need be. When playing them on the computer, it makes more sense to compress them so that they consume less disk space. There are two sorts of compression: lossy and lossless. There would be few who have not heard about lossy compression CODECs for music. While there are several sorts, the most common is MP3. Unfortunately, this entails removing some of the “unnecessary” musical information in order to make the file size smaller. This is fine on a low-fidelity playback system, but it degrades the music noticeably on a hi-fi even if you lack Golden Ears.
Lossless compression, on the other hand loses no musical information. A player with the appropriate CODEC will play the compressed file with no loss in fidelity. There are two such CODECs that The Git knows of: Monkey’s Audio and FLAC. Monkey’s Audio is Windows only while FLAC is supported on Windows, Mac and Linux. Here’s a comparison of file sizes using John and Beverley Martyn’s song, Primrose Hill:
|Original WAV file||31,089 KB|
|FLAC (level 7)||16,930 KB|
|Monkey’s Audio (extra high)||16,581 KB|
|MP3 (320 Kbps)||7,053 KB|
Monkey’s Audio and FLAC have varying levels of compression, with the higher rates taking longer to compress files than the lower rates. In each instance for the comparison, The Git chose the second highest as a compromise between time and space-saving. The highest level takes a fair while longer, especially when compressing multiple files, for a diminished rate of space saving. The MP3 is the highest quality lossy compression available and shows dramatically greater space saving, albeit at a noticeable loss of fidelity. dBPoweramp Music Converter is the “Swiss Army Knife” of file converters for the audio enthusiast allowing conversion between a large number of different file formats.
The Git is storing his sound files on a USB external hard disk connected to a Core-Solo Toshiba notebook purchased specifically for playing music. The internal sound card leaves somewhat to be desired in quality of output. The iMic external USB sound adapter, originally purchased for recording on the Mac Mini, also has a very low output requiring the power amplifier gain control to be advanced near to its maximum. A Soundblaster external USB sound adapter has been on order for some weeks now, but has failed to make an appearance yet.
For playing the sound files, The Git is using a minimalist application called Foobar. Foobar has replay gain support, but more than 6dB of gain causes distortion in the iMic. Hopefully, the output of the Soundblaster will be higher. Foobar beats any other sound player The Git has used into a cocked hat for simplicity and power. There should be more applications written so well. Using random track play of the whole music collection is a bit like having a radio station devoted to your musical tastes. Using random album play brings back memories of another life when The Git was a DJ who often played whole album sides.
Burning CDs from your audio files requires you to make a change in the burn settings. Nero and most other CD burning software automatically places a few seconds of silence between tracks. You will either need to turn that off, or eliminate your recorded silence from the audio files. Live recordings just run on produce unexpected silences between the songs if you don’t remember to change the burn setting.
The external USB Soundblaster worked well as expected, though only briefly. Not so very long after this was originally written, The Git built a dedicated media machine with lots of hard disk space and a digital TV card driving a 32 inch LCD TV. While the software that came with the TV card flatly refused to work (Pinnacle TVCentre Pro), Windows Media Center mostly works like a charm. Very occasionally, it decides that there’s no signal from the TV aerial, though it mysteriously knows what channel it’s tuned to and the program that is playing. Sometimes, restarting Media Center fixes the problem, sometimes it takes a restart of the PC. Recently, both these techniques failed to work and so The Git reinstalled Windows 7 and relevant software and the problem has yet to reappear. It would appear that bit-rot is still a feature of Windows!