|
|
|
|
 |
|
|
|
Detailed Steps Using Sound Forge XP Studio 5.0
Sound Forge XP Studio 5.0 (SF-XP) is the low end
version of the professional Sound Forge product by Sonic Foundry. The
XP version is priced for the home user and provides most of the features
that a home user would need in audio editing. As of this writing, Sonic
Foundry has released version 6.
- Get your compute system(s) set up. I digitize
my LPs on an old 200 MHz Pentium Pro system running an older Sound Forge
version 4.5. After the LP is digitized, I copy the files to my editing
system, a dual Celeron system running Windows 2000. I use seperate systems
for two reasons - to allow me to digitize one LP while editing another
and to make use of a higher quality sound card in the older system.
If you have more than one system, digitize on the system with the best
sound card. If you're not sure about the quality of your sound card,
search the web for product reviews. One excellent independent site is
PCAVTech.
- Hookup your turntable to your computer.
If you're not familiar with your stereo equipment, refer to its user's
manual. There are two important connections that make all this work:
- Connect the turntable to a phono input on
a preamp/receiver;
- Connect a line-level output of the preamp/receiver
to the computer sound card's line-in. You may need to get an RCA-to-1/8-inch
stereo plug cable. These are available at your local Radio
Shack
The general concept is simple - the turntable's
left/right output signals must be fed to a phono preamp. This can
come in several flavors - a dedicated phono pre-amplifier (usually
a high end product that takes a phono input and outputs a line-level
output), a general purpose pre-amplifier, an integrated amplifier,
or a receiver (stereo, surround, home theater). Do not
connect your turntable to a line-level input such as those marked
"aux" or those used by tape decks and CD players. The cartridge
on a turntable outputs a signal significantly lower in voltage than
line-level and thus requires more amplification. The phono input provides
this extra amplification as well as providing RIAA equalization -
a form of decompression that compensates for compression of the signal
originally recorded on the LP. All LPs use the RIAA curve and so must
be properly compensated on playback.
The line-level output fed to the computer can be a dedicated output
from the preamp or the tape-deck output. Any output that contains
the audio from the turntable and provides a line level voltage signal
will work.
- Prepare your computer. On a Windows system,
make sure you enable the Line-In connection and disable all other audio
inputs to avoid potential noise. To do this, open the Volume Control
applet in the SystemTray. Enable "Line In" only; mute all
other forms of input including the mic. Keep this applet open so that
you can adjust the input level control later.
- Prepare your LP for digitizing. Make sure
your turntable is setup properly. If you haven't touched it in a while
or if it's been packed away in the attic, give it a thorough inspection.
Make sure the cartridge tracking force is correct as is the anti-skating
force. For each LP you digitize, clean it well and clean your stylus.
- Adjust the input level. You may want
to do this for each LP (since the recording level can vary from LP to
LP) or find a good setting for all your LPs and live with it. In the
ideal case, you'll set this value once for all 33-1/3 rpm LPs and a
second time if you digitize 45's. Of course you may not be this fortunate
or you may wish to be extra diligent and optimize each digitizing session.
It's up to you. Either way, adjusting the input level is an iterative
process whereby you digitize some amount of audio, review the results
and repeat until you're satisfied.
Let's review what our goal is. We want to digitize the incoming analog
signal with as many bits as possible. In the ideal case, the maximum
signal (peak) for the LP (all sides) will be just under the maximum
value that can be represented by the data format. For 16 bit audio data,
valid values range from -32768 to +32767. Any value above or below this
range results in clipping and cannot be represented. If we digitize
at too low a level, we don't use all 16 bits and we end up with less
available dynamic range.
While the idea of fine tuning the input level may seem daunting, in
practice, we can usually find a setting that is suitable for most LPs
and we just keep it there. Even though different LPs have differing
recorded signal levels, most LPs do not have the wide dynamic range
that can be recorded with 16 bits of accuracy. The noise floor of even
a clean LP just isn't that low, and most of us with old collections
have LPs that are somewhat worn, even if we took great care of them
over the years. What this means is that even if you don't optimize the
input level setting, you have plenty of bits left over to handle the
music on the LP.
- Digitize each LP, one continuous side
at a time. Beware of skips - we can catch them in different places
along the process. A bad skip that causes an infinite repeat can be
caught when you digitize - your LP will never end and if you're not
paying attention your session will end when your recording timer expires
or when you just run out of disk space. You can also catch skips later
on when you edit the tracks or when you play them back during verification.
Naturally, the ideal place to catch skips is during the digitizing process.
As mentioned earlier, I use SF-XP. There are many other software packages
that are also suitable. Software cost varies. If you have a lot of LPs
to convert, you may want to consider a commercial package that offers
a time-tested tool. If you only plan on processing a few LPs, freeware
or shareware software may work just fine. I use SF-XP because it provides
me with an intuitive interface for recording and editing. The main feature
it lacks is direct sample editing, which I would have used if it was
provided, but was able to live without.
Here are the digitizing steps I used. Remember this on a Windows platform.
Your mileage may vary !
- Start SF-XP.
- Set the directory preference. This is where
digitized data goes before it's saved. I set this to the same disk
where the final audio files will be saved.
- Open the Record applet under the "Special"
menu.
- To set the input level mentioned earlier,
start playing the LP and adjust the slider in the Volume Control
applet to avoid clipping as indicated in the Record tool's meters.
As mentioned earlier, this may have to be repeated many times. If
you know the LP well, and you know that one passage is extremely
loud, play that part and adjust the input level using that as your
reference.
Note that you may be able to adjust the input level using the preamp's
volume control. Sometimes, it is advantageous to keep the sound
card's input near its maximum input level and use the preamp's volume
control to attenuate the level. You would do this if you felt that
the sound card has a poor hardware attenuator.
- Click "Prepare", lower the stylus
onto the LP and press the space bar to start digitizing. You can
also start recording by just clicking the red recording button but
I find it easier to use the keyboard.
Note that you can start recording before lowering the stylus but
if this yields a "pop" and causes the clipping indicators
to light up, be sure to clear the clip indicators so that subsequent
real clips can be seen.
You can preset the recording time if you want to leave it unattended.
Just set it to a time longer than the album side's total length
and come back later. I don't bother doing this; instead I use a
cooking timer to remind me to check my recording sessions!
- When you're done recording, stop the session
by hitting the space bar. The resulting waveform will appear in
a window. Make sure the clip indicators are not lit. If they are,
lower the input level and repeat the digitizing session. If you
think the maximum signal was too low (as indicated above the peak
meters), raise the input level and repeat the digitizing session.
- Exit the Record tool.
- Save the resulting waveform data to disk.
I name the files "side1.wav", "side2.wav", etc.
- Seperate the individual tracks on each
LP side digitized. Supposedly, some software packages will automatically
do this for you by detecting blank spaces between tracks. I never tried
any of these since I used SF-XP exclusively and it doesn't have this
feature. I must say that I'm skeptical of silence detection for several
reasons. First, my LPs have varying noise floors. Some are in pretty
good shape and relatively quiet. Others are almost 30 years old and
nothing on them is silent! Second, some LPs have passages that are as
quiet as the spacing betwen tracks. If the software is able to detect
true spaces, it would also pick up these false ones. Third, the noise
floor between tracks changes within an LP. Because the LP rotates at
a constant speed, the stylus has a faster linear speed at the outside
edge of the LP than it does at the inner edge. The faster speeds lead
to more noise. So any silence seperating tracks will be quieter on the
inner tracks. This makes detection more difficult. Lastly, many of my
LPs have no silence at all. Live albums are a good example of a continuous
recording. Even studio albums often flow from one song to another. Maybe
I'm just a hands-on person, but I prefer to take full control over the
task of identifying and seperating the tracks. Those of you with time
to spare may want to experiment with whatever automates software is
out there; maybe they do work after all !
Using whatever method you like, take each digitized LP side and cut
it into seperate tracks. You may want to post-process the tracks to
perform noise reduction or pop detection. How you cut the tracks is
up to you. Save each seperated track to its own file on disk Be sure
to trim the lead-in to the first track and the lead-out after the last
track on the LP side. The spacing between the tracks can be kept as-is
to reflect the original LP or you can modify it to your taste. For live
albums or albums where songs blend from track to track, you should leave
it as-is. For albums where each song begins and ends definitively, you
can edit the blanks in between the tracks to shorten or lengthen them,
or remove pops and ticks. Keep in mind that you still want some spacing
betwen the tracks when you play them back just so that they sound like
seperate tracks. For simplicity, name your tracks by number starting
with "1". This makes it easier to assemble them later when
it's time to burn the CD-R. Most LPs will have the length of each track
written on the LP's label and/or jacket - use this information to help
identify individual tracks. Caveat - that the song titles, song ordering
and song lengths on LPs are often wrong ! This happens more often than
you might think. I've seen A/B sides reversed, track lengths off by
one minute, missing tracks on the label, inconsistencies between tracks
labeled on the LP and on the jacket, etc.
Here a detailed look at what I do under SF-XP:
- Open the source file. This is the previously
digitized LP side.
In the example shown here, you're looking at over 20 minutes of
an LP side that contains four tracks - the track breaks should be
obvious. The top half is the left channel and the bottom half is
the right channel. Fully zoomed out, there are five obvious "pops"
(four on the left channel, one on the right).
- Disable Undo. Undo is a nice feature but
the application has to save steps and data as you edit and this
takes up too much time for me. If I make a mistake, its much easier
for me to re-read the source file and start over.
- I start from the back of the LP side - in
other words if an LP side has 4 tracks, I start with track 4 and
work my way back to track 1. This method leads to less data copying
under SF-XP - again it's a speed issue. I delete the lead-out samples
after the last track at the end of the side. If the LP was very
noisy to begin with, I'll mark a region at the end and do a fade-out.
This just gives a smooth fade before the CD-R ends, for a more pleasant
sound.
In the example shown here, the highlighted region (blue waveform
on black background) is marked for a delete operation. In this highlighted
region, you can see the "pop" at the very end of the waveform
caused by the stylus lifting off the LP during the digitizing process.
- Working backwards from the end, and using
the track times from the album jacket or LP itself, I locate the
begining of the last track. I do a mark-in and mark-out to define
a region, the do a delete (control-x). I create a new window ([menu]
File/New) of 44.1 KHz, 16 bit stereo data and paste the buffer I
just deleted (control-v). I immediately save this new window with
a name - usually the track number (such as "4.wav" for
the 4th track of the LP).
Shown here is the break between tracks 3 and 4. Notice that this
silence is not so quiet !
Here, we see the entire 4th track marked for cutting and later pasted
to a new window.
Finally, here we see the new window (lower) with track 4 pasted
in it. The original window (upper) still contains tracks 1, 2, and
3.
- Then it's back to what's left of the LP side
and I repeat the process - mark-in/mark-out/cut the next track,
open new window, paste and save as "3.wav"; and so on.
- When I get to the first track (the last one
to be processed for this side of the LP), I clean up the lead-in
to the first track just as I did for the lead-out. This means I
cut any junk before the first track, and may add a fade-in before
the first track begins if the LP is noisy.
- Burn CD-R. Numbering the tracks in numerical
order makes burning that much easier. Just drag the tracks in order
using your burning software. I set the burning software not to
insert a pause between tracks. Check your burning software to find out
how to do this. For example, if you're using Nero, set the "pause"
property of all tracks except track 1 to 0 seconds. By default
the pause value is set to 2. For albums that blend in (such as live
albums) you must do this. For albums that have seperate tracks,
having a pause (usually 2 seconds) won't hurt.
- Make CD-R and Jewel case labels. There
are 3 main parts to this process (a) the front jewel case cover (b)
the back jewel case cover (c) the CD-R label. Depending on how much
of a perfectionist you are, you can do this in a number of ways. If
you only want the cover artwork as a visual reference, a high resolution
image won't be necessary. The simple way to get a low-resolution cover
is to look up the album at an online site like Amazon
and save the images they provide as part of their web page. Then import
the image into an image editing program and resize it to fit a jewel
case cover. Print it out and you're done.
Another useful feature provided by online shopping site is the track
listings. I just copy and paste the track listings directly off the
web page and print them out as a text file. This becomes my back cover.
Naturally if you have an LP that can't be found online, you may have
to scan the LP jacket. Most scanner won't be able to scan a 12-inch
wide jacket so you'll have to scan in two passes and glue the two pieces
together.
Note that many CD versions of older LPs have "bonus" tracks
not found on the original LP. If you copy and paste the track listings
from an online CD retailer, be sure to double check the track listing
with that printed on your LP.
For the CD-R itself. I just use a Sharpie pen and write whatever I want
on the CD-R. If you're into fancy labels, you can print your own round
label and stick it on.
|
|
|
|
