Tip:
Highlight text to annotate it
X
Andrey Subbotin Mastering engineer
Today we'll walk step-by-step through compilation of a CD master and learn which operations
are involved in the process of producing a master for the pressing plant.
We begin with arranging the tracks on a timeline in the order determined by the band or by
their producer. It is not something that the audio engineer would decide.
We simply put them one after another in that order.
Next we can space the tracks with pauses that are deliberately longer than required,
for example 5 seconds.
Just like that.
We'll be doing a short master of 4 tracks.
What's next? The next step is cleaning starts and ends of each track.
The tracks are already mastered in 24 bits. We need to choose the right type of display in Pyramix...
And remove the excessive leading and trailing noise.
Pyramix is very convenient for this because it automatically normalizes the waveform display.
Once we make a selection, we can see the tail in great detail: the image gets normalized.
There is an interesting moment here during the fade-out. Let's listen once again.
This track has been noticeably compressed at mastering and during the natural signal
decay the compressor does not let the sound go, it holds it a bit longer than needed.
What do we need to do about this? We need to fix the decay curve, make it more natural.
Let's listen...
This is a regular fade-out that does not sound quite right here.
We'll slightly adjust it here.
Now it fades out beautifully. It can even be made slightly shorter.
Now the curve is more natural and the work of the compressor is not audible.
Now we'll do the same with other tracks.
There is something strange here... let's listen like this.
Same story here: the cymbal fades out, but the compressor lifts it back.
I think that the default fade will work here...
Even a bit shorter...
Just like that.
Then the operations are repeated. We've got a bit of noise again here.
Although we can hear the compressor's release here on the bass note,
we'll keep it because it sounds good.
And the cymbal fades out.
Moving on to the next track...
As with any live band, there's something in the beginning.
Here they have already had some fade-out. It may not be perfect,
but it was already there.
It was the last track of our example album.
The next operation is called song spacing -- it consists in determining the pauses between tracks.
Clearly, we could just space them with some standard intervals and finish the work.
But if we treat the album as an integral musical form and not just the compilation of tracks recorded by the artist during the
statement period, then the spacing of songs starts to play.
How is it done? Let's try to move them around guided by the heart, the mood.
Here is the end of the first song;
here is the beginning the second one.
Let's move the second one away and listen to the ending.
Here I want the next one to start. Putting it there and listening again.
A bit too far.
Nudging it slightly.
There we go! We could do the same with other tracks, but to show you different PQ coding
techniques we'll create a standard 2-second pause between the second and third tracks.
And here we'll try to make a crossfade.
For example, we may want the next song to start right here, on this note of the bass guitar.
It is possible to simply overlap tracks without the crossfade.
It comes out slightly wrong...
Let's make it like this. Not very meaningful musically, but this will just be an example.
By editing the fade,
we'd like to make the bass fade out slightly faster and not mess with the next song.
That is better!
Here we've got our example album properly spaced.
The next operation is PQ coding, i.e. setting the markers.
Pyramix is trying to set the markers automatically.
We'll let it do the job and look what happens.
The first one...
the second one...
the third one...
the fourth one...
and the final one denoting the end of disk.
We can see that the markers are set precisely at the beginning of each track. Why is it a bad idea to do so?
When the CD player is seeking for a track, it mutes the audio output. When the marker is found,
the output is un-muted. The time of un-muting on most (or even all) CD players is not zero
and it can vary on different players. So,if you want your track to play from the very beginning,
the marker should be moved slightly to the left. By how much?
By how much?
Strangely, there is no standard. In practice, the most widely used range is between 7 and 12 CD frames.
As you know, there are 75 CD frames per second.
So, we just turn on the offset and it can now be seen on the display.
The first marker has been shifted to the left by the whole second.
There is no big sense in this 1-second pause before the first track. It has a long history.
There is a legend that during the times when first CD players did not have a remote control,
the listener, after pressing the play button, should have had a time to return to his armchair before the music begins.
But nowadays the offset here can match the other tracks.
We've set them all to 7 frames. Again, Pyramix has set them automatically.
Let's look where some manual editing is required.
In this place, where we've been choosing the optimal track spacing, there is a starting marker, but no ending marker.
This is normal because Pyramix follows the Red Book standard
and the ending marker is omitted when the pause between tracks is less than 1 seconds.
The next pause has both the starting (it goes with an offset) and the ending marker.
Now to this place...
Pyramix does not know where to put the marker
and has set it at the beginning of the last track. So, if we start it on the player...
it will play like this. But I think that maybe it can be set slightly later.
No, this way we have lost the first note of the song.
Even earlier... It looks like this location will be optimal.
Now let's check how it looks in the player emulation mode.
Let's check what happens between the first and the second tracks, how the display looks.
Here is the second track and its time count begins.
Here, between the second and third tracks it will look a bit differently.
Since we have both ending and starting markers in there,
there will be a backward time count between them which was previously absent.
And the last transition... let's see how the label changes there.
Like that.
This concludes our work on this disk. We'll create the disk image used for cutting the CD-DA or a set of DDP files.