RECORDING the Hammond Organ

There are three different places where we may add the various effects of DSP. The first place is at the instrument itself. For instance, we take the mono signal right out of a B3, set up a DSP to do both Leslie simulation and add reverb, and then the stereo output of the DSP would become the input for two tracks on a multi-track recorder.

The second place to add these effects would be in the effects loop(s) of a mixer or multi-track recorder. The effects loop is simply a common output that we can access via the effect send controls on a mixer or recorder. Depending on the setting of the send control, a certain portion of the signal from any particular track will go to a single or mono output from the mixer or recorder to the mono input of a DSP. The DSP invariably returns a stereo signal. This stereo signal comes back into the mixer or recorder through the effect return jacks. (A classic electronic example of what goes around, comes around. Generally there will also be a control here as well to regulate the volume of the returned digital effect.

Some mixers have more than one effects loop, and many digital mixers and multi-trackers have both a standard, external effects loop such as I described above, and they also have an on-board DSP which works essentially the same way with the exception that it is a part of the recorder or mixer. This gives us the possibility of using two different effects to be applied either together or individually on different tracks. And of course we may first use individual, stand-alone DSPs for specific instruments and their dedicated recording tracks.

The final place where we can add an effect is on the mixdown when we have it in a computer and are doing our final processing, which we’ll look at soon. Effects added here will affect the entire mixdown. Typically the effects that we would do here would be: eliminating crackles and pops from noisy electronics, eliminating background 60 Hz power humming, equalization (which is a fancy word for adjusting bass and treble but in sometimes very specific ways) and also noise-gating, compression and normalization.

A noise gate is a device which does not allow any signal through unless the signal is above a certain level. The idea behind this is simply that the musical portions will be louder than background noise. Therefore, in the absence of a musical signal, the noise gate will not let the noise go through. When the musical signal is above the cut-off level of the noise gate, then the noise gate lets everything get through, however, the music is louder than the background noise and masks it so that you are not aware of it. This is very useful with many electronic organs, particularly Wurlitzer electrostatic instruments which are notorious in humid weather for letting signals from every tone generator bleed through, resulting in a constant background droning.

Compression changes the gain of a signal, taking very quiet portions and automatically ramping them up a little. Likewise, compression also automatically reduces really loud sections so that you hear quiet parts better and loud parts won’t blow you off your chair.

Normalization is the process of making all tracks on a CD, for example, approximately at the same average loudness or volume. For instance, once you put the CD into your player and sit back to enjoy the music, you don’t want to keep getting up and turning it up or down because different songs on the CD are either too quiet or too loud relative to each other.

Remember, at the recording end of the process, we have no idea what you will play the CDs on. We have no control once the final mixdown is applied to a CD. On the next two pages, we’ll examine the final steps of the process, seeing exactly what we do in the mixdown process and with the computer sound editor to get the final stereo recording ready for you to play on the CDs you buy or the music you download from the Internet.

Back in the days of analog recording and editing, some aspects of the final editing process were a lot more difficult than they are today. For instance, sometimes it would be necessary to cut up lengths of recording tape with particular portions of material and then rearrange them, or perhaps eliminate pauses and then splice the tape back together. Eliminating particular short-duration defects such as a snap or a pop could be very difficult.

Today, we can display accurate graphical representations of the audio tracks on a computer screen and then delete unwanted portions, add a little extra silence, add reverberation just to a final note or a brief excerpt and do many other things as well, and it is also possible to take sections of a final song that were recorded at different times and then string them together so that everything goes along smoothly and there is absolutely no interruption of the rhythm or “flow” of the music. What used to take hours and hours of very painstaking work can now be done in a few minutes on a computer. Also, there is no waste of material as there would be when recording tape would be cut up and sections discarded.

Here's a classic example. Let's say you are recording your playing of a traditional Hammond. You've come to the end of the first part of a song. You want to change a drawbar setting, turn off the vibrato for one manual, change the setting of the vibrato for the other manual from V3 to V1, turn on the percussion, and make a change to the pedal drawbar setting. How could you possibly do all of this and still maintain the beat and not interrupt things? The short answer is, you can’t. You have several choices. You could make just one change and perhaps make other changes later on. You could have a second person standing by to make some of the changes, or you could vamp the rhythm with the left hand and pedals while you make the changes with your right hand, however this adds extra measures and in general does not sound very professional.

However, if you are using modern recording methods, you simply stop recording after the first part of the song. Then you take your sweet time and make as many changes as you want. You play a little and see if you are happy with the changes, and maybe make a few minor tweaks. Then you listen again, decide you are happy with the changes, and then you resume recording.

A digital recorder can “stop on a dime” so to speak and resume again when you are ready. However, this is still a little tricky and you run the risk of a slight pause or hesitation as you move your hand from the recorder back to the Hammond. We should note, however, that on modern digital multi-trackers, there is a feature called auto-punch. Auto-punch lets you set a specific point on a recording, so that the machine will automatically switch from the playback mode to the record mode on a specific track or tracks. You can then listen to the playback after you've made all your changes and be immediately ready to resume playing at precisely the right time just as the recorder switches from playback to record. The multi-tracker when in auto-punch mode will let you listen to a portion of the previously recorded material so that you will know exactly when and where it is going to swap from playback to record, and thus you will know exactly where to resume playing.

This can still be a little tricky, however, and it's possible that you could still have a slight break in the continuity of the song if you are not ready to resume playing at exactly the right instant when the machine switches from one mode to the other.

In my opinion, a better method is as follows: if you store the first part of the song on a computer, make your changes, store the second part also on a computer; then you can look at graphical representations of both parts on the screen, and simply move the second part forward or back in time down to one millisecond until its beginning matches up exactly with the first part’s ending so that when you play it back, the first part ends correctly and the second part resumes exactly where it should; and all of the changes that you made are intact and there is no break at all in the rhythm.

This, by the way is not a new technique. This was done via physical cutting and splicing of recorded sections of tape back in the 1950s when tape recording was perfected. From our research, we learned that famous 1950s organist Ken Griffin used this method on many of his recordings to make mutliple changes and yet have everything right on the beat. The difference is that now, by moving virtual song parts on a computer, it's a lot easier and also more precise.

I am not aware of the above-mentioned auto-punch feature exisiting on earlier commercial tape decks. On pofessional machines with independent recording and playback tapeheads, there is a slight physical distance between the recording and playback heads, and the amount of time taken by tape to traverse this distance could introduce a slight glitch in the accurate timing of the music. I should preface this statement with a mention that although I have not seen such a feature on any tape recorder, however, I have not had any experience with an analog multi-track tape machine such as the one in figure four so I do not know with exact certainty if such a feature might have been included.

On the next pages, we'll look at a few specific examples, showing actual screenshots of sound editing in progress.