Featured Post

Monday, June 3, 2013

Synching New Tracks To Old Recordings With Cakewalk Sonar AudioSnap

AudioSnap is a great thing, if it works. I tried it many times in the past and never got the results I wanted.

All I can say is that a light came on the other day. It reminded me once again that all this technology can get in the way of remembering that "simpler is better."

I had almost written AudioSnap off as a tool for those who have audio tracks with heavy percussion defining each beat.

I think I mentioned in another post that creating backtracks for my brother (a singer) got me interested in sequencing. Any recordings that couldn't be "vocal removed" gave me another project to help me learn how to create sequences.

With all the karaoke tracks out there, there are still songs my brother likes to sing, and there are no tracks available for them. Or, the karaoke is in an impossible key for certain vocal ranges. 

He sent an original recording of such a song.

Before I go any further, what you read below will be foreign to you if you haven't at least played around with AudioSnap. You can take any recording and practice doing what I'm suggesting here. I would suggest a recording that starts on beat one of measure one. Something with a relatively steady beat. And maybe something fairly short.

I dragged the original recording my brother sent into an audio track in Sonar. The tempo was about 142, but it ranged from about 140 to 145.

The mix of the recording was poor in many places, but I tried looking at the "Audio Transients" (using the drop down below the audio track name which defaults to "Clips").

Bad results.

So I took the hard road and tried EQ to bring out the bass drum and bass parts. That was a complete waste of time -- the kick and bass came and went in the mix. The transients didn't seem to line up with the beats that well.

I thought, "if only I had a click track." Then a light came on, "DUH! Why not record a MIDI track with a bass drum hitting each beat of each measure?"

So, I created a MIDI track and set it to play the TTS-1 bass drum.

Then I turned off the metronome ("P" for Preferences, Project>Metronome -- deselect "Recording").

Armed the bass drum track, clicked the record button and started hitting the bass drum. I missed the first beat, but I had the recording set to start on the first beat. I could add the first kick drum data in Piano Roll View (PRV).

Before I go further, it's important to tell you that if your original recording starts on any beat other than beat one, you should insert some audio at the first of the original recording to fill in at least to where the first beat would be if there was audio in the whole first measure. For some reason, AudioSnap loves to consider silence as a measure zero and you cannot force it to do otherwise -- at least I couldn't.

If your added audio makes the first measure too long and it confuses AudioSnap, notice where AudioSnap wants to put the first beat and shorten your added audio to that location. Then right click your audio clip and select "Bounce to Clip(s)" to get rid of the extra audio data. Now AudioSnap should give you a first measure of the correct length for what it considers the average tempo of the recording -- and it should start measure/beat counting at 1:01:000.

After recording your bass drum MIDI track, you should expand the MIDI track so you can right click in the FX bin. Select MIDI Plugins>Cakewalk FX>Velocity, and click the "Set to" fader and set it to 127.

Now you have a strong bass drum beat. You can minimize the Velocity plugin.

Next, freeze the TTS-1. Bring up the synth rack -- I use "B" for that, and click the snowflake on the TTS-1.

Your TTS-1 output track will now have an audio click track with clean and definite beats. Select that audio clip and use the drop down to view the "audio Transients."

Open AudioSnap -- I use "A" to do that.

To use the Audiosnap window you have to have your click track selected.

Now click on "Edit Clip Map."

Here's where I had a problem that you might face. AudioSnap, though it guessed the average tempo correctly, wanted to put the measure markers as if they were half the tempo. Also, even though the beats were strong, I had to adjust the threshold in AudioSnap down to zero in order to see all transients. But, since the beats were strong, each beat had a transient.

To correct the markers on incorrect transients,  I clicked on the marker of the first beats of each measure, starting with measure two, and brought it in line with the correct transient for beat one of that measure. I continued on toward the end of the recording.

That's when another problem came up. I ran out of measure markers several measures before I was through lining everything up. Pulling the last visible transient toward the left locked Sonar down tight.

I had to re-record my MIDI kick track, and I made sure to record several measures of bass drum hits after the recording ended. This was so AudioSnap wouldn't run out of transient markers for me to drag. After doing that, I froze the drum track again and started dragging beat one to each appropriate measure.

Make sure to save your file regularly as you do this -- just in case.

Next, in the AudioSnap popup, I selected "Set Project from Clip." I clicked the drop down and chose "measure." The reason was that I knew at least the beginning beat of each measure was dead accurate. This causes a tempo map to be inserted into the project, with tempo changes whenever they are needed to keep the project file in synch with the original recording.

From this point on, you can create MIDI tracks or record audio tracks with the metronome activated, and whatever you record will be in synch with the tempo of the original recording.

You can now archive the MIDI/audio bass drum track -- or even delete them. It's great to be able to record parts while listening to the original recording -- and everything synchs up.

Note that you can do this same thing by recording a MIDI track with a beat on just the first beat of each measure. That results in fewer transients to have to deal with if you have to drag the measure/beat markers to the correct location.

I'm sharing this for what it's worth to those reading it. If you're new to sequencing, creating karaoke versions of songs is a great way to learn. It hones your ability to hear specific parts and block out others. It also helps critical listening -- and much more.

If you have some AudioSnap techniques you'd like to share, drop me an email or text file. I'll post it with full credit to you.

Finally, I apologize if this post is confusing. Any comments/suggestions on making it clearer/better will be appreciated and acknowledged.

The song I most recently used this technique on was "You Be Dead" from Duke Nukem II. You can find the results of my synching HERE ON SoundCloud. I also used the crossfading technique from the previous post.