The Mystery of the Mac Start Up Chime.

I first that to say this has been the most interesting part of this project.

My initial thoughts and inspiration for this project was to use the OS X Start Up Chime as a sampled instrument for the bass line. This idea came from an old article Tim Whitwell posted on his informative, no longer updated blog Music Thing. The article mentions a few tidbits on the iconic sound [which I will cover more below]. However, it was the article mentioning the Berlin music group Tranformer di Roboter’s cover of Michael Jackson’s “Stranger in Moscow” which was the inspiration to use the Start Up Chime as a bass line.

Jim Reekes created the iconic Start Up Chime. Jim was the developer who was a driving force behind a lot of the early Apple Sound Manager systems. One More Thing has a great interview with him. In doing research, you can find that Jim created the sound on a Korg Wavestation and [according to Wikipedia], the sound Jim used was a slightly modified version of the preset ‘Sandman’ patch.

At first I did want to do the exact same thing that Transformer di Roboter did. Grab sample of the Start Up Chime, map it to a sample instrument and call it a day. However, since the origins of that sound are actually quite well documented I thought I could do one better. From what I knew and learned about the making the Mac Start Up Chime I wondered if I could make a sound-a-like of the sound instead of using a straight sample. So I got to thinking. Since I personally own a Wavestation [the software version] that seemed feasible. This itself, leads to an interesting question involving sound-alikes. Yet, in trying to make the sound-a-like is where I think I made the most interesting discoveries regarding this project.

Before I go on, I need to direct your attention to a couple links. First, there is this YouTube video that has all the Apple Start Up Chimes from inception until 2012. Second, 99% Invisible recently did a great episode on sound trademarks, where they talk about the Start Up Chime. The biggest take away from the 99% Invisible episode is that sound trademarks are very hard to get.

Now, if you listened to the YouTube video from above of all the Mac Start Up Chime you will notice the tuning has dropped over the years. Jim Reekes, in the interviews states the original was a two handed C-chord [with a third on the top]. According to the Apple Start Up Chime Tradmark, the current chord is a G-Flat with Concert A tuned to 432.4.

So my initial plan was to:

Load up my Wavestation software synthesized

Tune the synth’s master tuning to 432.4

Input the correct MIDI notes for the G-Flat chord on a track

Hit Play

I would then have my own non-sampled sound-alike of the Start Up Chime.

However, the reality is much more complicated.

The first thing I encountered was when I played my WaveStation version of the patch is the current start up Chime is missing the distinctive ‘chiff’ attack of the Sandman Patch. If you watch the history of the start up Chime video above, you will notice the ‘chiff’ is present in the Quadra era, but disappears afterward. My guess is that Apple decided to remove the first half-second of the sound to save space in the boot memory. That solved one mystery. However, even though the texture was closer, it still seemed very thin and more importantly it was very out of tune with the information I had and pitch of the chord.

The voicing I played for the chord from the trademark never seemed to match the current Start Up Chime sound. After getting frustrated I decided to load the sound into Sonic Visualizer and see if there was something in the chord I was not hearing. The answer, yes there is something missing, or at least in the wrong order.

The picture above is melodic spectrogram of the Mac Start Up Chime. In layman’s terms, colored lines correspond to the strongest frequencies being sounded in a sound. In this case, they correspond to the sound being played. If you look at the voicing of just the first 4 notes [starting bottom-up], the most prominent notes are the root, a octave above the root, then then a fifth, then another forth or fifth [1, 8, 13, 17/18]. Strangely, it does not match the chord voicing listed on the Apple Trademark of the Start Up Chime. The trademark lists the chord as the root, a fifth above, a forth [giving the first octave above the root], and then the third [1, 5, 8, 11]. In fact in the spectrogram, it shows a clear absence of any note a fifth above the root [keys 68].

So at this point, I was a bit frustrated and baffled. After getting dinner with a friend, he suggested maybe I could write Jim Reekes and see if I could get any insight. I thought this was a goofy idea at first, but when I discovered he had a website, I decided to try anyway. To my pleasant surprise he responded a couple days later and verified pretty much everything I had come to suspect about the sound.

Probably, one of the most important pieces is that the Wavestation was only a part of the Start Up Chime. He mentioned it was actually a stack of several synths and patches. Although he did not remember the exact details, along with the Wavestation he remembered using an Oberhiem Matrix 6 analog brass patch to get a lot of the fatness in the sound. He also used various stereo effects, such as panning and phasing effects, as well as selecting notes and tuning the filters to enhance the sound along the overtone series, with the third at the top of the chord he played. So now had a better idea about the correct voicing of the chord.

He also stated the reason why the current chord is pitched lower then the original C-chord sound, is actually due to a bug by the engineers. Digital audio is supposed to be played back at a sample rate of 44,100 kHz. However, for various reasons the actually playback of the current tone is 44000 kHz, thus dropping the base pitch about a tri-tone to the current G-Flat.

Armed with all that knowledge I had enough information to generate a pretty fair sound-a-like. I ended up with was the following:

[OK, On second thought, that graphic is not nearly as impressive as I thought when I took the screen shot.]

The sound-a-like patch is composed of a Wavestation playing a modified Sandman patch that removes the initial ‘chiff’ element. A layer of Logic’s ES2 – playing a slightly modified Brass path, then two instances of Logic’s Retro Synth playing a modified version of the ‘Cheerful Melody’ path, and ‘Slow Swell Brass’. I then added an assortment of delays, other effects, and a Chord Trigger so that every time a note played, the correct voicing would play back.

In the end, I think I got something that is functionally pretty close to the original synth tone. I am sure if I have more time, I could get the tuning dead on, but with finals fast approaching one can only do so much.


Big Thanks to Jim Reekes for helping out too.


Edit:  I have added a couple of snippets.  One using the original Mac Chime sound…

and one using my recreation…


The Intel Bong

The Intel Bong.

Sorry about the delay, end of the semester and projects are adding up.

Ok originally, I wanted to write about the Mac Start Up sound first, but that has turned into an interesting journey unto itself. So I am going to hold off telling about that and start with the Intel Bong sound. I decided for the melody the Intel Bong sound would be a good fit. First, I needed to acquire a good copy. Strangely, despite poking a friend that works at Intel, I could not get a raw sample of the sound. So I ended up ripping the audio from a YouTube video. I even found a little history on the noise here.

What I found interesting about the sound and building a Sample Instrument is that the sound has been roughly the same for the past twenty years. The sound has the initial bong that is an Octave Chord and made up of sounds which are various tuned percussions. The sound then plays a rest of the short melody on a Xylophone or similar instrument with heavy reverb. Over the year some synth textures have been added to the background but it has roughly seemed to stay the same.

The largest problem with building the sample instrument was there were just too many synth textures going on to create a good loop. However, since I decided to use this for the melodic line this was not a problem. Because the melodies typical classical counterpoint, there were usually continuously busy not to need sustained notes.

In the end there was really nothing special about creating this Sample Instrument. This is most typical of what the inventors of sampling synthesis envisioned. You take a recording of a sound, and then map it to a key range and play your new sound. Unlike the use of Sample in terms of looping that arose with Hip-Hop.

In producing the melody line, I did a very informal survey just to see if anyone recognized the sound. What I found interesting, is even if people thought the sound might have seemed familiar, most people did not recognize the sound until I played the short Intel Jingle associated with the Intel Bong.

AS for the Alto and the Tenor lines, I settled on using the Vocoder version of the Intel Bong for the Alto, and a separate audio sample I found of a choir singing the bong for the Tenor.   I find the Alto the more interesting, because although I am not using the sound directly, I am using characteristics of the sound to modulate a vocoder’s synth.  So as this point, how would one claim that it is still the Intel Bong sound?  For the non-musically inclined, a vocoder is the stencil version of a sound.   In other words, it was like I made a stencil of the Mono Lisa, and then usedthat stencil and spray painted through it.


Edit:  To demonstrate that Intel does think there is importance in having a piece of signature audio, this is a short video I recorded of unboxing an Intel NUC computer, at the lab where I work.


Here is what I found interesting.  Intel build into the packaging of the NUC and optical sensor that triggers a sound chip.  Every time you open the box, you hear the famous Intel Jingle.  At best when a person unboxes a new NUC, they might hear it once or twice.  Intel still thought this most is significant enough to add the jingle to make it memorable.


Thoughts on Pachabel’s Canon

Thoughts on I-V-iv-V

One reason I wanted to use Pachelbel’s Canon, is because it has a very well known chord progression. The chord progression for Pachelbel’s Canon is I-V-vi-iii-IV-I-IV-V. As noted by comedian Rob Paravonian, it could be considered the first pop song.

There is some truth to that, because a shortening Pachelbel’s Canon chord progression one can easily derive the [in]famous I-V-vi-IV chord progression, which has been used in a lot of songs. Most famously parodied by Axis of Awesome in their song “Four Chords”.

However, one needs to know, that copying a chord progression does not mean a song has been copied. Music, like many arts forms, contains make aspects to define a work of art. In analyzing a typical piece of music one has to consider melody, harmony [which includes the chord progression], and rhythm.

I think a better question to ask is what is about the I-V-vi-IV progression that causes it to reappear in so many pieces of music. I think the chord progression encapsulates a journey, musically. I also think one of its strength lies in the fact that the first chord is the tonic, which works by giving the progression a center. In other words it is sort of the short musical version of the Monomyth.

The MIDI File

So the first thing I needed to do for this project was to get the music for Pachelbel’s Canon. Now, the initial thought was I could get a copy of the sheet music and enter the music myself. However, in the modern age we have the great invention of MIDI so I wondered if there was a MIDI file available. After doing a web search I discovered a few websites with MIDI files. Unfortunately, the first site [] did not have any MIDI files that matched the original arrangement. I did get lucky with the second site [] which did have a couple versions in the proper arrangement. In order to compare, I found a version of the sheet music here.

Now for sanity’s sake, I cut the Gigue section at the end of the piece. Mainly, because most of the time the audience never hears that part of the piece, and when people do they are confused by it. So what I am going to use is basically the generic 4-part arrangement that we as an audience hear at every wedding. The major difference is we won’t be hearing it 5 times, while we wait for the bride to arrive.