posted October 04, 2006 11:11 AM               

quote:

---

Originally posted by KC Stoever:
BTW, was I hallucinating last night, listening to Keith Olbermann's Countdown (tv was in the other room)? Didn't he have the "small step for a man" story at number one on his story countdown? I can't find mention of it in the provisional transcript of the show.

---

But first, time now for COUNTDOWN's top three newsmakers of this day...

Number two, Neil Armstrong. Long have we been told that the first man on the moon said, "That's one small step for man, one giant leap for mankind"? But he was supposed to say, "That's one small step for a man, one giant leap for mankind." Now an Australian audio analyst has cleaned up the tape of what Armstrong actually said, says Armstrong got it right, said., "That's one small step for a man." But the word was lost in the garbled transmission from the moon.

posted October 04, 2006 04:58 PM               

• One small step backwards
  
• One 75-millisecond step before a "man"
  
• Armstrong's abbreviated article: the smoking gun?
  
• Armstrong's abbreviated article: notes from the expert

posted October 05, 2006 02:57 AM               

quote:

---

Originally posted by Rick:
I, for one, still don't believe Armstrong actually said "one small step for A man." The cadence of his sentence, to me, does not include any break for a missing word.

---

posted October 07, 2006 06:21 PM            

Upon hearing in the news about the Peter Shann Ford analysis, I did my own evaluation of the NASA "One small step for (a) man" sound file, and determined that Mr. Ford's analysis does not support his conclusion.

Mr. Ford used $45 GoldWave software and only looked visually at the section in question, after applying noise filtering.

I utilized $400 Adobe Audition 2.0 audio production software (developed by Syntrillium Software, which company Adobe bought), and $700 Celemony Melodyne Studio 3.0 software to do my investigation, so I had somewhat better tools at my disposal.

In the questioned part of the audio file where the "a" is sought, there is a bit of noise, specifically three short noise spikes, and that may be misleading Mr. Ford. Mr. Ford does use noise filtering, but such filtering does not completely eliminate the noise spikes, due to their prominence. You can still see them.

I am providing some much clearer screen shots and audio samples which should demonstrate to the careful listener that no "a" is evident in the recording.

Here is the original recording as obtained from the NASA website.

Here is a version of that same recording after applying Adobe Audition noise filtering algorithms.

Here is the recording, with no noise filtering, slowed down to half the original speed, using Melodyne.

Here is the noise filtered version slowed down to half the original speed, using Melodyne.

Here is a version which uses Adobe Audition to slow down the "for man" phrase to one tenth of its original speed.

Note that in the 1/10th speed version above, Mr. Ford's "0.035 second" section would be magnified by a factor of ten, making it about a third of a second long (0.35 seconds). This would satisfy Mr. Ford's anecdotal comment that the original was "10 times too fast for the 'a' to be audible." I think that the attentive listener will discern that the "m" is already closed during that time.

Supporting screen shots are consolidated into one PDF file.

The first picture shows the spectrogram of the original NASA recording, including background hum and noise. The three noise spikes mentioned previously can be seen.

The second picture shows the filtered version, with the three noise spikes still visible.

The third picture shows the audio wave file with the "for man" segment stretched to 10 times its length.

The fourth picture shows the filtered version as Melodyne depicts it on a "piano roll." The vertical axis represents notes on the piano ("C3" is the "C" one octave below middle "C") the blue "blobs" are the wave file, and the thin blue line represents the musical pitch of his voice. Note that in the segment in question, where the noise is, the pitch cannot be followed by the pitch detection algorithm of Melodyne.

posted October 12, 2006 09:17 PM            

For convenience, I have extracted the first page text:

An analysis and rebuttal to Peter Shann Ford's conclusion about what Neil Armstrong said
Garth D. Wiebe, October 12, 2006

Here is what Peter Shann Ford did.

He made the claim that the "a" was present in "One small step for (a) man" based on the following methodology:

He used a simple, low cost audio wave editor (GoldWave v5.14) to examine the audio waveform.

First he used the editor to select the background noise in the waveform where Neil Armstrong was not speaking.

Then he used the editor's capability to filter the whole waveform based on that selected noise profile.

Then he zoomed in on the area in question and noticed that there was still something there. He assumed that this was the missing "a" sound.

What he saw was only the noise that existed there in the first place, although somewhat reduced by the noise filtering.

He said that the "a" sound was too quick to be humanly perceived, as it was "0.035 seconds" long.

Here is what I did:

I likewise created a filtered version of the original recording, in the same way that he did.

• Original (from NASA website)
  
• Filtered

I used a more advanced audio editor (Adobe Audition 2.0) to create a two-dimensional, color-graded frequency plot, called a "spectrogram" of both the filtered and unfiltered versions. In it you can separately see the hum (straight continuous lines), crackly noise (vertical intrusions), and his voice (long wavy lines consisting of his fundamental pitch plus harmonic overtones above it. (See page 2.)

I used a second, very state of the art utility (Celemony Melodyne 3.0) to plot his fundamental vocal pitch on a "note by note" basis on what is called a "piano roll." One can flip back and forth between the piano roll and the spectrogram and correlate them. Melodyne is not able to discern a voiced vowel pitch ("a") during the spot in question, because it detects only noise there. (See page 3.)

I showed that the noise was still there in the filtered version, of sufficient magnitude to mislead someone looking at just the waveform into thinking that the desired object was there. (See page 4.)

On page 2, I show exactly where the "a" is still missing in the spectrogram. You can see this clearly in the spectrogram, whereas a simple waveform cannot reveal it. I circled that place where the "a" should appear as a fundamental tone of his voice, beneath the noise, if it was there.

I used the aforementioned advanced audio editor to slow down the section in question by a factor of ten, so that Mr. Ford's "0.035 second" section is now 0.35 seconds long. This alleviates Mr. Ford's contention that "0.035 seconds" is too quick to be humanly discerned.

Listening to the slowed down version, one can hear that the "a" still cannot be discerned, and that the sound progresses evenly from the "r" sound to the "m" sound with nothing in between.

• Slow filtered version
  
• Slow unfiltered version