Holly Neill and John N. Maclean compiled and sent to us the detailed information about the Yarnell Hill Fire videos in which they found the radio conversations that they referenced in their analysis that we published January 19. In the videos, fragments of radio conversations can be heard from nearby radios as various people shot the videos. The words are difficult to decipher, as they just happened to be in the range of the video camera’s microphone, in the background, but they were not specifically planned to be recorded.
The videos were shot at the Yarnell Hill Fire on June 30, 2013 near Yarnell, Arizona. Nineteen members of the Granite Mountain Hotshot crew were entrapped and killed on the fire. These recordings, along with other investigatory data, may shed a little light on the circumstances surrounding that tragic event.
Information regarding how Holly and John developed their analysis is in another article, The Yarnell Hill Fire recordings — the back story.
As we promised on January 20, we are providing information about the sources of the data below. As far as we know, the first one is not in a public location on the internet, but many people have the Freedom of Information Act files, so perhaps soon it also will be available to everyone, if it is not already. We discovered today that Items 2 through 4 were uploaded to YouTube by Elizabeth Nowicki between January 1 and January 20 and are embedded below. We thank Ms. Nowicki for posting the files.
1. “DivA-Ops Musser”:
SAIT Investigation Record:AO5-20130630: AerialFirefightingStudyPhotosVideos F: PhotosAndVideos: Panebaker:Video:20130630_154232_fire_behavior_EP.MOV
2. “Marsh talking to Abel about making his way off the top”:
SAIT Investigation Record: F: PhotosAndVideos:A2520130630RobertCaldwellVideos:RobertCaldwell_IMG_0749_2389
*This video was posted by EN on You Tube on Jan 1, 2014.
3. “Marsh at house”:
SAIT Investigation Record:AO5-20130630: AerialFirefightingStudyPhotosVideos F:PhotosAndVideos: Panebaker:Videos:20130630_161620_VLAT_SPLIT_1_EP_MOV
*This video was posted by EN on You Tube @1700 on 1-20-14.
4. “Coming from the heel of the fire”:
SAIT Investigation Record: F: PhotosAndVideos:AO8-20130630BlueRidgeHotshotPhotosVideos: Yarnell_Gamble
*This video was posted by EN on You Tube on Jan 2, 2014.
Elizabeth posted Holly and John’s enhanced or filtered audio on youtube. Holly and John’s version does seem to say exactly what they reported. I still find it hard to believe that the “escape route” chosen was being referred to as “vertical cut,” which to me is more surprising than Marsh being at the “house.” What I suspect is that there are identifiable technical differences in the way H & J and WTTKT worked with the raw audio that led to the difference in “enhanced” versions. Perhaps WTTKT could state exactly what he did to his version, and with what, so that H&J can use that as a point of comparison/ control?
Reply to SR post on January 24, 2014 at 6:25 pm
>> SR said…
>> Perhaps WTTKT could state exactly what he did to his
>> version, and with what, so that H&J can use that as a point
>> of comparison/ control?
Sure. Be glad to.
I downloaded the original (full) AIR STUDY video in question
from Elizabeth’s public YouTube link.
I removed the raw audio track from it with a free public domain
program called AoA Audio Extractor. It was saved as a high
fidelity WAV format audio file.
I have access to any number of professional audio editing
programs… but for something like this I have discovered that
one of the best programs out there is totally FREE and in
the PUBLIC DOMAIN.
It is called ‘Audacity’ and I used version 2.5.
Anyone can use this FREE software and do exactly what I did.
You can download ‘Audacity’ 2.5 for Windows, Mac or *Nix here…
http://audacity.sourceforge.net/download/
I loaded the high-fidelity WAV audio file extracted from the
video into Audacity.
Now to work…
The most obvious ‘noise problems’ with that video/audio are…
1) The wind.
2) The cars going by.
3) General background noise.
So the first thing to do is ‘remove’ all three of those things.
Audacity makes it really easy to do that. It has VERY
sophisticated ‘Noise removal’ algorithms built in under
the ‘Effects’ menu. All you do is ‘isolate’ a section of noise
that you want to remove from the track… and then pick
a ‘Remove Noise’ option and apply your ‘noise sample’
to the entire track.
So first I isolated just some pure wind. Removed that.
Then I isolated some pure ‘car noise’. Removed that.
Then I isolated just some ‘background noise’. Removed that.
It was already sounding a lot better… but now we run into
the other problems in this audio track.
The conversation we want is the BACKGROUND radio
traffic being captured… and NOT the two men talking in
the foreground.
Here again… Audacity makes this easy.
You just treat the two men’s voices in the foreground as
‘noise’ and remove that just like the other three things.
So I isolated a section where only man 1 was talking, treated
it as ‘noise’… and did a removal pass on HIS voice.
Ditto for man 2. Just isolate ONLY his voice. Remove it.
NOTE: None of the things being described or targeted for
‘removal’ actually gets completely, totally removed… but
its pretty close if you set the sensitivity high on removal.
So now I’m left with an audio track that has had a lot of
things ‘removed’ and the background audio is already
clearer.
Now for the worst problem.
The background audio itself is still very ‘faint’ and hard to hear.
That’s where a number of other ‘Effects’ tools in Audacity
come into play.
First you run a number of ‘Amplify’ passes, but don’t increase
by more than 10db per pass or you start to bust the waveform.
After each ‘Amplify’ pass Audacity lets you then run a ‘hard
limit’ pass and a ‘Normalization’ pass to pull the higher decibel
peaks back into ‘normal’ range before amplifying again.
If you don’t do that… each amplification pass starts turning
the foreground audio into ‘white noise’… which you don’t want.
So you repeat this ‘Amplify’, ‘Hard Limit’ and ‘Normalization’
sequence say… 4 or 5 times… and now the background audio
is boosted a total of about 50db and coming out louder than
before… but it was so faint to start with it’s STILL hard to hear.
The final step was to just then simply really hard-boost the GAIN
for both channels almost to the point where it’s red-lining.
I believe I brought the gain up to 27db. (on both the left and
the right channels equally ).
This isn’t something you would normally ever want to do
because things start to distort… but in this case it’s the
only way to really bring that background audio all the way
to the forerground.
The by-product of such a drastic boost in the GAIN is that
the entire audio clip becomes VERY, VERY LOUD.
If you listen at full volume it will be nothing but distortion… but
with a GAIN of +27db, and a volume level of, say, half… you
get the best of both worlds. The audio is much clearer but
you don’t run the risk of hurting your ears or your equipment.
By the way… this ‘GAIN boost’ trick is exactly what they use
to annoy the crap out of you when you are watching television
and the commercials suddenly seem so LOUD. They aren’t,
really ( the waveform is actually about the same as the primary
audio signal reaching your TV ) but the big GAIN boost trick just
makes you sit up and pay attention… which is all they
(the advertisers) want, anyway.
So that’s pretty much it.
I should also mention that the original raw audio was recorded
with what appears to be a stereoscopic microphone that had
the balance somewhat off-center. It might have been set that
way… or the microphone might actually have been a little
broken. Regardless… that means that one of the audio channels
was originally louder ( with more waveform ) than the other.
In the H & M clip… they seem to have chosen to include just
the left channel… and just totally wiped out the right channel.
They thought that would be the equivalent of actually doing
‘noise removal’, or something.
Well… that works sometimes… but what they lost from that
conversion was a lot of the ‘fidelity’ since the original background
radio captures were recorded on BOTH channels, not just one.
Maybe they ‘merged’ the two STEREO channels into just
a mono channel and worked on that single channel and that’s
why the right channel in their clip is simply ‘dead’… Dunno.
Regardless…
I made no attempt to ‘go mono’ or just keep one channel or
the other. All ‘noise filtering’, ‘amplification’, ‘hard limiting’
and ‘normalization’ passes were done on the full original
STEREO audio track… as were the GAIN boosts… and the
final result clip is also full STEREO, unlike the clip that
H & M were listening to.
>> SR also said…
>> I still find it hard to believe that the “escape route” chosen
>> was being referred to as “vertical cut,”
It wasn’t. What he ( Eric Marsh ) actually says there is…
+0:54.09 to +1:03.81
Eric Marsh: Well the guys… uh… Granite… is makin’ their
way out the exact escape route from this mornin’… an’ it
heads… ah… (pause) SOUTH. ( slight pause )
mid-slope, cuttin’ over.
Actually… here is what I believe is the FULL transcript of that
video… including the part at the very beginning which is
what causes someone to actually call Granite Mountain
to ask them what their status was ( at 4:16 ).
What is actually captured at the start of this video is someone
confirming with ASM2 that he ( ASM2 / Bravo 33 ) can
actually SEE Granite Mountain ‘behind those hills’ and ‘on the
corner of the fire’… but ASM2 is concerned that they keep
saying they are ‘comfortable’ where he can see them.
ASM2 says that is ‘not credible’ ( based on the fire behavior
he is seeing at 4:16 ) and he INSISTS that someone
on the ground call Granite Mountain and at least insure that
DIVS A (Marsh) is ‘with them’. ASM2 was being forced to play
double duty as Air Attack at this point since Rory Collins
abruptly left the area a few minutes earlier… and ASM2 needs
to KNOW if ‘DIVS A’ ( or simply ‘Alpha’, as ASM2 says ) is
actually with GM where he can see them down there at 4:16 PM.
HUGE CAVEAT: This is MY best interpretation of the background
audio for that entire video. Your mileage may vary, of course, but
I would put money on the following translation for the background
captured radio traffic in this ENTIRE AIR STUDY video…
AIR STUDY VIDEO STARTS
+0:08.23
(Unknown): ASM2, Copy that… behind the hills is where you
place Granite Mountain?
+12.726
(ASM2): They’re on the corner just repeatedly saying that they’re
comfortable and that’s not credible. ASK him whether Alpha can
be placed WITH them.
NOTE: A few seconds later… Someone OTHER than who
ASM2 was just talking to ( someone with what sounds like a
Cajun accent? ) does exactly what ASM2 just TOLD them to do.
He calls Granite Mountain directly and asks them to report their STATUS. If the time stamp on the AIR STUDY video can be
trusted then it is now about 4:17 PM, just before the SAIR says
they are about to leave the two-track road and drop into the box
canyon at approximately 4:20 PM.
+0:40.52
(Unknown): Granite Mountain ( Five? )…
What’s your status right now?
NOTE: A pretty heavy accent on this speaker. Sounds like
Louisiana Cajun? What he says almost sounds like…
“Granite Montun… wuz yo status rat now?
ALSO NOTE: Even though the caller asked for ‘Granite
Mountain’ specifically, and not ‘DIVS A’… Eric Marsh responds
immediately on behalf of Granite Mountain before Jesse Steed
even has a chance to say anything. Jesse only ‘chimes’ in with
his own ‘status’ report about “we’re pushin’ our way down into
the structures” when Marsh is done speaking.
+0:54.09 to +1:03.81
Eric Marsh: Well the guys… uh… Granite… is makin’ their
way out the exact escape route from this mornin’… an’ it
heads… ah… (pause) SOUTH. ( slight pause )
mid-slope, cuttin’ over.
+1:07.18
( Another voice immediately adds to what Marsh said but it is )
( NOT Marsh this time. It sounds like Jesse Steed )
Jesse Steed: We’ll be pushin’ our way down into the structures.
+1:10.16 ( Another voice. Very quickly )
Copy that.
+1:21.17
( Unknown): Ten four. You with Granite Mountain right now?
+1:24.68
(Eric Marsh?): Uh… just checkin’ it out to see where we’re gonna
jump out at.
There appear to be some other background conversations
in this video before we hear the 11 ( ELEVEN ) shutter
clicks of the fellas in the foreground actually shooting photos
of the VLAT drop… but there is also HEAVY road traffic
at this point that will take a lot more work to filter out.
AIR STUDY VIDEO ENDS
WTKTT,
Thanks very much for taking the time for that comprehensive reply. I assume Holly and John are now in a position to circle back to the audio people they used and compare notes. At the end of the day perhaps the most important point, to me, is that there was communication during that time period. If Marsh were to be saying the “escape route” was “vertical cut,” to me that would imply a different set of training and judgment issues than simply “cuttin’ over” implies, but those are sort of micro points.
What I would ask of Holly and John, in addition to circling back to their tech people, is actually a field-note question. I know they have walked the actual area of the YHF with Sonny and Tex among others. Have they, also, actually hiked down a roughly comparable bowl choked with the same general mix of chaparral? Page 20 of this document, http://www.pacificbio.org/initiatives/fire/THE%20YARNELL%20HILL%20FIRE%20AND%20THE%20NEED%20FOR%20FIRE%20ADAPTED%20COMMUNITIES2013july17.pdf , contains one of the better composites of the bowl pre-fire. I am sure Sonny and Joy and other locals could also suggest nearby areas with similar dense brush and topography. And it is also likely that Holly and John have already done this.
I think that experience is relevant to interpreting the audio. It may at the end of the day be clear as a technical matter what is being said. It may also stay murky and require some judgment. If Holly and John have already walked down a similar brush-choked bowl and do think in their judgment that Marsh believing using sawyers to create a “cut vertical” escape route, through that bowl, was something Marsh was likely to do and say, then so be it. As to the effectiveness of doing this, I’d go so far as to say it might make sense to have a couple teams do test “vertical escape route clearings” through similar chaparral. We know from GM’s own extremely slow progress in that bowl that whatever cutting they may have done didn’t speed their progress much, if any. A slow walking pace is around 2 mph. GM moved at roughly a half-mile per hour. I believe that a couple fit and efficient teams trying to clear similar brush would likewise have a bear of a time. This could simply mean that GM made a spur-of-the-moment choice of tactic that could have worked elsewhere, but was the wrong choice that day. It could also imply, if doubt does remain over what was actually said, that it would be more likely for Marsh to have said “cuttin’ over,” as WTKTT suggests.
I would have to agree with WTKTT… its sounds like “just checking it out to see where we’re going to jump out at”. I was on the line that day, and while I did not hear this transmission, his statement and the way he says it makes sense to me. I don’t hear “at the house” at all. Further I never heard any transmissions about GMIHC going to protect structures. In fact, I was with OPS when GMIHC stated they were in good black and staying put.
Which OPS?
Field
Field OPS that day was Todd Abel ( according
to all documentation ).
Paul Musser was ‘Planning OPS’.
Was Abel aware that Musser had contacted GM
and requested them to send resources to
Yarnell?
Just listen to the enhanced audio clip.
Make up your own mind(s).
The transcript I have published is simply what I am hearing.
That’s all.
Regarding section (3) above entitled ‘Marsh at the house’.
The effort that was made to translate that section of audio is admirable…
but no one ever says the word ‘house’.
When the ( still unknown ) speaker asks someone ( supposedly Marsh )
“You with Granite Mountain right now?” at +1:21.17… the response is NOT
( as previously reported ) “I’m at the house where we gonna jump out at”.
Whoever responds to that question ( it MIGHT be Marsh but it that hasn’t been
verified. It could be Steed. ) does NOT use the word ‘house’ at all.
The actual transcription on that section of audio is as follows…
+1:21.17
Ten four. You with Granite Mountain right now?
+1:24.68
Uh… just checkin’ it out to see where we’re gonna jump out at.
An enhanced audio clip of just this particular section of the
background conversation is now available on YouTube.
WARNING: The GAIN is set VERY HIGH on this audio clip which
makes it easy to hear what is being said but be careful what
your local volume is set to before you play this clip. The
high gain setting makes it VERY LOUD.
The section of audio in question repeats SIX times and
only represents this part of the background conversation…
NOTE: The ‘ten four’ lead-in appears to be from
a speaker OTHER than the one who asks the question.
This appears to be just an overlap in conversations at
that exact moment.
+1:21.17
(Speaker 1) Ten four.
( A different speaker ) You with Granite Mountain right now?
+1:24.68
Uh… just checkin’ it out to see where we’re gonna jump out at.
Here is a link to the enhanced audio clip on YouTube…
http://youtu.be/eq0rtX5OgKI
In your ” Uh…just checkin’ it out to see”, you have 8 syllables in this part. There are clearly not 8 syllables in this part of the actual conversation on the audio.
It’s possible ( the way Marsh talks ) that the
2 words ‘checkin’ it’ come across as one word.
Southerners do that at lot. They eliminate ‘spaces’
between words sometimes.
Regardless… what do you think of the enhanced
audio clip?
Are you still hearing “I’m at the house”?
WTKTT,
Thanks for linking to that clip. It’s still somewhat difficult to hear, but much clearer than before. For what it’s worth, these ears hear the same words that you transcribe.
If you have time to do any of the rest of that video, that’d be terrific.
Reply to Sitta on January 22, 2014 at 4:24 am
>> Sitta said…
>> Thanks for linking to that clip. It’s still somewhat
>> difficult to hear, but much clearer than before.
>> For what it’s worth, these ears hear the same
>> words that you transcribe.
Thanks. I believe that is ( finally ) accurate.
I, myself, would put money on it, anway.
>> If you have time to do any of the rest of that video,
>> that’d be terrific.
Already done ( here )… but obviously with what just
happened I am only going to release new transcripts
after I review things about 100 times.
Example: There is even MORE background traffic
on there that has NOT been transcribed yet.
Early on… we APPEAR to hear fire command actually
ASKING ASM2 to confirm Granite Mountain’s location
‘behind those hills there’… and ASM2 does so.
Keyphrase: APPEAR to be hearing. Not final yet.
That’s going to be pretty controversial so I have to
make sure I have that right before ‘jumping any guns’.