Thursday, March 29, 2007

Walking through the Luneau mire

Come with me and take a frame-by-frame walk through the Luneau video. Maybe this way I can show exactly how incorrect are the claims by Sibley, Bevier, Collinson, etc. that the video shows diagnostic features of a Pileated, and no features that are inconsistent with a Pileated.

By the way, I think the convention of numbering these frames to 0.1 ms (33.3, 366.7, etc.) is silly. The extra decimal place is unnecessary, and I'll call these frames 33 and 367. The video from which these frames are extracted is copyright David Luneau.

We start here at frame 167, where the bird is first emerging from behind the tree trunk. I'm going to focus primarily on the bird's right wing, as that is generally the one presented best to the camera. Here we see this wing in mid downstroke, showing its underside. Two things to note here: First, the white of the underwing in the image is bordered on all sides by a darker fringe; secondly, the wing is strongly flexed at the wrist, giving a distinctive bowed shape.

Next is frame 183. Here the flexed shape of the wing is especially evident. There are also four distinct dark blobs around the edges of the wing: two on the upper (trailing) edge, one on the lower (leading edge), and one at the tip. And here we first see the pattern that holds true for the rest of the video. Note that each of these blotches aligns with a darker shadow in the background. Even the distinct dark wingtip is in fact aligned exactly with the lower end of a shadow visible in f167.

The next few frames are a blurry indistinct mess as the wing reaches the end of the downstroke and then undergoes the rapid folded upstroke. So we'll jump ahead.

Here at frame 233 the wing once again begins to show something other than gray smears. We see the large white underwing area bordered on the left and the upper right by darker areas. Again, these dark areas align with shadows behind the bird.

Moving on to frame 250, the beginning of the downstroke, one of the most dramatic changes is the great increase in the apparent size of the white underwing from f233. It seems the wing, though appearing fully raised in f233, was still partially folded. Also note that the wing is already flexed strongly at the wrist. This is one of the frames that has been said to show a dark trailing edge. I see four dark areas bordering the wing: one on the left edge adjoining the edge of the tree trunk, a distinct one at the upper left , a smaller one at the top, and a general diffuse dark area to the right of the wing (close to the wing's leading edge). Once again, all align with background shadows. Let's look at the next frame and see where they go.

Here in frame 267, at first glance it appears that the four black areas from f250 have traveled with the wing. First glances are deceiving. Look again, and you will see that the first black spot (the one on the left) is still where it was... attached to the tree trunk and separated from the wing. The second spot has traveled down the trailing edge of the wing, remaining aligned with the same shadow. The now much larger black area at the trailing edge of the outer wing is aligned with a much larger background shadow that the wing is now passing in front of. And the black spot to the right of the bird now appears to be a black leading edge to the wing. Yet again, all black fringes are aligned with and attributable to dark background areas. The strong downward bend of the wing at the wrist remains very clear.

In the next frame (283) the wing is surrounded by black beads. The large black fringe that was on the outer trailing edge of the wing in f267 has now shifted proximally (remaining aligned with the shadow) and broken in two. A new black spot has appeared on the outer trailing edge where (lo and behold) the wing has passed in front of another background shadow. And at the wingtip, a darkish wingtip is forming where (guess what?) the wingtip image has begun to interact with yet another background shadow. Meanwhile on the leading edge, the black spot has condensed, remaining aligned with the background shadow. All through the video we see this: Black fringing blobs that remain aligned with the background features as the wing moves in front of them. There is no reason to treat any of these black blobs as real. And of course, the wing remains strongly bowed.

Here in the next frame (300) on the right wing we see a prominent black wingtip. Look back at f283 and especially f267, though, and notice how the shape of that black wingtip closely matches the shape of the background shadow in front of which the wingtip is now passing. The other black blobs along the wing fringe have largely faded. Now, for a change, look at the left wing which here is probably presented as well to the camera as it ever is. It seems to show a gray fringe around both trailing and leading wing edges. More prominent thougs is the black wintip. This one looks real, as it is independent of the background. It is of course smeared into a smooth arc and its actual shape is not discernable. Another very interesting feature on this wing is the apparent zonation in the white: bright on the outer wing, dingier on the inner wing, possible a darker indistinct border at mid-wing. This has been described as showing a Pileated's "bright white primary bases," presumably contrasting with the duller white underwing coverts. However, this fails to take into account the bend of the wing at the wrist and the forward rotation of the wing in direct flight. The outerwing is held more vertically and rotated more towards the camera than the inner wing; this is likely adequate to account for the underside of the outer wing appearing brighter in this image. And in some other images the pattern is reversed.

Skipping through the blurry upstroke again to frame 350. Just as in f233, we see what looks like a fully raised wing, except that it appears short relative to the next frame so it seems to not be fully extended even if it is fully raised. And we also see a distinct black border to the left (presumably) trailing edge of the right wing, at least on its proximal portions. This is another frame that has been said to clearly show the Pileated's black trailing edge. And in this case there is not a background shadow to blame it on. The quite reasonable suggestion as to why it is most prominent proximally is that motion blur is least there. So in the next frame, we should expect to see it contract somewhat but persist on the proximal edge. And in the next frame, what we do see is...

(frame 367) Hmmm... in fact it has disappeared proximally and been replaced distally by a big fat black fringe on the outer wing and wing tip. And, as usual, this big fat fringe aligns with a big fat background shadow. There's no hint at all of the black trailing edge on the slowest-moving proximal portion of the wing. So what was going on in f350? Well, there are two things to consider. First, in f350 the wing was evidently not yet fully extended, and there is also the issue of the left wing which should be in the image somewhere. I can't account for exactly where that black fringe came from, but there's not much clear indication that it came from a real black edge on the wing when you look at this subsequent frame. There's no suggestion of it on frames 467 and 583 when the wing is once again in similar orientation. Another thing to note here relevant to the "bright primary bases." Here, it is clearly the inner wing that looks brightest. Once again, the vertical portion of the wing looks brighter than the more downward-facing (due to the bend at the wrist) portion of the wing, as would be expected.

In the next frame (383) the rapidly-moving wing is dissolving into a hook-shaped blur (hook-shaped from the continuing strong bend at the wrist). The big black border from f367 is reduced to faint hint, still aligned with shadows.

In the next two frames the right wing passes in front of relatively plain background. Both here (frame 400) and below in frame 417 it is worth noting that the right wing white underwing shows a gray fringe around all edges. What should we make of this?

Here in f417 the left wing also show a quite uniform gray fringe around all its edges. Note that though the wing itself is strongly motion blurred, the the fringe is more consistently narrow around all sides, because it is an imaging artifact not a part of the wing.

I could repeat this for every frame but it would get tedious. The pattern is established: black fringe artifacts that occur in association with background shadows along all edges of the white underwing; wings held bowed with a strong bend at the wrist throughout the downstoke, no consistent appearance of real black on the underwings anywhere but at the outermost wing tips. I do want to sample a few frames from farther on to clarify a few points.

here in frame 467 is the next time the wing is held upward, comparable to frame 350. Again, in comparison to the next frame, the wing is clearly not fully extended. Here we see a nice dark edge to the underwing: a leading edge. Of course, it aligns with the background shadow. No reappearance of f350's apparent broad black trailing edge.

A very interesting thing happens here in frame 483. As the wing extends and begins the downstroke, it aquires a quite prominent black leading edge that does not align with a background shadow. The large black spot near the wingtip does align with a shadow, of course. So this balances f350 in a sense: one each of black leading and trailing edges that are not readily accounted for by interactions with background shadows. It would be folly to base an ID of this bird on either.

And finally, I like this frame for its comic value. This is frame 700, the end of the upstroke two wingbeats later. I count at least three black wingtips among the two wings, and black on both the leading and trailing edges of the right wing.

To summarize and reitterate: The black edge features imaged around the wings of the bird in this video do not represent real features of the bird. There is no evidence at all that this bird has a black trailing edge to its wings. The conclusions of Sibley, Bevier, Collinson, etc. about this matter have been based entirely on misinterpreted and selectively interpreted image artifacts, and should be disregarded. Sibley was quoted as saying that Collinson had demonstrated just how Ivorybill-like a bad video of a Pileated can appear; in fact what they have both demonstrated is just how bad experienced birders can be at interpretation of bad video.

Of course, no evidence for a black trailing edge does not prove the presence of a white trailing edge in such a poor quality video. Both species have black markings on their underwings (other than the wingtips) and the video does not resolve any of these. I have generally promoted the conservative approach that nothing can be inferred about the locaion of any hypothetical black on this bird's underwings from this flight sequence. Others have argued, however, that the Pileated black trailing edge is generally much wider than the Ivorybill medial bar, and it should have been clearly resolved in multiple frames if it were present. Cornell's experiments with shooting videos of painted models using the same video camera under the same circumstances suggested this. Still, it remains that the distinguishing characteristics of the underwing pattern are not resolved in the video for the case of either species, at least in the flight sequence I just discussed. But...

There are a few frames of the bird's underwing before it disappars behind the tree. These are the controversial frame 33 and adjacent images. As I have said in previous writings, I tend to agree that what is seen here is the underside of an extended wing, not the topside of a folded wing. Let's look at these frames again in light of what we've learned from the rest of the video:

Infamous frame 33. Best interpreted (to my thinking) as the trailing edge of the underside of the bird's right wing, extended as the bird begins to launch from the tree. I see just white. The faint gray fringe is barely present, no stronger than the gray artifacts present around all the white wing edges in other frames examined above. Still, there might be an unresolved black trailing edge.

Move on to the next frame (50). The wing is travelling to our right and being eclipsed behind the tree. We are only seeng the rearmost few centimeters of the underwing. Predominant color? WHITE. Faint gray fringe is just like the artifacts seen everywhere else. If this were a pileated underwing there should be no white left at this point. In Sibley's interpretive sketches of these frames he has drawn a razor-thin black trailing edge to explain this. The trailing edge is so thin, we are told, because of defocus, white bleed, and motion blur. This is well and good, except than none of these will make white bleed around the corner from behind a tree, which is exactly what Sibley has drawn the white doing.

In the next frame the wing has vanished entirely from view; all we have left is the end of the tail. Before you get excited about those two black pixels on the edge of the tree trunk above the tail, look back at the previous frames and you will see they are a fixed portion of the tree. They are not part of the bird.

Indeed, the bird appears to have a white trailing edge to the underside of its right wing. The black trailing edge had every chance to reveal itself in these two frames, and was still a no-show.

I'm not going to repeat my whole arguments about flight mechanics again here; I've covered them extensively in earlier posts. That bowed wing shape is characteristic of this bird in every visible wingbeat and is very unlike the wing dynamics shown in every Pileated video I have seen. We can't say that it is Ivorybill like, but we can say it is not Pileated-like. I wish someone would produce some comparable take-off footage of other Campephilus species.

At the end of our walk we find that, in contrast to the widely promoted view of Sibley et al, the Luneau video does not at any point show any distinctive characters of the Pileated's underwing pattern; indeed there is a strong indication that the bird has a white trailing edge to its underwing. It also shows flight mechanics that are clearly and dramatically atypical for a Pileated Woodpecker. In the exact opposite of the views promoted by Sibley and others, we find nothing in this bird that is inconsistent with its being an Ivory-billed Woodpecker, and some features that are clearly inconsistent with its being a Pileated Woodpecker.

ADDENDUM: I've had a couple of anonymous requests to give my assessment of the second flight sequence in Mike Collin's 2006 Pearl River video. I just looked at it again (his enlarged, lightened version with the frames numbered). I have a difficult time making out any detail of the wing postures other than a general sense of when each downstroke begins and ends. I can't really resolve more than that. For the first 9 frames as the bird launches and rolls I can't even resolve that much very well. From frames 10 onward I make out a flap every 4 frames, or 7.5 beats/s, but no more detail. I don't feel that I can make any assessment of that bird's wing postures in flight in order to make any comparison to either the Luneau or known Pileated videos.

Comments: Sign them with your real name and stay focused on issues raised directly in this posting, please. There are many other venues for generalized, wide-ranging, and anonymous woodpecker discussions. Thank you.


Post a Comment

<< Home

Site Meter