UFO over Trindade Island: What is wrong with this picture?

by Tim Printy © June 2004

Updated March 2005 (updates in yellow)

Mirror, mirror.....

Back in 1958, The US Naval attaché, Captain Sunderland made the following observations about the images:

...the photographs furnished by the Brazilian Navy are unconvincing. Details of the land are extremely sharp but the disc is hazy and has little contrast and shows no shadow effect. It also appears that the object was inverted in photograph 2 compared to 1 and 3. Also the papers have mentioned extremely high velocities and there appears to be no lateral blurring as would occur with any reasonable shutter speed. (Hynek 240)

In his book, Hynek belittles much of Sunderland's comments and refrains from remarking about the inversion he noted. Over forty years after the case had been first examined, Martin Powell began to look at the Trindade images in a manner mentioned by Captain Sunderland and noted that he was correct in that the images were inverted.

By increasing the contrast and gamma properties of the image, P1 began to show a great deal more ‘noise’ around it, and after a while the similarities between the two images became quite clear.  Most of the features are quite individual in shape and would not be expected to occur in another image, in the same relative position, by chance.  And yet they are seen to occur in both photographs.  As the edge of one image is followed around and compared with the other, notable similarities are found.  For example, the ‘dome’ of the saucer (or the windshield of the aircraft) and the light area below and to the right of it are very similar in both images.  The area around the nose-wheel (which is enveloped in the ‘noise’) and the left half of the image are also similar in both pictures.  The reader should try viewing these images at an oblique angle; the similarities may appear more obvious this way.  As a result of this experiment, I would say there is a strong case for P2 having been an inverted copy of P1, as the Preparing Officer had claimed. (Powell)

Even more interesting is how Kentaro Mori reports that all the images appeared to be the same object:

The result is that we can now more safely declare that all four images are too much similar to be different images of a real physical object photographed as it travelled through the sky. We not only have rotation, of more than 180 degrees with no significative change of the image (from P1 to P2), but a mirroring of the image (P4 in relation to P1 - horizontal flip - or P2, P3 - vertical flip). This mirroring cannot be convincingly explained by a "flipping UFO", even one with "no shadow effect", as recently proposed by researcher Brad Sparks (cf UfoUpDates list) as a response for the acknowledged inversion of the image of it from P1 to P2. (Mori)

While Brad Sparks contends this was due to the UFO flipping upside down and shifting its orientation, it is amazing that, if it were tumbling as it flew, all four of Barauna's photographs would show the "ring" of the UFO as essentially level and from the same angle. With an object tumbling and undulating around, it seems likely that one photograph would show the UFO's ring from a different viewpoint. Instead we appear to always see a two dimensional view with the "ring" always visible from the side. Did the UFO always bank towards the photographer to present an "edge-on" view of it's ring?

Quantitative analysis or guesswork?

source: http://www.ritzcam.com/catalog/images/Roll28ERY.JPG

Much of the recent analysis, at least we are told, is being done from "first generation" prints from the original negatives. According to Barauna, he used a Rolleiflex 2.8E camera. The 2.8 series Rolleiflex models came with an 80mm F2.8 lens (field of view approx 38.6 degrees) and normally used medium format roll film (120). Kentaro Mori computed a field of view (FOV) of 41.1 degrees for the camera but this assumes the entire 6CM of the negative is used. This is an incorrect assumption since 6X6 film uses only 5.6 CM of the negative. This 6X6 negative introduces some problems because this ratio does not correspond to a standard 8X10 format for printing. As a result, an 8X10 print would never show the entire negative. Therefore, it is difficult to figure out the scale. One can only assume that the longest length of the print corresponds to the full length of the negative and the vertical portions of the negative were cropped off. However, Mori has shown that this is not the case when he produced prints from the same negative that show different fields of view!

Kentaro Mori was kind enough to allow me access to the scans of these first generation prints. Apparently, these are what Mori referred to as the "Gevaerd" prints. Unfortunately, just as Mori has noted, each print is scaled differently and enlarged to confuse the situation. One can only use pertinent land features to try and get the images to the correct scale. Kentaro made such efforts in his use of the prints to create his panorama. According to him, he had to rescale P2 to 80% of it's original size and P3 to 95% of it's original size. When I tried to duplicate his efforts, I discovered that my adjustments were about the same (79.2 and 94.2%).

Using the 38.6 degree horizontal width as a basis and assuming the widest print (P1) showed the entire negative (which it apparently doesn't based on the "Olmos" prints), I computed some approximate angular sizes of the UFO after the images were scaled correctly. In the case of P1, the horizontal portion of the visible print was 2740 pixels. If this was the entire 38.6 degrees, then this would create a conversion factor of about 0.0141 deg/pixel (38.6/2740).

P1 = 97 pixels * 0.0141 deg/pixel = 1.37 degrees

P2 = 120 pixels * 0.0141 deg/pixel * 0.792 = 1.34 degrees

P3 = 105 pixels * 0.0141 deg/pixel * 0.942 = 1.40 degrees

P6 = 54 pixels * 0.0141 deg/pixel = 0.76 degrees

I would consider these values a maximum size because if the prints do not show the full 38.6 degrees, then the resultant angular size would be lower. For example, if the number of pixels for a print from the full negative was actually 4000 pixels (using the same scale as above), the correction factor would be 0.0097 deg/pixel and therefore decrease the angular size of the images.

While, these measurements come somewhat close to the values presented by Brad Sparks on the UFO updates mailing list of 1.5, 1.5, 1.7 and 0.6 degrees respectively, Sparks values seem too large based on what I measured. Differences can be explained by varying measurements, different prints, and incorrect scaling by either myself or Sparks. It is possible that Sparks used an incorrect value for FOV such as using the 75mm lens found on the 3.5 Rolleiflex or he could have duplicated Mori's error in using 6CM for the negative size. I suggest that anyone desiring to check measurements can attempt them from the photographs at http://www.rr0.org/1958_Trinite.html. These seem to be what Mori refers to as the "Olmos" prints. While these are lower resolution than the ones Mori allowed me to have access to they can provide some method of duplicating work. When I attempted the measurements on these prints, I arrived at similar values to the ones above. The print with the largest horizontal dimension was P3, which was 995 pixels in width. The scaling for these pictures appears to be about the same between prints. Therefore, the 995 pixels would be the closest to the full negative and make the scale 0.0388 deg/pixel (38.6 deg/995 pixels). When the UFO was measured with this scale, I arrived at the following values:

P1 = 33 pixels * 0.0388 deg/pixel = 1.28 degrees

P2 = 36 pixels * 0.0388 deg/pixel = 1.40 degrees

P3 = 38 pixels * 0.0388 deg/pixel = 1.47 degrees

P6 = 20 pixels * 0.0388 deg/pixel = 0.78 degrees

I considered P1 a poor image and it is possible the measurement here was not very accurate because the edges of the "disc" were washed out by the overexposure of the sky in the printing process (this was not so evident in the "Gevaerd" print of this image). However, one can see that the values are very similar to the ones I got on the "Gevaerd" prints (within 0.1 degree).

Something worth noting is that the UFO appears to have the same approximate angular size in three of the four photographs. Even Sparks only notes a difference of 0.2 degrees between P1, P2, and P3. Recall that Baruna stated:

I shot two photos before it disappeared behind Desejado peak. My camera was set at at speed of .125 with an F/8 aperture, and this was the cause of the overexposure error, as I discovered later. The object remained out of sight for a few seconds, behind the peak, reappearing larger in size and flying in the opposite direction. It was lower and closer than before and moving at a higher speed. I shot the third photo. (bold emphasis mine) (Lorenzen 165)

Could Baruna actually be able to differentiate a size change of about 10%? The description implies there was a noticable size difference. Had the object doubled in size that would be noteworthy but the actual size variation indicates the object (if it were an actual craft flying through the sky) was about the same distance away in P1 through P3.

Based on other statements by Sparks, I measured the angular distance between P1, P2, and P3 and came up with the following values using the "Gevaerd" prints :

P1-P2 = 45 degrees

P2-P3 = 24 degrees

Again, I assumed the horizontal distance of the widest print was the entire length of the negative. Again, if the prints are not showing the entire field of view, then these values can be considered a maximum and not the minimum angle. These are similar to the values presented by Brad Sparks, who stated they were roughly 50 and 25 degrees respectively.

Brad Sparks used these values to determine how much motion blur was supposed to be in the photographs. If the photographs did show a fast moving object in the sky, the blur introduced by the motion of the object across the sky could be noted. This would only amount to fractions of a degree and figuring this out would almost be impossible because we do not know the duration of time between P1 and P2 or P2 and P3. Additionally, for P2 to P3, we are not talking a straight line distance because the UFO, according to Barauna, moved from P2 to behind Desejado peak and then back to the position P3. However, Sparks seemed confident he could actually calculate, with accuracy, the amount of motion blur to deduct from his measurements. According to him, these values were:

Photo 1 1.5 degrees (1.1 degs if 25 degs/sec motion blur is deducted)

Photo 2 1.5 degs (1.3 degs if 12 degs/sec motion blur is deducted)

Photo 3 1.7 degs (1.5 degs if 12 degs/sec motion blur is deducted) (Sparks. Re: Trindade Photos)

I note some possible errors in Sparks computations, which, we are led to believe, have been examined by this "Trindade research project" team:

  1. The average of two seconds he uses for his rate of motion is apparently based on the 14 second duration between P1 and P6. If one were to divide by 6, we would get 2.3 seconds. However, P1 would be taken at time zero and, therefore, the remaining five pictures taken over the remaining 14 seconds. This makes the average time difference equal to 2.8 seconds. This would change the angular velocities (based on Sparks values of angular distance) to about 18 and 9 deg/sec.
  2. There is a problem of the accuracy of the timing tests actually reflecting the true time between photographs. Based on Barauna's testimony it appears that each photograph (especially between P2 and P3) was not separated by the same time interval.
  3. According to Barauna, the UFO went behind Desjado peak and then returned to it's position for P3. Assuming he correctly identified the peak it passed behind, this makes the angular distance covered to be at least 100 degrees (angular speed of over 33 deg/sec) and not the 25 degrees (12 deg/sec) Sparks seems to imply. The resultant blur for 33 deg/sec during a 1/125th second exposure would be roughly 0.26 degrees.
  4. Sparks apparently doubled the extent of the blur for some unknown reason. If one uses 25 deg/sec with a 1/125 second shutter speed, the amount of motion during the photo would compute to 0.2 degrees (25deg/sec * .008 sec) of blur (vice 0.4 ) and 0.1 degrees (vice 0.2) for 12 deg/sec.

Important to note is that Sparks used the correction value of 12 deg/sec for P2 because the image at P2 becomes his mathmatical centerpiece in verifying the authenticity of the photographs.

Sparks notes that the "dome" of the UFO appears flattened 30% at P2 and therefore vindicated his computations of the angular speed.

Here is where quantitative science comes in, listen carefully now: The amount of flattening is a lengthening of the object in the horizontal direction of flight motion of at minimum approximately 0.3 degree of subtended arc (the dome is about 1 deg wide at widest, 30% of 1 deg is 0.3 deg), which occurs during the 1/125th second exposure time. Therefore, based on linear-motion-blur the object is traveling about 37 degs/sec across the sky (125 x 0.3 =3D 37.5 degs/sec). It's simple physics. (Sparks Re: Trindade Scientific)

I tried to determine how he arrived at the amount of "flattening" for the dome since he really did not explain how he calculated it. I noticed there is a difference in the width to height ratio between P1 and P2 in relation to P3 of about 25% which is close to his 0.3 value. How the "flattening" measurement verfies his values, is best explained by Sparks:

Completely independent of that linear-motion-blur measurement I measured the flight path of the UFO in the sequence of photos and found it moved very roughly 50 degrees between Photos 1 and 2 across the sky over the ocean and the Trindade island, as it started slowing down according to the witnesses' testimonies, and it is about 25 degrees between Photos 2 and 3, showing that indeed the UFO did slow down just as the witnesses stated. The Brazilian Navy reenactment with photographer Bara=FAna arrived at an estimate of about 2 seconds between each photo, hence the rough AVERAGE angular velocity of the UFO between the first two photos is about 25 degs/sec, slowing to about 12 degs/sec between the second and third photos. Therefore the PEAK angular velocity could well have been 37 degs/sec or more depending on the variations of speed, to average out to 25 degs/sec over the entire interval between the first two photos. (Sparks Re: Trindade Scientific)

Sparks is incorrect again in some of his assumptions and statements:

  1. Barauna stated the object was moving much faster at the time of P3 and I could not find any published witness statement which stated the object "slowed down" between photos 1, 2, or 3 as Sparks claims. Even in the Hynek interview Barauna does not say it slowed down but stated it was moving at a relatively low speed between P1 and P2. Barauna only states the object was moving faster at P3, which means that it should have had a high degree of flattening during this picture. It exhibits no "flattening".
  2. Sparks again makes the error in calculating that the time difference between photographs is 2 seconds when, based on Barauna's statement about his time tests (6 pictures in 14 seconds with the first photo taken at time=0 and the sixth taken at time=14 sec), it should have been closer to 3 seconds and not "about 2".
  3. Sparks makes the incorrect statement that the object only traveled the distance between the two locations shown in P2 and P3. Sparks should be familiar with the terms "displacement" and "distance". The "displacement" would have been 25 degrees and give an AVERAGE angular velocity (which he does state) of 12 deg/sec (according to his calculations) but based on one of Barauna's statements, the object's distance traveled was much further. If Barauna correctly identified that it passed behind Desajado peak and not the unknown peak in the photograph, then this length of travel could have exceeded 100 degrees! Trying to calculate the true motion blur at the time the photograph was taken would not use the average angular velocity but the actual angular velocity.
  4. Sparks also contradicts himself. In his original corrections for image size, he used a correction for P2's angular size of 12 deg/sec. Now he states at the time of P2, the object was moving at 37 deg/sec! If he had used 37 deg/sec for his motion blur correction, he is suddenly faced with a problem. The UFO image size for P2 would be less than P1 meaning it would have to be farther away at P2 than at P1! This directly contradicts the testimony and the published flight path (which is incorrect anyway).

Using two different values for the angular speed at the same moment in time and then stating they confirm the other makes one question the "science" behind what Sparks is really doing. Of course, he could make a case again for his flattening measurement if he used some of the corrections I mentioned. For instance, he could now say the distance traveled between P2 and P3 was 100 degrees in 3 seconds or 33.3 deg/sec. Using this he can now state it verifies his "flattening" calculation of 37 deg/sec because the UFO started to speed up at time P2 (instead of slowing down). This, however, contradicts Barauna's statement that the object was moving much faster at P3, which shows no "flattening". Perhaps the UFO chose to become motionless the exact moment Barauna took his P3 photograph.

All of this motion blur described by Sparks seems to have escaped close examination by APRO's analyst, Hopf, who stated,

It is interesting to note that the outline of the object is quite sharp in views Nos. 2 and 3 where it had slowed down, but slightly blurred in Nos. 1 and 4 where it was going at a higher speed and the shutter setting was not high enough to freeze the motion. (Lorenzen 167)

Hynek also remarked about how there was minimal blur in the photographs during his interview with Barauna. Sparks seems to imply both Hynek's and Hopf's observations were flawed although he never mentions their comments. Instead, he seems more interested in criticizing similar comments made by Captain Sunderland of the US Navy, who also noted the lack of any blur. My personal observation is the UFO does not appear to be significantly blurred by motion at all and certainly not as much as Sparks' calculations imply. Most noticable is P3 (which Barauna implied was when it was going the fastest and was closest) where the outlines of the UFO appear pretty sharp and there is no significant "flattening" (which by all accounts there should have been). In my opinion, what Sparks has proven is that while there should be noticable motion blur in the photographs, the UFO appears relatively sharp and clear. This directly contradicts the story being told by Barauna, Filho, and Viegas that the UFO was moving at high speeds across the sky.

Sparks went even further in discussing his measurements on the photographs and made the astounding claim that he had proof that the UFO was "flipping over" as it flew:

I have measured what appears to be rapid rotational motion at the extreme edges of the object, amounting to something on the order of 400 to 800 degs/sec around a central axis, and the angular velocity of rotation changes from photo to photo (hence my 400 to 800 degs/sec figures). (Sparks Re: Re: Trindade)

When John Rimmer read this, he was somewhat amazed by the figures and stated:

What on earth is this all about? How can you possibly calculate the rotational motion of an object from a few fuzzy photographs? How can you possibly deduce that the angular velocity of rotation changes from photograph to photograph? What evidence do you have the object was even spinning? (Rimmer Re: Trindade)

Sparks never responded directly to Rimmer's question (that I could identify from the list archives), which prompted Rimmer to ask the question again while Sparks was responding to Kentaro Mori, "How can you possibly work out the speed of rotation - always assuming there is any rotation - from a series of still pictures?"(Rimmer Re: Trindade Scientific). Once again Sparks never directly responded to this question. However, he did hint at how he was able to determine it by stating to Mori:

The telltale giveaway that this is rotational blurring (as with an object flipping over) is the fact the blurring is greatest at the extreme edges of the object and that the image at those edges is dispersed up-and-down vertically in the direction of the rotational motion rather than laterally as with rectilinear bulk motion of the object in flight (Sparks Re: Trindade Scientific)

Apparently, Sparks is measuring the amount of blurring at the edges of the object's "ring" and comparing it to the center. Looking at the scans that Kentaro Mori allowed me access to (which were scanned at 300 pixels/inch for approx 6 million pixels), I could not readily identify these indications of "rotational blurring". The amount of blurring of the edges compared to the center of the "ring" would have been extremely small in 1/125th of a second and near the limits of resolution. One could suggest that Sparks is trying to pull data out of the noise (caused by film grain, camera and printer lens resolution, scanner resolution, etc). Maybe this is why he would not give any exact values or methods in how he arrived at his conclusions.

Since Sparks refrained from giving any details on how he arrived at his values, John Rimmer becames somewhat frustrated and stated, "This is getting into the realms of fantasy. How on earth can measurements of this precision be produced based on what Brad Sparks himself admits and vague and blurry images?" (Rimmer Re: Trindade Scientific). One must recall that in the Warminster photographs of 1970 (which were a staged hoax), all sorts of "scientific" measurements were made by UFO organizations from a series of photographs. These measurements turned out to be completely wrong. Sparks, apparently unwilling to give details to those who asked about his calculations, was happy to suggest that he had made an amazing discovery that vindicated the photographs. Instead, he may just be finding data that fits his interpretation of the events in question.

A more interesting aspect of Sparks analysis is how he has recently stated that as early as September of 2002, he was aware of issues that brought into question the fourteen second time period of the event. He now suggests that the time period is 1-3 minutes between photographs one and six because of problems with the photographs concerning wave patterns. If this is the case, why did he perform all these calculations based on the fourteen second time frame in late 2003? Using one time frame to verify his "quantitatvie science" and another to explain away problems with the photographs makes one question the objectivity of such analyses.

All of Brad Sparks measurements, as well as mine, can be considered suspect based on what we know about the prints used. As Kentaro Mori has mentioned, many of them are enlarged and cropped to the point one can produce errors in the measurement that would throw off any precision. Without the negatives, one can not make a very accurate measurement at all. I am sure Sparks "expert" analysis will provide plenty of numbers and analysis but if he is as careless as he apparently was in his statements on UFO updates, one might really question his motives. Perhaps John Rimmer's remark that, "This is not quantitative science, it's guesswork dressed up with impressive looking figures" (Rimmer Re: Trindade Scientific) seems most accurate when it comes to these measurements. What is apparently indisputable at this point is that the UFO image was the same rough apparent size for three of the four photographs (Sparks has them only differing by 0.2 degrees while I measured them to be much closer). This means the UFO was either the same rough distance from the ship (something Barauna denies in his description of the flight path) during most of the encounter or there is something wrong with the images in the photographs.

Works Cited

Hynek, J. Allen. The Hynek UFO Report. New York: Barnes & Nobles, 1997.

Lorenzen, Coral E., Flying Saucers: The startling evidence of the invasion from outer space. New York: New American Library, 1966

Mori, Kentaro. The amazing UFO. Available WWW: http://www.geocities.com/airsmither/trind.htm

Powell, Martin J. The Trindade Island UFO: A detailed study of photos 1 and 2. Available WWW: http://homepage.ntlworld.com/mjpowell/Trindade/Trindade.htm

Rimmer, John. "Re: Trindade" 5 November 2003.UFO Updates Mailing list. On line posting. Available WWW: http://www.virtuallystrange.net/ufo/updates/2003/nov/m05-013.shtml

-. "Re: Trindade Scientific Update - Rimmer " 11 November 2003. UFO Updates Mailing list. On line posting. Available WWW: http://www.virtuallystrange.net/ufo/updates/2003/nov/m11-009.shtml

Sparks, Brad. "Re: Trindade Scientific Update [was: Trindade] - Sparks" 7 November 2003. UFO Updates Mailing list. On line posting. Available WWW: http://www.virtuallystrange.net/ufo/updates/2003/nov/m07-010.shtml

-. "Re: Trindade Photos A Fake? - Sparks" 26 November 2003. UFO Updates Mailing list. On line posting. Available WWW: http://www.virtuallystrange.net/ufo/updates/2003/nov/m26-041.shtml

-. "Re: Re: Trindade" 4 November 2003. UFO Updates Mailing list. On line posting. Available WWW: http://www.virtuallystrange.net/ufo/updates/2003/nov/m04-005.shtml


Next: Trick or truth?

Back to: Trindade Island index

Back to My skeptical opinion about UFOs