R v Brease[2013] QCA 249

SUPREME COURT OF QUEENSLAND

[1] MARGARET McMURDO P:  The appellant was convicted after a ten day jury trial in the Brisbane Supreme Court on 22 May 2012 of murdering Stephen John Dickson at Brisbane on or about 7 October 2009.  He has appealed against his conviction on two grounds.  The first is that the trial judge erred in admitting evidence which the appellant describes as circumstantial evidence of identification.  The second is that further expert evidence received from Emeritus Professor John Trinder on 30 August 2012, which was unavailable at trial, demonstrates that to allow the appellant's convictions to stand would result in a miscarriage of justice.  In respect of the latter ground, the appellant has applied for leave to adduce evidence in this appeal from Professor Trinder and a certified surveyor who assisted the professor in preparing his reports, Brian Edward Donnelly.

[2] Before discussing those grounds and the application to adduce further evidence, it is necessary to have an understanding of the evidence at trial.

Evidence at trial

[3] A cleaner found the deceased's body at about 5.00 am on Monday, 5 October 2009 near a pillar underneath a building at 100 Wickham Street, Fortitude Valley, where he had been sleeping overnight.  That building and the neighbouring building at 108 Wickham Street are bounded by Gotha Street, Barry Parade, Warren Street and Wickham Street.  The deceased's body was found on the Barry Parade side of 100 Wickham Street, nearest Gotha Street.  He had suffered head injuries.  When paramedics arrived, the deceased was unconscious but breathing.  He later died in hospital as a result of his head injuries, the most serious of which was a large star-shaped laceration to the left of his parietal occipital scalp.  The injuries were consistent with being caused by a long, thin object such as a baton or a Mag-lite torch used with severe force.

[4] There were no eye witnesses and CCTV footage did not capture the lethal assault.  Police found a bag strap with a clip broken off at the crime scene.  The bag was never found.[1]

[5] David McEvoy, a forensic recording video analyst with Queensland Police went to 108 and 100 Wickham Street.  The security cameras at 108 Wickham Street were not operating from 11.58 am on Sunday 4 October until 7.52 am on Monday 5 October 2009.  The security cameras at 100 Wickham Street were operating during that period.  He assisted in downloading the footage from the camera of 100 Wickham Street for that period.  He also assisted in downloading CCTV footage from about 60 Brisbane City Council Citysafe cameras in the surrounding area.  He purchased hard drives on which he downloaded this footage.

[6] One camera overlooked the crime scene, but as a result of the positioning of a light, nothing could be seen on the footage.  Another camera was positioned at the Barry Parade end of the building, directed towards the crime scene although it did not capture the actual position of the deceased.  Police identified a suspect from the footage taken from the security cameras at 100 Wickham Street.  A DVD prepared from the footage from the camera contained four significant pieces of footage.  The footage was tendered[2] and contained an "i-Watch" icon.  The first two depicted a man in a dress shirt entering the crime scene at 3.43 am and then leaving at 3.44 am.  The second two appeared to depict the same man in a t-shirt at 3.54 am.  He looked around before removing a long cylindrical object from his trousers and re-entering the crime scene.  At 3.57 am he is shown leaving the crime scene carrying a bag.  In all instances the suspect appears to be wearing shoes.  The footage established that no other person was seen entering the crime scene after 1.10 am, when earlier footage showed the deceased arrive, until 3.43 am.

[7] Police officer Gutteridge examined the Citysafe footage from a camera at the intersection of Duncan and Ann Streets.  He found images of a person of a similar appearance to the suspect walking along the Chinatown Mall side of Ann Street in the direction of Brisbane City at 3.19 am.  Other footage showed this person leaving The Beat nightclub at about 2.45 am in the company of a security guard, later identified as Alexander McInerney.  Mr McInerney identified the appellant as the person depicted leaving The Beat.

[8] Alexander McInerney gave evidence that he was a friend of the appellant and helped him obtain a position as a crowd controller at The Beat.  Shortly after the police interviewed him, he asked the appellant as they walked along a road near Mr McInerney's house, what was going on.  The appellant said "he had done a bad thing and he had murdered somebody, but it was a bum and he had the whole rest of his life to live."[3]  On another night, together with Neil Hatfield who did not give evidence, they talked about the episode when they were all out drinking.  Mr McInerney could not remember the exact words, but the appellant said that "he did do it."[4]  In cross-examination, Mr McInerney denied that he had made up this confession to avoid being charged with supplying dangerous drugs and the resulting loss of his security provider licence.  He agreed he had been found guilty in the past for supplying cannabis.  He was confused about the dates on which the appellant made these confessions.

[9] Benjamin Raus was working as a security officer in October 2009 at Cloudland nightclub in the Valley.  He had then known the appellant, who was working at the Empire nightclub, for about four or five months.  They were friends who spent time together on weekends and shared boys' nights.  Following a police search of Mr Raus's house on 3 December 2009, he went with them to the Valley police station.  He was shown the footage depicting a man in a dark dress shirt walking in and out of the crime scene.  He recognised this person as the appellant "just by his stature and just like the way he walked, he walks sort of like with a sort of slunch (sic) and just his hairline on the CCTV footage."[5]  He agreed the footage did not clearly depict the man's face.

[10] A couple of weeks after the police interview, over breakfast in the Brunswick Street Mall, he and Mr McInerney confronted the appellant about why police were pulling everyone in for questioning.  The appellant said not to "worry about it any more, that he's going to handle it and the police … shouldn't bother anyone any more."[6]  Maybe a month or a couple of weeks later, police were still calling in others, including Mr McInerney, for questioning.  Mr Raus telephoned the appellant from his house.  Mr McInerney was present and they asked him, "What's the go?  You said it was all over and down [sic] with.  You said you had it handled.  Why are we being called in?"[7]  The appellant responded:

"… that he was walking after a night out, walking down the road and he seen a guy laying down next to a bag and he said that he went back to his car and found a torch in his car and went back to the bag and the male and took the bag. … He said that he struck the male with the torch and took the bag. … He said that he found an amount of money inside the bag, the amount I'm not too sure about, possibly around $3.50, $3.80 mark sort of thing."[8]

[11] He did not speak to the appellant again.  In cross-examination he agreed that, when the police showed him the footage of the man entering the crime scene, he said, "I would say I'm about 65 to 75 per cent sure it was [the appellant] by his walk."[9]  He was not certain that it was the appellant.  When the police raided his house on 3 December 2009, he was concerned that they found an ecstasy tablet and a stolen phone in his room.  He had since "dealt with that matter with the police and taken the correct actions".[10]  He denied that he made up the confessions to avoid trouble with the police.

[12] Police found a blanket at the crime scene and took tape lifts from it.  The DNA analysis of two of the tape lifts revealed mixed DNA profiles which resolved into major and minor DNA profiles.  Each major profile matched the deceased's DNA.  Each minor profile, which was partial only, matched the appellant's DNA.  One sample giving a minor DNA profile was analysed twice.  On the first analysis, the probability of it coming from someone other than the appellant was found to be 1:160,000.  On the second analysis, the probability of it coming from someone other than the appellant was 1:75,000,000.  The DNA profiles matching the appellant's were small in amount and at a low level.  The interpretation of low level DNA in a minor DNA profile requires some caution.  The evidence also established that DNA can be transferred from one person to another, and more than once, and that both the deceased and the appellant spent considerable periods of time in the Valley area.

[13] The prosecution called evidence from those who knew the appellant to the effect that he regularly parked his car in the vicinity of the crime scene.  Of most relevance, Timo Saikkala said that more than 20 times in 2008 and 2009 he travelled with the appellant to the Valley to go clubbing.  The appellant usually parked in Warren Street but if he could not find a park there he would park in Barry Parade.  This area provided free parking at that time of night, was well lit and was an easy walk to the clubs.  When they returned to the car, they would always go along Brunswick Street to Wickham Street, cutting through to Warren Street; they did not go down Ann Street.

[14] Emeritus Professor John Fryer gave evidence.  He was now retired but was formerly Head of the School of Engineering at the University of Newcastle.  He had a Bachelor of Surveying with Honours and a University Medal and completed a PhD before working in national mapping.  As a university lecturer, he developed an expertise in photogrammetry, a specialist area of surveying which is the science of taking measurements off photographs.  He had researched and published extensively in this area.  He lectured in photogrammetry for 30 years and had ascertained measurements from photographic material, CCTV footage and the like for about 40 years.  He had conducted 90 forensic investigations, about 80 per cent for the prosecution and 20 per cent for the defence, involving the determination of heights, widths, lengths of shoes and weapon sizes in respect of 250 persons of interest at about 750 separate CCTV sites.

[15] Queensland police asked him to analyse the height of the suspect in the footage of the crime scene.  He spent some time examining the footage frame by frame.  He extracted still images which would allow him to make measurements.  These images depicted the suspect standing against items or landmarks which he could positively identify at the crime scene.  He took surveying measurements at the crime scene.  He used an extension step ladder to take photographs from the same position as the camera which took the footage.  He placed his camera against and right in front of the CCTV camera lens to replicate what was depicted in the footage.  He calculated the probable height of the suspect in the footage at 3.43 am and 3.44 am as 186 cm and in the range of 184 cm to 194 cm.  He calculated the probable height of the suspect in the footage at 3.54 am and 3.57 am as 187 cm and in the range of 184 cm to 190 cm.  He used two methods of calculation.  First, he used trigonometry to calculate the suspect's height against the actual measurements of items and landmarks at the crime scene.  Second, he made a direct comparison between the suspect in the footage and the photographs he took which depicted a surveying staff at various points in the crime scene.

[16] Professor Fryer's cross-examination was brief.  He described both methods used by him as equally reliable.  The range of his calcualtions was because of the uncertainties in, for example, placing the top of the head of the suspect depicted in the footage at an exact vertical point above where his feet touched the ground.  That is why he allowed 3 cm each side of the mean average height calculated.  This range demonstrated the outer limits of the estimated height, in the first instance from 183 cm to 189 cm and in the second from 184 cm to 190 cm.  Professor Fryer stood by his conclusions which he emphasised were based on accepted mathematical standards, allowing for standard deviations.

[17] A photograph of a police officer measuring the appellant's height without shoes was tendered.[11]  It showed the appellant was 186 cm tall.  The photograph depicted the top of his hair as being about 2 cm above his skull.  The 186 cm measurement was to the top of his skull, not the top of his hair.

[18] Police officer Stirling gave evidence that the time stamp in the footage depicting the suspect was two and a half to three minutes behind actual time.  It followed that the footage depicting the suspect strongly suggested that the fatal assault on the deceased occurred between 3.56 to 4.00 am.  Police officer Stirling drove from the corner of Warren Street and Wickham Terrace to the appellant's Indooroopilly address.  It took five minutes and 20 seconds to travel to Milton Railway Station and 13 minutes and two seconds to reach the Indooroopilly address.  He also timed the walk from the crime scene to the intersection of Warren Street and St Pauls Terrace and back to the crime scene, a period of five minutes and 26 seconds.  Police officer Van Der Poel gave evidence that it took him nine minutes to walk from the crime scene to Hartley Street, pause for 20 seconds, and then walk back.

[19] The appellant owned a white 2003 Ford Falcon sedan.  CCTV footage of the only four white vehicles to pass Milton Railway Station at the relevant time was tendered.[12]  The white car passing the station at 4.13 am was identified as a Ford Falcon built between 2002 and 2005.  Four still images of this car passing the station at 4.13 am were tendered.[13]

[20] Ryan Delarue, a surveyor, became friends with the appellant in 2009 and they shared a unit at Indooroopilly.  On Sunday, 4 October 2009, they went to the local pub to watch the NRL grand final.  They returned home about 7.30 or 8.00 pm.  Before Mr Delarue went to bed, he prepared his work gear to commence a 2.00 am shift.  When he woke shortly before 2.00 am to leave for work, the appellant's car had gone.  About $50 to $80 of the appellant's $240 rent contribution was missing from the pin board where he had left it.  The appellant phoned Mr Delarue at about 4.00 or 4.30 am.  He was "pretty drunk"[14] and wanted Mr Delarue to let him into the unit.  Mr Delarue explained that he was at work and told the appellant how to enter the unit through the back courtyard.  When Mr Delarue returned from work, the appellant's car was in the garage and the appellant was asleep in his room.  He later asked the appellant for the missing cash for the rent and thought he provided it the next day.

[21] Police contacted Mr Delarue on Friday, 13 November 2009 to search the unit.  About a week later, he went to the police station to look at the footage of the suspect.  The following conversation occurred:

"Did you recognise anyone in the footage?--The characteristics of the person in the footage did have similarities to [the appellant].

What similarities?-- His - I guess the way he walked and his height and posture and things like that.

Is there any particular piece of that footage that demonstrated more than others?-- Most notably probably the second one when - the guy coming in with the T-shirt when he stands around for a good portion of time.

All right?-- I think that's the third one.

The man in the T-shirt entering-----?-- Yes."[15]

[22] In cross-examination, Mr Delarue agreed that it was very difficult to make out the suspect in the footage and he could not definitely identify the appellant as the suspect.  He had been living with the appellant at that stage for about seven or eight months and was very familiar with his physical appearance.

Ground 1 – the circumstantial evidence of identification

The appellant's contentions

[23] The appellant contends that the primary judge erred in admitting the evidence from Mr Raus (set out at [9] and [11] of these reasons) and Mr Delarue (set out at [20] and [22] of these reasons) as to the similarities between the appellant and the suspect in the footage.  Relying on the High Court majority's observations in Smith v The Queen,[16] the appellant contended that the opinions of those witnesses as to the similarity of the appellant with the suspect's height, stature or hair line was irrelevant and inadmissible: R v Christie;[17] Harriman v The Queen;[18] Festa v The Queen.[19]  Alternatively, if admissible, the judge should have excluded the evidence in the exercise of her discretion on the ground that it would create a risk of prejudice disproportionate to its rational probative value.  The judge was required to evaluate the weight the jury could rationally attach to the evidence and the extent of the risk that the jury would give the evidence disproportionate weight: Dupas vTheQueen.[20]  In ruling that the impugned evidence was admissible, her Honour did not make clear that she undertook that exercise.  The nature and quality of the footage was such that any opinion based on it as to similarities between the suspect and the appellant was of slight probative value.  It presented such a risk of the jury giving it disproportionate weight, despite the trial judge's warning to the jury, that it should have been excluded.

The trial judge's ruling and jury directions

[24] Her Honour determined that some evidence of the kind which the prosecutor sought to lead was inadmissible.  Her Honour determined, however, that the impugned evidence of Mr Delarue and Mr Raus was admissible as evidence that the appellant's gait and general appearance was similar to that of the suspect depicted in the footage.  It was circumstantial evidence falling within the third category of admissible evidence described by McHugh J in Festa v The Queen.[21]  The evidence was not sought to be adduced as positive identification of the appellant.  Her Honour then referred to evidence from other witnesses which she considered had very little probative value and indicated that she would exclude it on discretionary grounds unless the prosecutor wished to submit to the contrary.  The prosecutor accepted that ruling.[22]

[25] The judge gave careful directions to the jury as to the use to be made of the impugned evidence.  They could not convict the appellant unless they were satisfied beyond reasonable doubt that he was the man who attacked the deceased.  They could not be satisfied of that, her Honour emphasised, unless satisfied beyond reasonable doubt that the suspect in the t-shirt depicted in the footage was the appellant.[23]  There was no direct evidence identifying the appellant as the person who attacked the deceased.  Her Honour reminded the jury of the observations of Mr Delarue and Mr Raus when shown the footage depicting the suspect and directed them to:

"approach this evidence of Delarue and Raus with caution. Before accepting the evidence of either of them you must be satisfied that the witness gave the evidence truthfully and that the evidence is reliable."[24]

[26] She reminded the jury that Mr Raus was questioned by police about an ecstasy tablet and a stolen phone located in his bedroom.  They should consider the quality of the footage depicting the suspect which, they may think, was not particularly clear.  They may think the length of each clip of footage was short.  They should consider the features mentioned by Mr Delarue and Mr Raus and consider whether, for the reasons they stated, there was such a similarity between the suspect and the appellant.  They should consider how well the features upon which Mr Delarue and Mr Raus relied were depicted in the footage and whether those features, either individually or in combination, were unusual.  Her Honour reminded the jury that the impugned evidence was not evidence of identification.  If they accepted Mr Delarue's evidence, it was no more than evidence that the gait, height and posture of the suspect were similar to that of the appellant.  If they accepted Mr Raus's evidence, it was no more than that the gait, stature and hair line of the suspect was similar to that of the appellant.  If they accepted the evidence of either or both, the similarities were pieces of circumstantial evidence to be taken into account and weighed with all the other circumstances the jury found proved in deciding whether they were satisfied beyond reasonable doubt that the appellant killed the deceased.

Conclusion on this ground

[27] The appellant places considerable reliance on Smith[25] but the present case differed from Smith.  Police officers who knew Smith purported to positively identify him as the person depicted in security photographs recorded during the robbery with which Smith was charged.  They therefore purported to identify him as the offender.  The High Court unanimously determined that the police officers were in no better position than anyone else to compare Smith and the person depicted in the photographs as the robber.  The plurality's observations in Smith were made in the context of purported identification by recognition.  The impugned evidence in the present case was not a purported identification but statements from Mr Raus and Mr Delarue about similarities between the image of the suspect in the footage and the appellant who was well known to them.  The plurality[26] identified this distinction, noting that a witness's evidence of identification by recognition of a person depicted in photographs or footage from a crime scene may be relevant and admissible:

"if it is suggested that there is some distinctive feature revealed by the photographs (as, for example, a manner of walking) which would not be apparent to the jury in court, evidence both of that fact and the witness's conclusion of identity…"[27]

[28] Their Honours, however, qualified that observation by noting that other considerations may nevertheless arise warranting the exclusion of such evidence, for example rules relating to opinion evidence and discretionary grounds for exclusion.[28]

[29] The trial judge, consistent with the approach of the High Court in Smith, determined that the impugned evidence was admissible, not as positive identification or recognition evidence, but as circumstantial evidence which the jury might, in combination with other evidence, accept beyond reasonable doubt as demonstrating that the suspect depicted in the footage was the appellant.  See also McHugh J's observations in Festa.[29]  Both Mr Raus and Mr Delarue knew the appellant well and were familiar with his height, posture, gait and hair line at the time. His height, posture and gait would not necessarily be obvious to the jury observing him sitting in the dock, and hair styles often change over time.  Their opinion on these matters was admissible.  That said, trial judges and prosecutors should take care in admitting or calling evidence of this kind where it is not persuasive and of questionable weight.  The impugned evidence, however, is not in that category.

[30] It is true that her Honour did not in terms state that she would exercise her discretion to admit rather than exclude the evidence because of fairness considerations; the appellant did not ask her to do so.  But when her Honour's ruling is read as a whole I am satisfied the judge did not consider it was unfair to admit the evidence.  After ruling the impugned evidence was admissible, her Honour immediately indicated that she would be exercising her discretion to exclude other evidence of this kind on the basis that it had little probative value.[30]  I would infer from the fact that the judge did not exclude the impugned evidence on discretionary grounds that she did not consider it should be excluded on that basis.  That would have been an unassailable exercise of the discretion.  In any case, as her Honour was not asked to exclude the evidence on discretionary grounds, there has been no error of law.  Her Honour carefully and accurately warned the jury of the limited use to be made of the evidence and the possible weaknesses surrounding it.  There is no reason to consider the jury did not act on those directions.  The appellant has not demonstrated that a miscarriage of justice has arisen under s 668E(1) Criminal Code.

[31] For these reasons, this ground of appeal is not made out.

Ground 2 and the application to adduce fresh evidence

[32] The appellant's second contention is that fresh evidence adduced at the appeal demonstrates that the appellant's conviction amounts to a miscarriage of justice so that the appeal must be allowed, the convictions set aside and a new trial ordered.  A consideration of these contentions first requires a review of the further evidence.

The further evidence

[33] Professor Trinder's affidavit filed 9 April 2013 exhibited his curriculum vitae.  He holds a Bachelor of Surveying and a Master of Science and a PhD in photogrammetry. He was President of the International Society for Photogrammetry and Remote Sensing from 2000-2004.  He has had a long and distinguished career at the University of New South Wales where he remains Visiting Emeritus Professor at the School of Surveying and Geospatial Engineering.

[34] He explained that on 3 November 2010, well before the trial, he was instructed by the appellant's lawyers to give his opinion of Professor Fryer's report of 9 March 2010 relating to Professor Fryer's calculation of the height of the suspect in the footage.  Professor Trinder prepared a report on 15 November 2010.  He identified some concerns and was briefed to visit the crime scene and undertake measurements of the suspect in the footage.

[35] After doing so, he prepared a second report on 11 January 2011, still well before the appellant's trial in May 2012.  This second report included the following.  He extracted multiple images of the suspect from a copy of the footage in which the suspect's head and feet were both visible.  By comparing those images to 169 control points identifiable on the images from the footage and at the crime scene itself, and measuring the size of actual items depicted in the footage, he determined the suspect's height on 19 images.  From one piece of footage, he estimated the suspect's height at 1.898 metres.  From another piece of footage, he estimated the height at 1.901 metres.  From another piece of footage, he estimated the height at 1.906 metres.  The overall mean height of the suspect was therefore 1.915 metres.  These calculations established that 68 per cent of observations placed the suspect's height in the range of 1.895 metres to 1.935 metres and 96 per cent of the observations placed the suspect's height in the range of 1.875 metres to 1.955 metres.

[36] He ensured the accuracy of these conclusions by identifying 28 control points on the extracted images and by determining the exterior orientation (position and tilts) of the camera which took the footage using version 5 of Photomodeler software.  This mitigated against distortions in the geometry of the images typically caused by the camera lens, the electronic imaging processes and any subsequent processing.

[37] Professor Trinder opined:

"Having visited the site, I am convinced that [Professor Fryer's] methods can lead to significant inaccuracies in the determination of the height of the person.  Given the height of the camera above the ground, the scale variations at the locations of objects used by him as references to determine the height of the male person can vary significantly.  A small error in locating the reference objects will lead to significant error in the determination of the person's height.  As well, the scale of the reference objects varies continuously along their length.  Professor Fryer's determinations of the height of the male person are 1.86m and 1.87m.  These figures are outside the range 1.875m to 1.955m, which is two times the standard deviation from the mean of our determinations of the person's height, and therefore they are statistically significantly different from the determinations from this report."[31]

[38] Some time later, when Professor Trinder was working on another case, he became aware that the geometry of pixels in footage such as in ex 22 is often distorted by the video post-processing software.  It is necessary to compensate for this distortion with custom-developed software.  The footage in the present case was processed using i-Watch software which significantly changes the geometry of the image, resulting in the pixels being rectangular rather than square.  The i-Watch software can vary the size of images with some being 992x812 pixels and others being 704x576 pixels.  He had not used the custom-developed software which largely compensates for these distortions in his second report to the appellant's lawyers.  Professor Trinder then made new calculations using the appropriate software and prepared a further report with the assistance of surveyor, Mr Brian Donnelly, who adapted the software, measured the position of the control points on the images, processed the data based on the measurements and the ground control points and re-determined the suspect's height.  Professor Trinder interpreted the results of those computations and prepared a third report dated 30 August 2012 which he immediately sent to the appellant's lawyers.  This was well after the conclusion of the appellant's trial.

[39] In the third report, Professor Trinder stated that there was no difference in the heights of the suspect in the four pieces of footage he examined.  The suspect's mean height was 1.8 metres with an estimated accuracy expressed as a standard deviation of 0.04 metres.  Sixty-eight per cent of observations placed the suspect in the range of 1.76 metres to 1.84 metres.  Alternatively, using two times the standard deviation, 96 per cent of all observations placed the suspect in the height range of 1.72 metres to 1.88 metres.  These height estimates differ from those in his second report because his estimate now compensated for distortions in the pixel size caused by the post-processing of the video images in the footage with i-Watch software.  By contrast, Professor Fryer estimated the suspect's height as between 1.86 metres and 1.87 metres, measurements at the extreme end of the height range now calculated by Professor Trinder and with a computational estimated probability of less than 10 per cent.

[40] In his evidence in this Court, Professor Trinder opined that the measurement of single images from photographs or stills from footage was the least accurate way to measure, but in this case the only suitable images from which to take measurements were from single images extracted from the footage.  Measuring something by scaling it against other known objects depicted in the image could be done with accuracy only where the suspect's position could be accurately measured with no more than a 5 cm variation.  Apart from one image where the suspect was standing on a mat, it was not possible to do that in this case.  It was far more accurate to measure something in these images from the footage by taking into account perspective.  This requires the calculation of the perspective centre which is inside the camera.  It was not possible to determine the exact position of the camera which took the footage as the camera was enclosed by a box with glass protection in front.  Professor Fryer's placement of a digital camera in front of the camera could only provide an approximation of what the camera originally filmed.  It was not possible to exactly replicate the position of the camera or its tilt.  By contrast, Professor Trinder calculated the perspective centre by using the position and angle of the camera, together with identified points in objects depicted in the images, effectively enclosing the area to be measured with control points within the image.  He then used collinearity equations[32] to calculate the height of the suspect.  His initial calculations for his second pre-trial report failed to take into account that the pixels in the images were rectangular not square.

[41] Professor Trinder explained that he chose a still from the footage which allowed the position of the suspect's feet to be calculated to an estimated accuracy of 2 to 3 cm.  He then determined the coordinates of the suspect's head, assuming that the head was vertically above the point from where his feet were measured.  He made that assumption because he did not have sufficient information to positively determine that the head was vertically over the feet.  Unless the person was "really stooped",[33] he considered that this assumption was accurate to within 1 or 2 cm and that this was the most accurate method of determining the suspect's height from the footage.  This approach was well-established in photogrammetry and had been used for over 100 years.  It involved the use of accurate control points to locate the position of the camera and its tilt, corrected the scale errors in the pixels in the footage and it determined the position of the suspect with adequate accuracy.  He used this process to measure the suspect's height in a number of still images from the footage and then calculated a mean, that is, an average height taken from all observations.  His height estimates included the height of any shoes the suspect was wearing and extended to the very top of his hair.  His calculations did not provide a precise measurement: the images were not good quality; the lighting was extremely bad; and it was difficult to identify the exact location of the suspect's head.

[42] In cross-examination, he agreed that whether the images depicting the suspect were measured by his method or by Professor Fryer's, it was important to identify with precision the point of contact with the suspect's feet on the ground and the precise point of the top of the suspect's head.  He agreed that an error as to the position of the top of the head could result in a changed height calculation or 3 or 4 cm.  Similarly, an error as to where the suspect's feet met the ground could result in an error of 3 or 4 cm.  The accuracy of his calculations as to the suspect's height depended on the assumption that the head was precisely vertically above the feet.  He could not exclude that the suspect may have been slouching or had his knees bent in the images but he avoided using images where the suspect was obviously slouching.

[43] When he prepared his second pre-trial report, he allowed for a 3 cm variation to his estimated height to accommodate for the fact that the suspect was walking but he did not do this in his most recent report.  Had he done so, it would have altered his estimation of the suspect's height by 3 mm.  His calculations produced a range of heights; this was indicative of the difficulties he faced in calculating the height because of the many variables.  He could not exclude that the suspect was 186 or 187 cm tall but this was only a 10 per cent probability.  The suspect could have been a little taller if he was not standing erect in the images.  He emphasised that he chose images which portrayed the suspect upright.  He agreed that he did not know the ordinary posture of the suspect.  The images were particularly problematic around the edges, especially as to where the head finished and whether the feet were on the ground.

[44] Mr Donnelly deposed to his qualifications and explained his role in assisting Professor Trinder in the computation of the findings recorded in his reports.

[45] Ms Lauren Phelps acted for the appellant as his solicitor throughout the committal and trial.  She engaged Professor Trinder prior to the committal hearing to prepare his first report commenting on Professor Fryer's report and his second pre-trial report.  On 23 July 2012, two months after the appellant's conviction, Professor Trinder informed her by email that he had reviewed his computations of the suspect's height in the footage after finding that he had wrongly calculated heights in other cases.  He re-computed the suspect's height in the footage using appropriate software.  She forwarded the email to Mr Bradley Heilbronn of Legal Aid Queensland who had carriage of the appellant's appeal.

[46] Mr Heilbronn deposed that he received Ms Phelps's email on 24 July 2012 and on 23 August 2012 he requested Professor Trinder to provide a further report which he received on 30 August 2012.

The appellant's contentions

[47] The appellant contends that Professor Trinder's evidence is fresh evidence because it did not exist at the time of the appellant's trial and could not have been obtained, despite the diligence of the appellant's lawyers.  It is reasonable to assume that the suspect in the footage was wearing shoes.  The appellant must be taller than 186 cm when wearing shoes.  Professor Trinder's expert evidence that the mean height of the suspect depicted in the footage with shoes is 186 cm means that the suspect cannot be the appellant who is 186 cm without shoes.  Although it is not impossible that the suspect was the appellant's height, it was highly unlikely.  The judge directed the jury that the appellant could not be convicted unless they were satisfied beyond reasonable doubt that the person in the footage was the appellant.  Professor Trinder's fresh evidence made this unlikely as there was a 90 per cent probability that the suspect was not the same height as the appellant.  The unavailability of the fresh evidence at the appellant's trial has caused a miscarriage of justice in that there is a significant possibility that the jury, acting reasonably, would have acquitted him if the fresh evidence had been before it in the trial.

Conclusion on this ground and the application to adduce fresh evidence

[48] Fresh evidence is evidence which either did not exist at the time of the trial or which could not then with reasonable diligence have been discovered.  See Ratten v The Queen;[34] Lawless v The Queen;[35] R v Katsidis; ex parte Attorney-General (Qld)[36] and R v Spina.[37]

[49] Professor Trinder's most recent report and the opinion evidence it contains did not exist at the time of trial.  When he and Mr Donnelly gave the appellant's lawyers their second pre-trial report, they were not aware that the images in the footage were distorted by i-Watch software in the download process from the security cameras.  He did not then hold his present opinion.  They made fresh calculations upon which their most recent report was based only after the trial.  The appellant's lawyers had obtained reports from Professor Trinder well before the trial and, quite reasonably, relied upon those reports at trial.  They were not informed of Professor Trinder's revised opinion using correcting software to calculate the suspect's height from the footage until well after the trial.  I consider that as Professor Trinder's revised opinion evidence did not exist at the time of trial and could not have been discovered by the appellant's lawyers with reasonable diligence, it is fresh evidence.

[50] In determining whether to allow an appeal against conviction based on fresh evidence, this Court must determine whether there is a significant possibility (or that it is likely) that, in light of all the admissible evidence, both the fresh evidence and the evidence at trial, a jury acting reasonably would have acquitted the appellant.  See Gallagher v The Queen;[38] Mickelberg v The Queen[39] and Spina.[40]

[51] There may be cases where precise expert evidence can be given by way of photogrammetry but this is not one.  The evidence of photogrammetry, both at trial and in the further evidence in this appeal, was necessarily inexact because of the imprecision of the variables upon which it was based.  The footage and the stills taken from it upon which the photogrammetry evidence was based were poor quality and the lighting was far from ideal.  Professor Trinder's most recent opinion as to the height of the suspect depicted in the images was dependent on where the image of the head finished; whether the feet were touching the ground; and whether the head was positioned vertically over the feet, that is, the suspect had perfect posture.  The suspect in the footage was wearing some sort of shoes but it is not possible to make any reliable assessment as to the height of the shoes.  There is no evidence as to the height of the shoes the appellant was wearing on the night of the killing.  The appellant's height was measured at 186 cm without shoes.  The photograph of the taking of that measurement[41] shows the measure positioned against the appellant's head and that his hair was about 2 cm above his scalp.  The photograph clearly shows that the measurement is from the top of his scalp, not the top of his hairline.  The appellant's friend, Mr Raus, referred to the appellant's tendency to "slunch".  Whilst this evocative expression is not one known to the English language, its context makes clear that it is intended to convey the two concepts of slouching and hunching.  Mr Delarue also commented on the appellant's posture and way of walking as being similar to that of the suspect in the footage.  If the appellant's posture was not ordinarily perfectly upright, this is another factor throwing doubt on the accuracy of Professor Trinder's fresh evidence of the suspect's height.

[52] These many variables meant that Professor Trinder's fresh evidence was not a precisely accurate mathematical calculation.  But in any case, it provided for a 10 per cent possibility that the suspect's height may have been up to 190 cm.  Even at its highest, the fresh evidence did not clearly demonstrate that the suspect's height was in a range into which the appellant's height did not fall.

[53] In determining whether the fresh evidence demonstrated that, if it had been before the jury, there was a significant possibility the jury would have acquitted the appellant, this Court does not look at the fresh evidence in a vacuum but considers it with the other evidence at trial.

[54] The appellant was in the vicinity where the deceased was killed at about the time of the lethal assault.  DNA consistent with his DNA was found on a blanket at the crime scene.  While he and the deceased were often in the Valley area and DNA can be transferred, there was no evidence of how this may have occurred.  Two friends of the appellant, Mr McInerney and Mr Raus, gave evidence that he confessed to the crime.  There was no contrary evidence that the appellant did not confess to them.  Although their credibility was attacked, it seems implausible that they would fabricate a confession from a friend to a killing merely to avoid relatively minor trouble with the law.  Professor Trinder's evidence was imprecise and even at its highest did not demonstrate that the appellant was not the suspect depicted in the footage.

[55] After reviewing Professor Trinder's fresh evidence, together with the evidence at trial, I am unpersuaded that there is a significant possibility that, had Professor Trinder's fresh evidence been available at trial, a jury acting reasonably would have acquitted him.  It follows that the application to adduce further evidence should be refused and the appellant's second ground of appeal is not made out.

Orders

[56] As the appellant has not made out either ground of appeal, his appeal against conviction should be dismissed.  His application for leave to adduce fresh evidence should be refused.  I would make the following orders:

1.Application to adduce further evidence refused.

2.Appeal against conviction dismissed.

[57] FRASER JA:  I agree with the reasons for judgment of McMurdo P and the orders proposed by her Honour.

[58] PHILIPPIDES J:  I agree for the reasons given by the President, that the application for leave to adduce further evidence should be refused and that the appeal should be dismissed.