Some technical notes on the Grand Finale of HBO's "The Jinx," from the perspective of a sound recordist.

Warning: spoiler alert. You may wish to avoid reading this if you live in a cave and haven't yet seen The Jinx.

Andrew Jarecki's riveting series for HBO, The Jinx, seems at present to be well on the way to fulfilling the ultimate goal of investigative journalism, the righting of wrongs. His subject, Robert Durst, the millionaire scion of Manhattan Real Estate royalty, was a free man when Jarecki began his research into three separate murders that Durst may or may not have been involved with. Today, just a few days after the  airing of the sixth and final episode, Durst is behind bars, accused of first degree murder. Most of us working in the documentary arena dream of moments like this, when the ethereal work of filmmaking has an immediate and real-world impact. But instead of reveling in the moment, Jarecki and his collaborator Marc Smerling have gone silent, invoking, like many a corporate entity of recent decades, the notion that it would be imprudent of them to comment because they are likely to have to testify in court. “Given that we are likely to be called as witnesses in any case law enforcement may decide to bring against Robert Durst, it is not appropriate for us to comment further on these pending matters. We can confirm that evidence (including the envelope and the washroom recording) was turned over to authorities months ago,” they announced, in a statement. Although it would seem to me that another option would be to tell the truth now, and then tell the truth again later, in court, the filmmakers have cancelled all further interviews and press appearances. 

The press hate to be shut down, and journalists hate little more than having their interview cancelled, but the backlash against the creators of the The Jinx has nonetheless been extraordinary. From Gawker to the Guardian, my entirely unscientific analysis is that as much as half of the press coverage has more to do with when Jarecki learned what and what he did with the knowledge than it does with Durst's guilt or innocence. The snark also began before the filmmakers cancelled all appearances, with questions about the "timeline." Because the final interview with Robert Durst occurred "a couple of years" after the earlier ones of 2010, Jarecki is accused of having sat on some extraordinary audio, presumably recorded in 2012, in which Durst appears to confess all. For as long as three years. While Durst roamed the streets. Jarecki's response has been to say that "many months" passed between the filming of that last interview and the discovery of that audio. How can such a thing have come to pass?

Durst and Jarecki, in a screengrab from HBO Go.

There are two potentially inculpatory bombshells revealed in the final episode of what, in stepping back and taking a retrospective overview, is some ostensibly documentary television that is extraordinarily difficult to distinguish from the mood, feeling and structure of a fictional scripted narrative crime drama. The first is the comparison of two hand-written envelopes addressed to different addresses in Beverly Hills. Both writings are in block capitals. Both misspell "Beverly" as "Beverley." Both appear to have been written by the same hand. One envelope once held a letter from Durst to his since-murdered friend Susan Berman; the other told police where to find Berman's corpse. Durst admits he wrote the first, and denies writing the second. Then, at the very end of the episode, apparently at the very end of the interview, Durst goes to the bathroom, where he is recorded muttering to himself apparent admissions of guilt. It is the ethics and "correlation to reality" of this section that most interests me from a technical standpoint as a sound recordist. 

Jarecki holds a photograph, or photocopy, of the "two Beverleys"

The "timeline" questions about the Jinx are important, but they essentially boil down to wondering when Jarecki and Smerling told the police what they knew. In other words, despite believing Durst to be dangerous, did they privilege the schedule for the release of the series over the safety of the public? Did they, ethically or not, postpone the inevitable moment when their relationship with Durst was going to change from one of friendship, collaboration and mutual exploitation into one of antagonism and betrayal? The jilted LA Times' Questions we'd like to ask The Jinx's Andrew Jarecki and Marc Smerling is typical of the quite numerous articles that question the filmmakers' priorities. It is literally a list of questions, much harder questions than I think Meredith Blake would have asked if the LA Times had actually been able to conduct an interview.

Durst's numerous nervous tics include belching, blinking and yawning as if bored. Note the lapel mic, prominently visible at the top center of Durst's shirt.

Handwriting experts will battle over the "Beverley" envelopes in court, but to understand what happens in the final two minutes of The Jinx, it may be necessary, for once, to call to the stand experts in the recording of documentary film audio. How is it recorded? What equipment is used? When are those recordings listened to? How are those recordings used? How do they actually get into the filmmaker's workflow?

A quite murky and poorly-defined extra-wide shot that will not be earning anybody any cinematography Oscars. Durst faces Jarecki in the lower right-hand corner. Note the microphone pointed towards them from the ceiling. I would estimate it to be a good six or seven feet from the subjects.

These final two minutes are presented as having happened at the very end of Jarecki's very last interview with Durst. After the scene in which Jarecki confronts Durst with the two envelopes, we cut to a high-angle wide shot, in which much film equipment is visible; the cameraman, who we have just seen passing behind Durst's head, back on the other side of the table (the edit leaving insufficient time for him to retake his place, incidentally); a boom microphone on a stand, very high up, almost against the ceiling; a lamp, also clamped up high. In this shot, Jarecki is holding a photograph that is not visible on the table in the shots that make up the sequence of the envelope confrontation. It appears to be a photograph of a couple. It is certainly not the image of the two "Beverley" texts that he was holding two shots prior. In a feature film this would be a continuity flaw, a failure by the script supervisor. In a documentary, it slides by. Film editing is all about collapsing time in a coherent way, and this sequence clearly shows that some unspecified amount of time has passed between the envelope conversation and the wide shot, in which the two men thank one another and say goodbye. (This sequence analysis also answers one of Meredith Blake's LA Times questions; "Did the interview end immediately after you confronted Durst with the handwriting samples, as depicted in The Jinx?" Meredith, the answer is "no.")

This camera and its position are themselves odd. The images from it are much blurrier and darker than those from the "primary" and "secondary" cameras used to construct the envelope sequence. It is a sort of "record" camera, in that it sees as much of the room as possible, without any artistic intention. In my entirely speculative estimation, it was likely a small "Go-Pro" style camera placed in a high corner, perhaps just on the off chance that Durst might lose control of himself and attack Jarecki. Even as a rarely used third angle it is a bizarre shot, given that it sees the appurtenances of filmmaking. The already extremely wide position of the microphone on the boom indicates to me that it has been installed in such a way as to be invisible to even the widest possible shots from the two cameras set up across the table from Durst and Jarecki.

 At the end of this shot, Jarecki appears to remove his own microphone.

What happens next is also quite odd, cinematically. The wide shot holds as the two men shake hands and get up. Durst walks forward, around Jarecki, and out the bottom of the frame as Jarecki asks for someone to locate "Bob's" bag. Someone else offers Durst a sandwich-to-go, which could be taken as a rather nasty inside joke, given that Durst was once arrested in Pennsylvania for shoplifting one. At the end of this shot, Jarecki can be seen removing his own wireless transmitter. We then jump-cut to another shot from this same camera, with Jarecki no longer in frame, and a couple of presumed technicians, who we have not seen before, suddenly visible, coming around the far end of the table. Is one of them the sound recordist?

From the moment of this odd jump cut there is only an illusory synchronicity between the sound and the picture. Once we see the two technicians, what we primarily hear are the private murmurings of Robert Durst, no longer visible. I would argue that the jump cut here, from one, synchronous moment in this strange wide shot, to another, with Durst's voice coming from some unspecified area off camera, is used precisely to create the illusion that Durst is speaking at the same time as we see and hear a light being switched off in the interview room. It is important to note that we hear the light being switched off, meaning that in constructing the scene, the editors took audio both from some source within the interview room, perhaps the overhead boom, and also from Durst's lavalier microphone now out in the hall, as he looks for the bathroom. The two have very consciously been mixed together, presumably by Coll Anderson, who has the audio post credit on the series. The sound of the lamp being switched off reinforces the idea, for which there is no actual basis, that Durst's monologue and the shutting down of the set are happening simultaneously. We then see an intertitle card:

You are now going to accuse me of pedantry, of harping on semantics, but this is a strange card. Microphones do not record. Tape recorders record. Actually, in the present day film business, a combination of software, hard-drives and compact flash memory record. The filmmakers have thought very carefully about what images they are going to put in what order in the telling of their story; I'm going to assume they also thought carefully about how to phrase the crucial intertitle that introduces Robert Durst's earth-shattering, allegedly self-incriminating statements. Let's imagine for a moment that we are sitting in on a meeting in which this phrasing is being discussed:

As a sound recordist I might raise my opinion that sound recordists get insufficient credit for supervising the most important half of the image-sound combination that make up the modern film. Why not make up for that in this intertitle? How about "The sound-person continues to record Bob while he is in the bathroom."? 

No, someone would point out, that reinforces our ongoing intrusion during his going to the toilet. It gives us agency.

I should mention that approximately 90% of the interview subjects I have wired in my more-than-twenty-year sound recording career forget within minutes that they are wearing a microphone. Only crack professionals who are interviewed almost daily sometimes remember. Very often it is only when they go into the bathroom and find that there is a small wire leading from their belt or pants pocket up to their shirtfront that they realize someone may be listening to their every fart. It is quite common for interview subjects to then come see me and ask me to turn the transmitter off before they return to the toilet. On a recent shoot with former president Bill Clinton I was informed by his staffers that I was not to use a lapel mic, presumably because of the risk of recording an unauthorized or unguarded moment before or after the official interview. Obama, Bush, and Ronald "we begin bombing in five minutes" Reagan are only a few of those who have been burned by so-called hot mics. Jarecki told Charlie Rose that Durst knew that his microphone was always on, but knowing this and remembering it are two entirely different things.

How about: "The tape is still rolling as Bob goes into the bathroom?" No. First of all there's no tape these days, and we can't exactly say "The hard-drive is still rolling...."

OK, then, what about "The audio is still being recorded as....?" I suppose that might work, but it would be nice to avoid using the passive voice.

The obliviousness of the New York Times and the Washington Post with regard to my profession should embarrass both of those venerated publications. Here's The Post's absurd sentence about the bathroom confessions: "The camera crew had already packed up from the day's interviewing but the recorder kept rolling as Durst went to the bathroom." For their part, demonstrating a breathtakingly gullible interpretation of how this film was edited, the NYT writes: "Near the documentary's end, the filmmakers were packing up their equipment when Mr. Durst asked to use the bathroom. He did not remove his wireless microphone as he closed the door, however, and began to whisper to himself."

This is already dragging on, but perhaps I'm getting ahead of myself. Most people who aren't in the film business don't have a firm grasp of what it is that a sound recordist does. I'm often asked if people hire me to come along and record the music that will be used in the film, or the sound of the birds in the trees or the passing traffic. In fact, these are usually the things I am trying to avoid recording, in the context of an interview, or a scene with several people talking to one another. Viewers don't understand that in order for a speaking human being to be heard crisply and with clarity on film, that person, in the vast majority of contemporary settings, needs to have a microphone very close to their mouth. Let's say, ideally, within 18 to 24 inches. Someone needs to put that microphone there, and the audio coming from it needs to be monitored at all times. (For this I charge approximately $90 per hour, if you are looking).

Back in the long-distant days of film, sound recordists were blissfully independent. We attached our microphones to our tape recorders with wires, and our only worry about the camera was whether it would see our boom poking down into the top of the frame, or our lavalier microphone peeking out from behind a shirt placket. We recorded the sound, and the camera recorded the image. In post-production, the two were synchronized. In this scenario, it was important for sound to begin rolling before the camera, and to continue to roll, if possible, after the camera cut.

Throughout much of the 1990s and the "aughties," as film became too expensive and video became more sophisticated, most audio was recorded on the camera. As a sound recordist my microphones were now attached to my portable mixer, with the camera doing the actual recording, whenever the cameraman was "rolling". I monitored the sound, but did not myself have control over the recorder. I would either have a couple of wires leading from my mixer to the camera, or transmit the audio to the camera wirelessly. Cameramen generally hated this arrangement, because the wires risked impeding their movement, and the "wireless link" option demanded that they do some level of monitoring to be certain the sound was coming through okay. In this scenario, which prevailed in television from about 1990 until at least 2005, and persists on many lower-budget productions, essentially zero audio was ever recorded when the camera was not rolling. Durst's mumblings would never have gone down on tape.

In a sense, we have now gone back in time. Because of the widespread adoption of portable multi-track hard-disc recorders like the Sound Devices 788T, and thanks to the proliferation of Digital SLR cameras, which shoot high-quality video but sound something like a broken walkman, sound recordists today, and during the period The Jinx was filmed, once again generally record the sound independent of the image. Although the video camera visible in the final shots of The Jinx is capable of recording decent sound, the use of a multi-track machine allows the recording of a large number of discreet audio tracks, offering more flexibility in the final mixing of the film.

The credited sound recordists on The Jinx, neither of whom I know, are Tim Hayes and Paul Marshall. I suspect that either of them would have been recording on some kind of multi-track machine. We could ask them how it all went down, but I am certain they will have signed deal memorandums ensuring not only their salaries but their silence. I expect them to be subpoenaed.

There are three microphones visible in the scene: the boom, the lavalier on Durst, and a lavalier on Jarecki. Each of these would be recorded on its own track. In order to aid the synchronization of sound and picture, the sound recordist would then send either one or two tracks of "scratch" audio to be recorded on the camera. On lazier or lower-budget productions, or quite commonly still on network television productions, the editors might well take the sound from the camera tracks rather than bother synching up the files from the hard-disc to the camera images. In this case I would only believe that this had happened if the recordist told me that he had sent Jarecki's and Durst's microphone signals to separate channels on the camera. The boom is so far away from the scene that it can only be there to record room ambience or to provide marginal backup audio in the case of a catastrophic failure of one or other of the lavalier microphones. It is quite possible, although lazy filmmaking, for the editors to have listened only to the audio recorded alongside the camera images, and to have ignored any audio recorded after the cameras had cut. Nonetheless, this recordist would have immediately alerted the producers to the bombshell that had just come into my headphones.

It is important to understand something about the transmitter packs to which these lavaliers are attached. I can see Jarecki's in the near final shot, but I cannot determine its make or model. Often just called "wireless," these body packs transmit the signal from the subject's lapel mic, or lavalier, to the sound recordist. Because of FCC regulations on their transmitting power, they typically have an out-of-doors "line-of-sight" range of approximately 100 yards. In dense urban environments with lots of other wifi activity their ranges can be significantly diminished. Walls also have an impact.

In my sound recording setup, I use Lectrosonics wireless systems. A receiver for each wireless is attached to the recorder, and I record the signal only when I choose, that is to say, when I press record. At the obvious end of the shoot, or interview, I will generally cut (press "stop"), if it is abundantly clear that we are finished. In an exceptional circumstance, and recording a final interview with Robert Durst probably qualifies, I won't cut at all until the director makes it very clear that he considers the filming complete. But in that case I will be sitting in front of my rig, with my headphones on, listening to what I am recording. Jarecki, or at least the media's interpretation of what he has told them, would have me believe that whoever recorded the Durst interview walked away, or started packing up his gear, while still recording. This makes no sense to me whatsoever.

If we accept that the Durst confession actually did happen after the close of the interview, rather than at some other time, it is clear that Jarecki thought the interview had concluded. How else do we account for him removing his own microphone? He can't have known that Durst would hang around long enough to use the toilet. The sound recordist could be excused for cutting, and even for beginning to wrap up his cables. Had I been the sound recordist, I would probably have been attempting to get the okay from a producer to remove Durst's microphone, in order to prevent him from walking out of the building with it. Durst's bag, after all, is being gotten for him, and he has left the room. If we accept the film's chronology, the lights have been turned off. The shot of the darkened, empty room then holds during Durst's entire lengthy bathroom monologue. This is another oddity; it even seems manufactured. I have never been on a film set where the film lights were turned off without the "house" or overhead room lights being promptly turned on. When the film lights are turned off, wrapping is about to commence, and wrapping is not done in the dark. It makes me suspicious. There is something slippery here. When we hear Durst finally say "killed them all, of course," the room dramatically darkens even more, as if the filmmakers flicked off a couple more switches out of frame, and have left the building, with all their gear still installed.

A dark and empty room full of film equipment.

There is one possibility that alleviates, but does not entirely remove my skepticism about how and when this audio was recorded, listened to, and noted. It also could account for the strange phrasing of "Bob's microphone continues to record...." The sound-person might have use Zaxcom wirelesses. These differ from Lectrosonics in that their transmitters simultaneously transmit audio to a wireless receiver and record it locally. They make a time-stamped backup of all the audio fed into them by a microphone for as long as they are switched on, regardless of whether or not the sound recordist is listening or recording. The idea is that if the subject wearing one walks, runs or drives out of range, it is still possible to recuperate their audio. I've heard they are useful for extreme sports.

I have a question for Meredith Blake to add to her list. I would like to know what brand of wireless Durst was wearing, and, if it was a Zaxcom, how anyone ever came to listen to the files that were recorded on it, given that they would be imagined to serve only as emergency backups of the production audio recordings on the hard-drive? If it wasn't a Zaxcom, the whole scenario is very fishy indeed.


Unknown said...

This is an amazing and damning analysis Richard. Another thing is that if it was a Zaxcom mic/transmitter and it has a time stamp, as you point out, then that would clarify when the recording was made. There seem to be more and more examples of questionable narrative constructions in documentary, that cross ethical lines. I'm reminded of Catfish, which raised a whole set of questions of its own. I imagine that if the material is subpoenaed a lot will be clarified, either by the material itself or sworn testimonies of the sound recordists and editors.

They say it's a cold world said...

Hey Paco, thanks for commenting. My link to this over at facebook seems to have generated quite a number of interesting comments, and I'm thinking of copy and pasting them over here. It is annoying that all facebook discussions rapidly fade into history without really maintaining a connection to the links they refer to.
I haven't used the Zaxcom systems much myself, but from my understanding they are time code jammable. I don't know if they would have a user bit field in the metadata where they would stamp the date, but I would imagine they do. Of course both date and time would have to have been jammed into the unit by the sound recordist, but certainly the capability for those files to be carrying the exact date and time of their creation exists, assuming a Zaxcom was indeed in use. Although I certainly appreciate the high praise of your phrase "damning analysis" I do want to be clear that I thought the Jinx was great, both as entertainment and as cinema and as an oblique view into the mechanisms by which the rich move through the justice system in a very different way than others, especially people of colored others. I really wrote this because the discussion around the film seemed to ignore the very existence of sound recordists, at the same time that it was totally clear to me that the audio so key to this episode and in fact the denouement of the entire series only was captured because of a sound recordist!

John V. Fleming said...

The Smoking Microphone! This is a brilliant essay, and so deft in its technical explanations that even I can understand it. Possible conclusions go in different directions, but several of them are quite disturbing. I don't think that the documentary makers are going to be the only ones called to testify on one side or the other. Get ready for a new role: Expert Witness.

Unknown said...


I loved your article and do feel your pain about the shameful way sound recordists are neglected by the general public at large. Also, I do share your suspicions that the filmmakers may have been more than a little disingenuous in their explanations of what happened, or truthful about the "timeline" in which it happened.

Still, as a filmmaker and editor myself, I might suggest another scenario.

As you, have noted, the badly–framed, “high and wide” angle was most probably filmed with a GoPro. Let’s assume that was the case. Typically, a GoPro (especially one mounted to the ceiling) would have been switched on to “record” with a remote or with the GoPro IPhone app. Because, the data rate at which the camera records is pretty frugal, typically, one would turn it on and let it roll for hours. Now, it is possible to feed the GoPro with the “scratch mix” and so, it is possible that is the source of the so-called “confessional” audio. The recordist (as you point out yourself) might have been waiting outside the bathroom to retrieve the wireless microphone before Durst left the building and thus would not have been monitoring that recording.

Moving forward to the edit, the audio on the GoPro, in all likelihood, would have never been listened to by the editor. I know, I would not have listened to it ever, and I have been editing for over 30 years. (Maybe this also explains why I have never been able to convict anybody in any of my films, but I suppose that discussion of my shortcomings will have to wait for another article.) Typically, an editor would have used either the mixed audio from the scratch mix, or had it been provided, the two split channels from the wireless microphones routed to the camera and left it at that. Perhaps you would have left further audio work to the sound-editing department?

So, it is entirely possible that an assistant editor, or some other sound dept. person, “discovered” this missing audio months later while scanning through the footage, and the direction of the film changed dramatically after that. (Technically, the audio might have been made visible in the waveform display on the audio track and been discovered that way.)

Still, as a matter of cutting style, I would have referenced this wide shot earlier in the sequence so as to reveal its existence to the audience and thus foreshadowed the big payoff. I might have helped their case had they done the same.

Nevertheless, I still feel that something is fishy about Jareki’s explanation of the facts to the press. Indeed, the entire build-up and faux-rehearsal of the “ final interview” that is presented in the final episode of the Jinx rings hollow and, to my view, does not fully stand up to the smell test. It’s too bad because it throws doubt on the entire series and places this work firmly in the camp of docutainment that HBO seems to have embraced with a passion. We were all entertained for a few hours and let’s face it, none of us really liked, or rooted for Durst anyway. Who cares (except for the dead) if any of it was true?

They say it's a cold world said...

Ernie--Great to hear from you. I think I can safely say that in terms of thoughtfulness and time invested, yours is amongst the top comments ever posted to this blog in its roughly 9 years of existence. I'm intrigued, because you are offering another quite plausible explanation of how that audio, unmonitored, could nonetheless have been recorded. I'm not sure I buy it, but I like it. This may be outing MYSELF as a lazy sound recordist, but I think given the bizarre angle, with its implicit high-up-in-the-ceiling corner mount, and given the fact that I would already have been feeding audio to two cameras on the ground I think I would probably have said something like "you are never going to use this shot unless Durst leaps out of his chair and starts knocking over the other cameras, and you are never going to use the audio under any circumstances, and there isn't room to strap an IFB to your rig up there in the corner, so I'm going to let this camera be mute." I will be surprised if we learn that there was audio fed to the Go-Pro.
There's another thing, which is that personally, I'm unlikely to pull a wireless off of someone while they are potted up on the mixer or recorder. Generally I am going to pot down the signal from the receiver, and also turn the receiver off, before I ever get to the "talent" to remove their transmitter. It doesn't have to happen that way, but best practice in the sound world is definitely to turn all your input signals down before you start unplugging mics, etc.
I'm 100% with you on the concerns about the docutainment!
And, let's get together before another fifteen years go by!

Unknown said...


Nice to hear from you as well. If I had your e-mail, I would like to send you the trailer to my newest film.

As far as the Audio to GoPro, Yes, I agree that sending audio to the unit would be a real pain in the ass. Still, I always like to have audio -- any audio -- on every source, so at the very minimum, I can check the sync against the master. In this case, I would make sure the microphone on the GoPro was on, and hope that I would get enough level so at least I could make out the dialogue being spoken. Having a guide track of any sort makes syncing so much easier and on some edit systems, can make it automatic. (Plural Eyes on FCP7 or built-in syncing on FCPX)

As far as the the shot in question, if you go back and look at the action taking place through the window on the exterior of the building, It looks like it's not at sync sound speed but rather a weird surveillance camera speed. -- Strange!

They say it's a cold world said...

Hi Ernie--I think I have your email, and just sent you one. What you say about the Go-Pro is interesting. I also took a look when I was writing the post. The final shoot being essentially an empty lock-off, I wondered if there was any possibility it had been looped, since the filmmakers might not have known they would need a nice long chunk at the time they shot it. (That issue is probably obviated by your earlier comment, that the Go Pro would have been cut and rolled via remote control, iphone or whatever). To me it looks like reflected traffic smearing in the glass. But if it is indeed off-speed footage, and it was shot that way, that's a further reason I would have been unlikely to put any of my sound down on that camera....

They say it's a cold world said...

Hmmmn. No. On that last comment I spiralled into tech geekdom. The "Go Pro" footage where we see Jarecki and Durst is on-speed, so if the static shot is off-speed, that must have been done in post, for whatever reason. That would be a strange choice, certainly.

Scott B. said...

Great analysis Richard. I didn't see The Jinx, but have worked with Producer Marc Smerling & His Prod. Co. on countless Commercials and Music Videos in the past. I heard about The Jinx from an LA Times article written the day after the show aired which lauded The Jinx Prod. for their journalistic prowess and then went on to taut the commercial viability of documentary filmmaking.
As a 25+ year veteran film technician, I know the sequence of events at wrap and on many occasions We've waited for the Sound Recordist to complete their jobs prior to wrapping the equip. Once We get the "go ahead" from Sound & Prod., We restore the "House Lights", turn off the film lights and wrap out the location. This is generally the rule regardless of the scale of the Prod. The sequence of events You've laid out definitely do not pass the smell test from a Gaffer's perspective as well. It will be interesting to see how this plays out.