Wolf Seeberg - Video Assist Equipment for Sale or Rent
Wolf Seeberg - Video Transmitters

File-based sound-recording
Copyright Wolf Seeberg, 2003

Features recorded on Fostex PD-6 - - magless dailies using PD-40- syncing in Avid - direct sound file transfer to Avid 

Mixer Checklist: 

  • Make sure your Avid version can import the format of file you record on DVD. Older Avids can do only four tracks audio max! Only the very latest  can do 24 bit. You want to be in control of any bit reduction, so plan ahead! For efficient audio import with all metadata and timecode information you need vers 10.5 on of Film Composer. You may be asked for different bit rates for the Avid and for ProTools. 
  • Do a “one line” ( a graphic chart of the media path and all its parameters) of the signal path and media thru post and online or film release. Have everybody sign off on it. 
  • If there is a real smart producer at the helm he would do a “test” thru to on line or final mix to see that everybody is on the same digital page. 
  • You may not need a TC slate. The Avid currently needs sound files with TC though. 
  • The editors need to be technology friendly. There may be a need for an extra Avid. This is NOT a smooth self-evident workflow as of 2003. Many extra phone calls will need to be made to some “technology consultant”. Chris Munro has shepherded more Fostex projects than anybody by summer of 2003.  I don’t know where to point you if you are working with the DEVA recorders. Ed. 
  • Results: Overall there will be a money saving, speedier better dailies, a better final product, better control by soundmixer as well as editor, and faster post all round. 

Chris Munro chris@soundstation.demon.co.uk talks about practical field experiences with file based recording: 

The first thing we need to think about is that not all file based recording principals are the same. Although file based recording does give us a lot of advantages I can only talk really extensively about my experience with the PD-40 and PD-6. One of the things about PD-6 that’s as important as the fact that it’s file based recording is what’s contained in the meta data. We use what’s known as a file per take system so every take that’s shot is on a separate file. The file name of that take is the slate and take number. So let’s say your slate number is “123 Apple Take 1” then that’s what the file name becomes. It is named by you, as you enter this on a keyboard for convenience or the front panel if you are over the shoulder. There’s a fast key on the keyboard to bring up the screen. You enter the scene number. The default is take 1 (you can change that) and from then on every take is automatically incrementally advanced. This is incorporated in the meta data. 

The meta data is imported into the Avid along with the files in the BWF file format. A lot of people are concerned about using SD2 over BWF. Avid has pretty much discontinued SD2 now. BWF is certainly the best format to use. Primarily because of its meta data and timecode capabilities; it allows original location TC to be imported along with the audio. Whereas with certain types of SD2, you import audio into a Pro-Tools with its time stamped with the timecode of when you’ve imported it and not when you originally recorded it. That’s somewhat of an oversimplification but BWF gets over these issues and is a good common file format between the acquisition on a PD-6, the import into the Avid film composer and finally the post onto a Pro-Tools. Meta data is what carries the time code information, the file name information which is the slate & take, but it also has another area that you can complete which is the default track names. Within default track names you can actually label what is on each track. You probably wouldn’t want to do that with every scene or slate that you do, so it remains in default until you change it. It is important that you record a finished mix track, a track used for dailies and the picture editor which should be track one. You’d probably label that track one as mix. Track two can be used for offstage lines, off-lines , because generally a picture editor wants to edit with just one track, but it’s useful for him if he can access off-lines easily and he knows they’re always going to be on track two. You can as a default label track two as off-lines. I’m not sure how many characters you’re allowed on a label. I use short labels, so haven’t run into limitations. 

 The mixer is also able to convert all of this information in to an ALE (Avid Log Exchange ) file if he wishes. This is a simply a text file that can allow you to easily print out sound reports with all of this relevant information. 

When you’re using PD-6 in the six track mode, I tend to use one and two as cutting copy sound, sound for dailies and the picture editor to cut to, and I tend to use the other four tracks for the elements that make up that mix. That’s ideal for most post operations. It means the picture editor has one track he needs to cut to. The editor would find it too time-consuming to cut with more than one or two tracks. There’s a great temptation to record everything on the set and leave everything unmixed, It means that we can use every track to have a separate microphone on and allow someone else to mix it later. I understand how that can give great flexibility but in fact it’s a disadvantage because what has to happen is that generally only one of those tracks is used for the editor to cut with and it may not have all the dialog, or worse still all the tracks are combined together arbitrarily by an assistant. They’re not mixed; they’re just combined into the Avid and the editor uses that as his cutting copy. The problem is that nobody knows what the quality of the audio is like. And because of time constraints what very often happens is the ADR has to be spotted and decided before the dialog editor has had a chance to really go back to all the original tracks. This leaves the dialog editor has no choice but to cover his ass and say ADR anything that is doubtful, or label up a whole load of stuff for ADR, and that’s unfortunate because had you been able to access those tracks you’d find they weren’t necessary to ADR. The benefits of putting a one track mix down means that a lot of time the sound mixer would have made a good mix and the sound editors wouldn’t need to go back to the elements. A lot of that is due to the fact that the mixer is more comfortable taking chances knowing if he makes a mistake, it’s recoverable. We don’t want to make a mistake so often we’ll play safe, which might not be the “best” solution. 

It’s very important to decide how you’re going to work and to stick to the same track setup throughout shooting. Always track one is your editor’s mix track and the other tracks are your elements. In your default file name you always try to put the same things on the same track numbers. That’s very helpful in post, because sound editors immediately know where to look. They get an idea of how the production mixer works. The easier it is for them, the more you will shine. The jobs that I’ve been happiest with have been where I have worked with the same post people. They know how I work and I know how they work and hopefully you complement each other. You give them what they need and they let you know what they need and where and how they want to find it. 

Maybe we should touch on why a picture editor doesn’t want to track cut on a number of tracks. If you provide six separate tracks for the picture editor, because he cuts checker board style, he actually needs twelve tracks on the Avid. That leaves him very little space for all the additional stuff he has to add to the cutting copy. Like temp music, temp FX, maybe voiceover. He runs out of tracks and while it’s possible to cut picture with more than eight tracks, it’s very cumbersome. Eight tracks is the maximum number that most picture editors want to use. Usually they have an analog external mixer with a limited number of inputs. If you give them four tracks for editorial, that still means that all those eight tracks are covered and it doesn’t give them any extra, so the ideal that most picture editors want to have is one or two tracks. So if you can think about mixing as if you were mixing on a mono Nagra to that one track (as you probably all used to), and knowing that that’s undoable by people going back to your other tracks. It’s the best solution. It’s easy on the editors and on post sound. 

If you’re recording 6 tracks on the PD-6, you have 28 minutes total per disk, depending on your sample rate and bit depth. At present we are very much influenced by what the film composer is capable of and remember the newest 2003 versions are not widely spread and won’t be till 2006. So right at this moment I’m shooting 48K sample rate , and only recording in 16 bits . Shortly I intend to record in 24 bits. I’m recording on my current (2003) project in 16 bit for continuity purposes. The Avid film composer is very close to being able to work in the 24bit format which will be better for all of us. However, recording production sound in 16 bit works pretty well. It works well with DAT and it’s worked pretty well with the situation so far. It means that material is imported effortlessly into the film composer and also those files can be transferred directly to ProTools in post production. The important element of this is that we need to remember that that’s the whole aim, so we don’t have to go back in post production to the original disks. There’s seldom time for that. It should be that the sound editor is able to take the files from the film composer, by whichever method he chooses, via OMF, by taking a drive, by networking, etc. The important thing is the sound editor is able to access all those multi-track files, including “out takes”, immediately. Often the biggest restriction on the dialog editor is time. There’s almost always pressure to a) come up with an ADR list (because actors need to move on to the next project) and b) to make a temp mix. Producers very often start a post sound department within a very short time before they need to come up with a temp preview of the film. So there’s a lot of pressure on the dialog editor to find out takes and replacement lines. If it’s difficult to find, then they often have no alternative than to mark stuff up for ADR. 

The meta data stays with each sound file and the sound editor can go back to the original disks, if required. There’s a very good program from a company called Gallery called Metaflow http://www.metaflow.info/ᅠ which was written up in the Editors Guild magazine May 2003. 

http://editorsguild.com/newsletter/MayJun03/mayjun03metadatametaflow.htmlᅠ

That’s a very good article that explains Metaflow which makes the Avid film composer interface much better than just the EPC with the Pro-Tools. It uses the meta data to carry this information across. It’s a meta data tracker. And it gets you back to the location file, the disk number, etc. The point is we’re always reliant on how the Avid assistant has imported the material. 

However we work, the most important point is that the Avid assistant does not delete any files. Even though those files may not be used by the picture editor at that point, they should be kept within the film composer. The PD-6 deals with this by having a couple of track modes. The most useful are the 4+2 mode now in 2003. The 4+2 means that you’ll have two separate files in each recording, so when you record, two files are made. There’s a file that has just tracks one and two in it. It has another file which has the other four tracks in it. I recommend using the two track file for the picture editor. The other four tracks have the components of that mix. The two files are distinguished purely by the way they are labeled. The file name for both is also the slate number. So let’s work with slate number “123Apple take 1”. It will look like “123A_1” for the 2 track file and “123A+1” for the 4 track file. The convention is that one uses a capital “A” for the shot and where you have the scene number which is perhaps “a123” you use a lower case “a” to distinguish the scene number. This is just a convention I’ve found that works and I’m sure other editors and mixers have different conventions. 4+2 track mode is particularly useful because many versions of film composer will only allow a maximum of four track files to be imported and would not recognize a five or six track file. With the two file procedure, both files are recognized. The latest versions of Avid ( past 2003) will be able to import up to an 8 track file. When that’s possible I would probably recommend using the 5+1 mode, where track one is again the cutting copy or the picture editor’s sound track and you have all five of the other tracks to be used as components or elements of that mix, 

I often record backgrounds to the other tracks . For instance let’s assume we’re shooting at a location with only one boom, so we really only need one track. Perhaps another boom for the off-line, so now we’re using just two tracks but perhaps something like a police car or fire truck goes by with the siren blaring. For the show that might fit in very well, the director may have no problem with hearing a siren in his master. The problem of course would be when you cover the close-ups and there would be no siren. The way we’d get over that is that we would have put another microphone in another room, perhaps on the upper floor, which was discreet from the dialog, you’ve recorded no dialog on it, but you recorded clean background. You can put that on a separate track on the master. The intention being that when you cut the close-ups in you could cut them in with that background track only on it that’s continuing the sirens throughout the whole scene. That’s just another way of working that are new to us with multi-track recording and rely ultimately on ease of use by post. 

Timecode, Userbits, Playback & Userbits – syncing with clap sticks: 

Everything about film is about compromise and what works best. There are always ways to use more tools, make things more complicated to make our department safer or better. Usually better tools impact on to different departments. The more we try to do, the more we try to use, sometimes the more complicated we make things for ourselves and others. I think this whole multi-track idea still sounds complicated at the moment but the whole key to this is simplicity; to keep it as simple as possible. With that in mind let’s look at timecode and timecode slates. Timecode slates are generally disliked by the camera department (they are heavy and fragile and expensive). The more cooperation we get from them, the simpler it makes our lives. Most editors have so far been happy with standard clap sticks in the major feature world anyway. Unless anyone has a preference, my advice would be not to use digital slates on features. I’ve not been using them for the last couple of productions ( and they were all large budget films). 

I prefer to use record run timecode for a number of issues. It’s generally easier to handle. In file based recording we no longer need these 10 seconds of pre-roll. So that’s not the reason for it, but it’s very useful to have on side one of your 8 centimeter DVD Ram disk timecode which starts with one hour and on side two have timecode which starts at two hours. It makes it very easy for someone in post to recognize the source of the material. It also makes it easier for editorial if they know what sort of timecode to expect. That’s my main reason for using record run timecode. Time of Day doesn’t have any major advantages unless we were working in some kind of multi-camera situation. 

User Bits: 

I have another issue about user bits which may be alien. I don’t generally see the advantage of including the dates within the user bits. In many ways I find it a disadvantage when I’ve been trying to work with auto conform programs, trying to use the timecode to re-conform.  One of the issues I’ve found about putting dates in is that it’s not a consistent number. Sometimes in the month you have 31, 30 or 28 days. Months go up to 12 rather than 10. It’s not a format which easily interfaces with the way that we work within computers and a metric system where we expect to see things in tens. The important thing is if you have the reel number within the last three digits of the user bits , then you will be able to identify one timecode from another on a roll. The reel number is what will identify the roll number. The reason why people generally put the date in is you may have the same timecode come up the next day and you need some way to distinguish that. It’s all the more important when you’re using record run when you might be using one hour timecode for every disk. The reel number does this. 

On a PD-6 we have something other than that which is called event numbering which is also on the PD-4. It’s like a counter at the start of a project which you can set at zero at the start of a project. Every time the machine goes into record it will increment the first five numbers of the user bits. So you would start off with user bits on the first reel that would have seven zeroes and a one. When you did your first recording it would have 00:00:10:01. That means it’s the first recording on roll number one, the last three digits being the reel number. The second recording you did would then have 00:00:20:01 because we’re still on reel one. What I’m saying is this number would increment throughout the whole project up until you have a five digit event number which is 99,999 different events. The advantage is if you were able to keep a database of these event numbers in some way, it would mean that you could identify the source of every recording, every clip, rather than a reel. Every time you rolled or went into record you could identify the source of that just by looking at the timecode UB. So even if you transferred something in analog or any other way as long as you carried your timecode along with audio there would be a point of reference to recognize the source of that material. Also, using this means that absolutely every recording has a completely unique recording number within the user bits. Though The PD-4 does this I’m aware that few mixers in the US use this option 

People have gotten locked into using time of day and giving 10 seconds of run up that used to drive directors crazy. Using record run you don’t need that because you would have your pre-record on the previous take which is why we first started using record run. Record run was used much more in the UK than it was here in the US. One of the reasons had to do with the type of digital slates being used. The US ones worked on a crystal lock where you jam generators together every four hours. Though these are widely used in the UK Ambient Recording in Germany have a slate with a built in receiver to display a transmitted timecode. I found this to be a much more accurate way of working. Event numbers is something I pretty much established. Though its  in the PD-4 manual, but like everything in manuals they’re never explained terribly well. 

In PD-6 one thing that will be quite controversial is something called Circle Take . It allows you to mark your circle takes. I think that’s an advantage if you’re going to be using a mag route or telecine route and not importing directly into the film composer. When you’re importing directly into the film composer, the circle take mode is redundant. You should not need to mark circle takes. Because the whole idea is that the AVID assistant imports every take whether it be a circle take or not, so that the sound editor is able to access out-takes. In fact there’s a disadvantage in labeling circle takes because in effect it actually changes the file name by putting the character in the file name to identify it as a circle take, as that character is not required within the film composer it could well be deleted. This would mean the files within the film composer may not necessarily relate exactly to the file names on the original disk. 

Correction of file names on the PD-6: 

Very easy. Again there’s a quick key format. You hit F1 on a keyboard. I would suggest that when people are working a cart operation, always try to use a keyboard; it makes life a lot easier. You press F1 you’ll get into the edit filename menu and you have to be on the file you want to edit and you can edit the filename. Editing the filename automatically will edit the information within the meta data, so when the file is imported into the Avid both the new filename and the meta data are imported at the same time. 

False start routine: 
Fostex may slightly change the way the false start routine works in the background. We spoke about the way that you put in default file names in the slates and the take number field that will automatically increment for each take. But what happens when you have a false start? It will really mess things up because it will have incremented the default filename and it’s not a file you want. So Fostex has a false start operation. You hold down clear and press standby on the machine and it will enter the false start routine. You’ll be prompted as to whether it is a false start. Pressing enter will delete that false start permanently and reset the default file name back to the previous take number. It shouldn’t be used lightly because although you are only able to use it on the last file you recorded, it does delete the file permanently. you can never have two files with the same name, because if you were to manually delete the file it is only put in to a “trash” bin..  so you would be unable to record a file with that same filename again. what you need to do is change the filename before you delete it. For example, add FS before the file name before you delete it, so that would allow you to go back to the original take number. The PD-6 false start routine is very quick and easy to use, and if you’re sure that it definitely is a false start then there’s nothing wrong with using that. Because I’m comfortable with the machine I use it all the time, but I would hate for someone to delete a file by mistake. When you use the false start routine it is NOT recoverable, but a manual delete would go to a bin and therefore would be recoverable. Fostex engineers are working to change the false start routine to make false start files recoverable, but At present (2003 July) they are not recoverable. 

How long does it take from a cold start to record? I think it’s 12 seconds on the PD-4. It could easily be 30 seconds on the PD-6. It has to search the disk, it has to recognize that there’s a DVD RAM disk, it has to look for what files are on there. On NP-1s you would be powered up all the time in an over-the-shoulder documentary situation. 

Run around guys, over the shoulder guys want to change from one battery to another without interrupting power. So the best way is to have an external battery with a 4 pin XLR on it, you can always have it plugged into the external socket so you can take the internal and charge it with no interruption. 

How long does it last on lithium? Two hours is quoted. I have my doubts and I would like to look at that in the real world. The PD-6 operation has been changed slightly in the latest version. They found that they could shut down the DVD RAM drive quite often instead of it running all the time. That saves a lot of battery power and saves heat, so it has become more efficient. They’re also looking for other software improvements to make it even more efficient. 

Pre-record and post record time buffer: 

The pre-record buffer can be pre-set for anything up to 10 seconds.  Most machines only allow you to use the pre-record buffer when you’re using Time of Day timecode because of the difficulty of joining the timecodes in record run. The PD-6 pre-record buffer works whether you’re using record run or time of day simply by filling up the buffer with timecode information and will give you a continuous timecode. If you are rolling time of day with pre-record buffer, what happens when you hit the stop button? If you’re in Time of Day when you hit the stop button, it will stop. It just has to store its timecode. Then it’s very easy because when you start to record again all it does is timestamp the next recording with the new time of day. It can’t do that in record run because it needs to jam the timecode to the time recorded on the previous take. So what it will actually do is temporarily extend the previous take by the number of seconds selected in the pre-record and then it will steal that bit of timecode back to go in front of the next take. So it will effectively give a continuous timecode, but it allows you to have timecode on your pre-record buffer. When you’re in record run timecode and you hit the stop button – let’s say you have five seconds of pre-record selected then it will take five seconds before the machine actually stops recording. However, because the buffer would be once again active, you could hit record again and go immediately into recording again because it will store that amount of time in the buffer until the machine is ready to go into record again and then it will dump it into the recorded file. So that’s important – there’s no delay there. I think with some other HD recorders there was/is. But I don’t know. My understanding was you couldn’t use pre-record when in record run but I’m no expert on other HD recorders. 

What to do with bad disks ? 

I have to admit that in the early days I have had bad disks, generally not from Maxell. The great thing is the difference between recording on a DVD RAM disk as opposed to a hard drive is that you know immediately you’ve got a bad disk when you start to record on it. It will tell you that you have a bad disk because you have read write verify as part of the recording process. A bit like listening off tape except that the machine is listening off tape for you. It’s verifying everything that’s being written as it’s being written and it will immediately tell you if it’s not been written correctly. Unlike a hard drive where you could record a whole day, a whole week’s work and it could be recorded perfectly well, but through a problem with the hard drive you may not be able to read that information. For me the great benefit of a DVD RAM disk is that if it’s been written on, you should be able to get it back, and you know bad disks immediately, generally when you’re formatting them. Using Maxell 8cm discs I have not encountered any disc problems now that manufacturing and QC is established but The best way to deal with any bad disks is to give them a full physical format before you use them.

Perhaps I should explain this whole idea of formatting disks... They are formatted in UDF. That’s a Universal Disk Format, which is the format that Fostex has chosen to use because it makes it usable within the Avid and Pro-Tools system. There’s a big advantage in using this UDF file format because it’s equally at home in the PC and MAC environment. No accessory software necessary. You have to format the disks before you use them. There are two ways of doing that: the full physical format which is somewhat like formatting a hard disk. It will give a full format to the disk but the downside is that it takes something like 19 minutes to fully format.. However in normal circumstances one only needs to do a fast format like what you’d do on a floppy disk. This takes 20 seconds and is the format that I always use. 

Undirected production mixers: 

Most production mixers have the best intentions. They want the best for the film. Sometimes they’re not well directed. Post doesn’t tell them what they want and, unfortunately, there are not very many mixers that have much post experience, so they still believe that what they are giving post is what they need, but that might not be true. Also a mixer uses equipment that he’s most comfortable with, like a Nagra, not realizing why down the line it’s very difficult for post to deal with those tapes and regrettably it’s sometimes easier for them to ADR than to go to a lot of trouble archiving and recovering those tapes. So the location mixer actually hasn’t done himself any favors. It’s very important for the production mixer to interface with post people and for each to swap information on who needs what. 

An example of this is that DAT grew only because post asked for it. They insisted on it, because it was too difficult to work with ¼” tapes. It meant that what they were doing was transferring ¼” out to DAT. Even people using Nagra D shooting in wonderful 24 bit and everything else probably didn’t know their material was being transferred to a DA88 because editors could find it quickly and the DA88 had serial control that interfaced with their workstations. I think the mixer would have had heart attacks if they realized it and that’s probably why post rarely told them what they were doing. What’s happening with the PD-6 is this: I get less enquiries from production mixers than I do from editors because it makes things work so much better for them. In fact I have yet to encounter an editor who, having tried this route with the advantages of magless dailies, has wanted to regress to older working methods. I also get a lot of enquiries from producers because if it works better in editorial, it saves producers a great deal of money. We ought to talk about where that money is saved, and it starts right at the very beginning. ¼” tapes for instance are not cheap. Fewer and fewer companies are making professional standard ¼” tapes. You could use six reels a day. Chances are you’re going to use one DVD RAM disk. I believe they cost something like $18 and it’s going to come down. So right from the start the stock price is going to be lower. 7” ¼” @ 7 ½ = 16 minutes = $8.00 (x3= 3x16 = 48 minutes $8 x 3 = $24). DVD RAM disk 2+4 files at 16/48 = 36 minutes = $18. The need not to use mag, that’s a major issue. Can you imagine using 6 cameras? You have to do a mag track for each of those cameras. It’s not unusual for a major feature film to use a million feet of mag. Just think about the transfer costs. That excludes what it costs to get it to the transfer house and back again and all the other storage and handling issues. Getting it retransferred when a select is missed. All of those issues. Avoiding mag also saves the producer having to run a film and a nonlinear cutting room. When you use mag you have to use two cutting rooms, film & non-linear, in parallel. Obviously, it means more staff and more equipment. 

Note from ed.: Picture editors really are happiest if they get only ONE audio track. If there is a second one they have to spend twice tha time reviewing the tracks and making choices. A picture editor will rarely have the time to do this and sound will suffer because of all the pressures picture editors are under. At the same time picture editors wil want the assurance from a competent reliable soundmixer that the other 4 tracks are always there and are available to the sound editors latter. Picture editors can live with the fact that the one track they get will have mistakes and slpoopy mixes on it if they have confidence in the 4 backup tracks. It is rare they have the time or inclination or skill or proper monitoring environment to come to an actual judgment on the other 4 best racks. 

Another advantage is the editor gets absolutely all the audio into his Avid immediately. Using the scenario where he telecines his film MOS direct to an Avid hard drive and syncs up entirely within the Avid, rather than the blind syncing method, means he has access to the material before anybody even sees dailies. The problem with the use of mag is everybody has to see dailies before editorial can then send the stuff off to telecine to be put on tape and then it has to be digitized to the Avid. It makes it very hard for the picture editor to keep up with production because he’s always waiting to get the material in. Whereas directly accessing the material means he can get it in very fast. He can see earlier how things are going to cut and he can advise the director if things are working before the director’s through on the set. 

Dailies are finished the next day earlier and can generally be viewed at lunch: 

You have two scenarios. You have what’s known as direct to the Avid hard drive , digitizing of the film in telecine, with the Avid media station. In that scenario what would happen is the film goes to the lab for processing overnight. It probably comes out of the lab at 5-6 a.m. First an assistant would order the rolls into a dailies order, i.e. the selected takes the director wants to see would be laid into dailies rolls, depending on whether the lab had printed only the selected takes or everything. It then would go to telecine at the lab. Let’s assume they printed the circled takes only. What the assistant may want to do is put all the A camera stock together, and all the B, so you could sort the rolls at the lab. Other directors may have no preference, they’ll be happy for the lab just to put it on rolls, send it directly to telecine. Using an Avid media station picture only is transferred directly to an Avid hard drive, at the same time being transferred to a Beta tape MOS. The hard drive goes to editorial, they put it on their system. They’ve previously imported the audio direct from the DVD RAM drive into their system, now they sync up within the Avid using the clap sticks. Typically that’s a half hour to 45 minute operation. It is not generally speeded up by using timecode. For the advantage that the camera department is gaining by not using timecode slates and the good will that mixers get. If camera departments know that, they do their best to get a clapper on it, no matter what it is. That saves them from using timecode slates. Camera departments are only too willing to get the clappers on. It makes their lives easier.  Again we get back to compromise and cooperation. 

Once all takes are in sync, all the assistant has to do is export. He’s put all the audio files into the right order. They match the pictures files. He’s made the composition. He exports that composition to another DVD RAM disk that will run in sync with the original film dailies. He puts the picture on the projector which is connected to a Fostex DV40, slaves to the projector. Presto, we see sync dailies on film.. Some projectors automatically output timecode starting with 00:00:00:00 at the punch. Many of the Kinetones have the capability to give timecode out when they run and certainly the Arri-lock pro gives timecode out; it doesn’t give bi-phase unless you have a separate interface. But it’s no major problem to get timecode out of the projectors if they have a Biphase output. There are a number of manufacturers making Biphase to timecode converters. So in that scenario, making lunch time dailies is no problem. And that’s the key, to get lunch time dailies. 

Let’s say the telecine was unable to telecine MOS dailies in time for whatever reason, then in editorial we would take the picture and put it onto a kem. On the kem we would look to see where the picture slates actually clapped and make a note of the footage and frame number on the roll, and give that information to the Avid assistant by phone if necessary. He would use a method known as blind syncing , where he would then put the sound clap for that take in exactly that footage in the Avid, make his composition and export to the composite disk in the previously described way. For safety’s sake we could have a Fostex DV-40 slaving to the Kem so that we could take that composite DVD disk to the Kem and check that everything was right before dailies were sent to projection. We would then send the disk and the film to telecine later on for film to be telecined to a beta tape which would then be digitized into the Avid film composer in the normal way. 

Note from Ed.: An editor who recently used this system was admant that a third Avid in the editing room is really needed to do this additional work. The second Avid is used by the traditional 1 st assistant to do all the usual housekeeping chores that a first does on a normal film. So there is additional labor involved for a syncing assistant. This wouold be most efficient if there was a portable laptop around that could this this work and its little fixes in various editing spaces. 

One of the questions often asked is how do we make producer’s dailies tapes ? How do we make studio tapes if we’re stuck with digitized MOS? What we do is when we digitize directly to the Avid hard drive we make an MOS beta with the timecode on it. We then lay the audio back onto that beta from the Avid. You’ve only got one sync mark at the beginning of the reel. Lay that on and then you can make all the producer or studio tapes from that beta. Lay it in by telecine in one piece. Some editors have loved the idea of making the studio tapes from within the cutting rooms. It means cost savings in telecine allows them to have an extra intern or assistant. It also means they have absolute control of everything that goes out to the studio. Different editors like to retain different amounts of control. Some like to have the control to know exactly what’s going out, and who sees it. Other editors don’t like to have their cutting rooms turned into a telecine – pre-edit bay. At the end of the day, of all of these methods, nothing is set in stone. 

SoundStation Ltd, advise the producer on sound methods, we provide magless daily setup for probably most films that happen in the UK and Europe. We provide the recorders. We talk to producers, directors. Do work flows. What we tend to do is set up a work flow for each production or we’ll adapt a work flow. Production will approach us and say they’d like us to provide DV-40s and a magless dailies service and we would tailor it exactly to how the producer, the editor and the production sound mixer wants to work. That’s what we at SoundStation do in addition to our usual role of providing ProTools and other NonLinear audio systems. 

www.Editstation.com

DV40 and six tracks: 

DV40 at present (June 2003) is a four track machine with a separate card (5050 card) that has to be dealer fitted, which makes it capable of reproducing six tracks with PD6 DVD disks. Outputs are analog and AES/EBU. 

Expansion Chassis News: 

PD6 - the extra inch below to be released 2004: 

Fostex is working on an expansion chassis. Fostex recognizes the difference between the mixer who wants to work as on a cart or over the shoulder. There are the guys who work on documentaries and on TV and wear the machines over the shoulder. The recordist wants the machine to be as small as possible with as few things hanging off of it as possible. He may be working in video mode where he’s using the mixer section of the PD6 to use the multi plug out to go straight to the video camera. He may be using his disk purely as a master. Maybe he might work in the 4+2 mode where he sends two or four tracks to the digi-beta but uses more tracks as his master. He would be very happy working in a light weight mode. He would be restricted in disk time with a small disk, but probably no more restricted than the length of a beta tape. I think they’re 50 minutes now. The F900 is 56 minutes, DV-pro is 160 minutes. Using both sides of the disk, it may be that each disk correlates to each tape on digi-beta. However, for the production mixer working on feature films which is much more my area, I feel that they might like to record more commonly not onto the 8 centimeter disk, but onto the bigger size DVD RAM disk which gives us much more recording time, which is important when we get up to the higher sample rates and higher bit rates, but also is preferred in editorial because the disk doesn’t have to be removed from the cartridge. Therefore Fostex is about to release an expansion chassis for the PD6. I anticipate the expansion chassis will record directly to a full size DVD RAM disk. It will also have additional benefits like perhaps a power supply and better meter. I expect to see the expansion chassis imminently. 

As far as I’m concerned, hard disk wasn’t the way I wanted to go. I know there’s lots of talk about where’s the backup and so on and so forth. For me backup is only backup if you’ve got a completely separate recorder. If you record to a hard drive and then that is mirroring down to a DVD RAM disk or whatever, if something goes wrong with the hard drive before it’s mirrored, then you’ve got only limited backup. You’ve only got backup if you run two completely separate machines, and let’s face it, then you’ve only got backup if you’ve got two separate power supplies. How far do you go? In my mind, what is a huge advantage is recording directly to DVD RAM disk is the write verify operation, where it immediately verifies what you’ve written. You know what’s on there. It is a physically changed disk. The laser has physically changed the DVD RAM disk to make a master recording. I like that with write/verify you know it’s going to make sure it’s reading the same information that was written. 

Limiters: 

It has very interesting channel limiters. Which are user definable. You can set the ratios and the level of limits on each channel and they work in the analog section. I’m pleased with the limiters, and also particularly with the mic channels. The mic channels are just not in any way comparable to the PD4 which I guess is using 10 year old technology. Technology has really changed here. It’s much less power hungry. I’m very happy with the mixer actually on the PD6. One needs to work out the way that works, not just as a mixer, but also the way that it routes track to the tracks that you want to allow flexibility for the editors. It works in a different way from the way most mixers work. You always mix from the left out to the right. So generally, you’d expect the clean tracks to be on the higher numbers, on track seven or eight. This works so that you put the mixes on the lower numbers. That’s purely for editorial convenience, because when you import directly into an Avid you’ve taken all the tracks in, you really want your mixed tracks to appear first in the smaller track 1+2 file. 

Protools: 

The PD6 disks are said to work OK with Pro Tools ver 6.0 on OSX. No problem with 6 track files. Avid Adrenaline running on OSX, to be released in 2003, is said to be able to import up to an 8 track Polyphonic file. 

Unfortunately ProTools still has a 13 hour timeline limit (in 2003), making it difficult to auto-sync using time-of-day timecode. 

The BWF format allows ProTools to use the original TC and not a new TC generated on import. The preferred working method however would be to OMF files, to take a drive from the film composer or even to have ProTools on the same network. There would be no conform as in a telecine synching or mag synching process. The original import of files to film composer will be of good quality and all “out takes” will be imported. All available tracks for each take will also be imported and available from the Avid even though the picture editor may choose to edit with only one track of production sound. This is why it is important that the production mixer creates a mix to track 1 and uses tracks 2-6 to record the elements of that mix pre-fader. However, the current restrictions of the film composer will only allow import of files containing up to 4 tracks this is catered for within the PD6 by using the 2+4 recording mode which actually records 2 files for each take the 2 track file containing the mix on track 1 and perhaps OS dialogue on track 2. The 4 track file has the individual microphones that make up the mix. It is important that picture assistants do not delete any of the files that the picture editor is not using as they will eventually be required by sound editorial. 

In & Out tags (handles) are always available for the reasons outlined above unlike in an auto-conform process where the handle length is specified. 

PD6 pre record buffer: 

The way the pre record buffer works (depends on your software version of course but note that all software updates are free and downloadable from www.fostexdvd.netᅠ) is that when you press the pause key, the pre-recorder buffer "empties" onto the disc for want of a better description i.e. records the remaining 10 seconds or so. However in very early beta versions it would do that and nothing else i.e. you'd be stuck for 10 secs after the end of the take if you had a 10 sec buffer on. However, now what happens is that if you press REC during the buffer emptying/recording process, it continues to empty and record to the disc, but immediately (a 0.5 second gap perhaps) starts to fill up again so you won't miss any of the new take. e.g. if you had a 10 sec buffer and were 5 sec through the process, you'd have a 5 sec pre-rec on the new file. There won't be any TC overlaps of course. 

(Because we generally prefer to use Record-Run TC, so that for example we can have 01hour TC on side 1 of the disc and 02 hour for side 2, the pre and post record function calculates TC and avoids overlaps. The way that this works is that at the end of a recording the machine enters a post record mode to account for TC needed or the next pre-roll. However, the pre-roll buffer is utilized to capture audio from the next take if the record function is entered during the post-recording period. So there is no wait in the recent versions of software.) 

Avid requirements: 

Avid requires that all files have TC though it is not vital to use a smart slate. Avid can not handle more than 4 audio tracks at a time. Bit depth and sample rate varies between versions. 

The future formats (post 2003) that will be used in the Avids for audio files will have more meta data space for meta data. 

Chris Munro is a director of   SoundStation, 116 St. Margarets Road, St Margarets, Twickenham, Middlesex, TW1 2AA, Great Britain.  Much of the material here is copyrighted by him. The company advises and consults with producers as to the best way to record, edit and mix sound on motion pictures. The company rents audio and non linear audio editing equipment and facilitates soundpost. SoundStation Ltd is based at Twickenham Film Studios www.editstation.com. 

General PD-6 and PD40  info at http://www.fostexdvd.net/

  Home    Products    How To Order    Video Assist Services    Literature    Contact 

 

© Copyright 2005-2008 Wolf Seeberg Video - all rights reserved