Although all of us now work with digital sound and picture files, our industry has been slow to change the way we manage and use information about those files. At the same time, incompatibilities between picture and sound workstations have prevented us from easily sharing that information. The result has been widespread frustration for picture and sound editors and their assistants, who must still do a great deal of their housekeeping work by hand, laboriously reassembling audio from EDLs or searching through stacks of paper for alternate takes or channels in original recordings.

If you consider the actual picture or sound as “data,” then information about that data is “metadata.” Traditionally, metadata has been stored separately from the data itself — a reel of quarter-inch tape with a separate sound report is an example. More and more of us have been using databases — both custom and off-the-shelf — to organize information about recorded picture and sound, but often that information still starts with pencil and paper. Now, new tools allow us to create metadata digitally and store it as part of the picture or sound files, which can streamline the entire post production process.

For example, a new generation of direct-to-disk audio recorders such as the Zaxcom Deva, Fostex DV40 and PD-6, HHB PortaDrive and Aaton Cantar save audio as computer files. But these recorders can go a step further, storing metadata in the same file, so that both data and metadata can be delivered to the picture-editing team in one neat package. At the same time, the venerable Edit Decision List (EDL) is slowly being supplanted by workflows based on the Open Media Framework (OMF).

Working in tandem, these technologies will allow us to set up new kinds of digital workflows where the original audio, together with all its associated metadata, can go from production recordist, to picture editing room, to audio post production — simply, automatically and without degradation.

Broadcast Wave Files

A good example of data/metadata creation and control is the Broadcast Wave File (BWF). It’s the good old WAVE audio file, but with enhancements for our industry, including specific data “chunks” that can hold a wealth of important information. Once stored in the file, applications that know where to look for the metadata have quick and easy access to that information. (For the complete BWF specification, see www.ebu.ch/tech_32/tech_t3285.html.)

Modern Avid Film and Media Composers are now able to import BWF files recorded on non-linear field recorders. These recorders use the “broadcast audio extension” data chunk to store metadata. At the very least, this includes the production timecode, which is automatically entered by most recorders. If the recordist is able to input scene and take data or any extra notes such as microphone designation, that can also be stored in the BWF file. When these files are imported into an Avid, they are converted into 16-bit OMFI audio files with the same metadata, which is also automatically placed into appropriate bin fields.

MetaFlow's Master Clip Window
Figure 1. MetaFlow’s master clip window, showing clip name, scene, take, notes and the Avid source bin, all imported automatically from metadata. The layout mimics the look of an Avid bin. MetaFlow’s playback engine is shown in the lower half of the window.

But how do we get this data out of the Avid and into our DAWs? Let’s look at another Avid product: the near-ubiquitous Pro Tools. Pro Tools and Avid may be made by the same company, but Pro Tools has never been able to get at all the metadata inside an OMFI audio file. Fortunately, Gallery Software (www.gallery.co.uk), which makes the indispensable Session Browser, has just released a new product called MetaFlow (www.metaflow.info). It is designed to access, organize and utilize audio metadata and can solve some of the stickiest problems in OMF workflows. It can also handle some important nuts-and-bolts tasks that have previously been extremely frustrating to sound editors, such as searching for alternate sound takes or alternate channels, and converting a 16-bit Avid audio sequence into a Pro Tools session containing the original 24-bit production sound.

Searching for Alternate Takes in Your 16-bit OMF Library

In the July/August 2002 issue of this magazine, Larry Schalit talked about how he creates a custom database to search for audio by scene and take. This is an invaluable tool for any dialogue crew on an OMF-based show, but it requires considerable manual labor from the picture and sound crew to get the data in all the right places.

Master Clip List
Figure 2. MeatFlow's master clip list showing individual channels associated as a single master clip.

An Avid stores its data not only in its bins, but also as metadata in its OMF files. The trick is getting at that metadata, and this is what MetaFlow was designed to do. It can create an OMF-searching database simply by pulling metadata out of OMF audio files, OMF sequences or Avid bins. Once the metadata is in the MetaFlow database, a user can search for a setup and quickly find all alternate takes. It is then possible to listen to any channel of a take and spot one or all channels directly into your Pro Tools session (Figure 1).

Alternate Channels with a Mouse-click

MetaFlow can also use metadata to link multi-channel recordings together into what the program calls “master clips,” borrowing Avid’s terminology. With the product’s Session Expander, an editor can look at regions in a Pro Tools session, match them up with their cousins, and automatically enable access to alternate channels. For example, if 25G/1A was recorded in four channels (boom, radio mic 1, radio mic 2, mix), but the picture editor only used the “mix” channel in his or her cut, MetaFlow will create Pro Tools regions for every other channel in the take. If the original region in the OMF sequence was called “25G/1A,” MetaFlow will rename the region in Pro Tools “25G/1A^4.” You will be able to access the other three channels, “25G/1A^1,” “25G/1A^2,” and “25G/1A^3” by command-clicking with the selector tool (Figure 3).

BWF and the Elusive 24-bit OMF

Though they’re made by the same company, Avids can only use 16-bit sound, while Pro Tools can accommodate 24-bit material. This means that even if audio was loaded digitally into the Avid, it may not be as good as the original recording, which today is often 24 bits deep. This has presented a frustrating “gotcha” to productions employing a 24-bit OMF workflow. They have either relied upon hacking the Pro Tools sessions with a hex editor or manually reloading and resyncing every 24-bit audio file to match their 16-bit counterparts from the Avid.

Avid Timeline

Protools Timeline
 
Figure 3. A sequence in an Avid Film Composer and the same sequence in a 24-bit Pro Tools session. After processing the session with MetaFlow, alternate channels within a take can be accessed with a simple command-click.  

But, if your production audio was recorded using a metadata-friendly medium like BWF, MetaFlow can do much better, because it can easily compare the metadata from the 16-bit Avid OMF files to that of the 24-bit BWF files and find common data points. It can then automatically convert a 16-bit OMF sequence into a perfectly matched 24-bit session that points to all the original 24-bit production files. (This only works if the file lengths are identical, which means that audio must be imported into the Avid directly from a file-based recorder, not loaded analog from tape.)

Metadata Workflow in the Real World

Here’s a brief explanation of how metadata and MetaFlow are being used on Jersey Girl, the upcoming Kevin Smith film, currently in post production.

Production audio was recorded digitally in 24-bit. Using a Deva with a Cameo mixer, the production sound recordist embedded metadata, such as scene and take information, into the audio, which was recorded as BWF-P files, with one file per take, each containing as many as four tracks. (The “P” stands for polyphonic or multi-track.) The Deva names each file DDDNNN.bwf, where DDD represents the shoot day (in Deva terms, the “Partition Number”), and NNN is a sequential take number for that day (the “Segment Number”). For example, the 12th recording on the 15th day would be “015012.bwf”. At the end of each day, the audio was copied to a DVD-RAM disk, which was then delivered to the picture-editing department.

These BWF-P files were then imported into the Avid in audio-only bins, one for each DVD-RAM from production. For each take, the Avid created a single master clip pointing to up to four physical sound files. The associated metadata was included in both the Avid-created OMF audio files and the bins. The audio was then autosynced, and data was checked and corrected as needed in the bins by the picture assistant.

Audio post was given a copy of all the BWF-P files recorded during production, a copy of all the 16-bit Avid OMFI audio files, and a copy of the original audio import bins.

Because Pro Tools cannot work with interleaved files (containing more than one channel of audio), all the 24-bit BWF-P files had to be converted to monophonic, 24-bit BWF-M files. MetaFlow does this conversion with a simple drag-and-drop, and the metadata is automatically copied into the separated, monophonic files.

MetaFlow then imports all the BWF-M files, as well as all the 16-bit OMFI audio files. Once ingested, the files show up in MetaFlow’s database windows (Figures 1 and 2). The program connects the individual files into multi-track master clips, which can then be browsed, searched by scene and take, auditioned within the MetaFlow environment and easily spotted into your Pro Tools session.

For Jersey Girl, we made great use of the MetaFlow’s SessionExpander function. The picture department sent a 16-bit OMF sequence, which was then translated into a 16-bit Pro Tools session via Digidesign’s DigiTranslator. This session was fed into SessionExpander, which re-linked all the 16-bit OMFI audio-files to their 24-bit BWF-M siblings, creating a new 24-bit Pro Tools session (Figure 3). At the same time, SessionExpander provided one-click access to alternate channels not used in the picture editor’s cut. The program also provides a straightforward way to examine all the metadata associated with any particular file, through its Media Files window (Figure 4).

These Are the Early Years

MetaFlow's Media Files Window

Figure 4. MetaFlow’s Media Files window provides editors with a wealth of data about their audio files. Here it is showing all channels for 34/1, as both OMFI and BWF files.

MetaFlow is currently the only product that utilizes metadata to streamline the audio workflow from production through post production. As such, it allows sound editors to easily make use of original 24-bit audio, compare takes and manipulate multi-channel recordings — all tasks that were heretofore frustratingly difficult and time-consuming. It bridges the Avid and Pro Tools environments and it’s the first product that begins to realize the potential of metadata to empower a new kind of automated workflow for production and post production.

It is important to note, however, that MetaFlow is useful only to shows utilizing OMF. If your sound workflow is based on EDLs, it will not be of any help. However, for sound departments working only with 16-bit OMFI audio, MetaFlow can function as a database that will greatly simplify access to your production sound. And for those who want 24-bit OMF, MetaFlow is indispensable.


Erich Stratmann is a sound editor working in the
San Francisco Bay area. He can be reached at email

Phil Benson is a post production supervisor for View Askew Films.
He can be reached at email