Film Sound Restoration

Because it's being pushed. Voice Isolation separates speech from everything else, and past a certain point it takes parts of the voice with it, heard as a watery, phasey quality with soft consonants. Backing it off trades the artefacts for the noise you were trying to remove, which on a delivery is rarely an improvement.

Once a recording only holds together at full strength, the damage is in the source and no setting will hide it. That is the work we take on, as fixed-quote projects.

Enhance Speech applies one model across the whole clip, so where noise, hum and reverb are present together it trades one against another. Separating them is the part a single control cannot do, and it's where the results diverge.

Past that point the damaged sections need repairing individually rather than processing harder. Send the file and you will get an assessment and a fixed quote before anything is booked in.

Often, yes, within limits. Where a recording clipped briefly on peaks, the waveform can be reconstructed and the harshness removed so the dialogue is usable. Where the level was so hot that the speech is distorted throughout, some of the original is simply gone and the honest outcome is an improvement rather than a repair.

Send the file and you will get a straight answer on what is recoverable, and a fixed quote for the work.

Usually, yes. Modern spectral noise reduction builds a profile of the noise itself and subtracts only that, rather than filtering by frequency. The limit is how much of the dialogue sits underneath the noise floor. Where speech is buried rather than merely accompanied by hiss, removing the noise removes part of the voice, and the honest answer is that some of the original will not come back. Restoring in several gentle passes causes far less damage than one aggressive one.

Yes. Source separation can extract dialogue, music and effects as discrete stems from a finished mix, which is how a usable M & E track can be produced for a film whose original session elements no longer exist. Results depend on the density of the mix. Separation is very good on speech over music and less clean where effects and music share the same frequency range.

Yes. Optical soundtracks are read from the film itself, either from a print or a soundtrack negative. The recording is captured at high resolution, then cleaned of the dust, scratches and density variations that show up as crackle and distortion. Optical tracks from the 1930s onward can usually be recovered to a listenable and often very good standard, and variable area tracks tend to survive better than variable density.

If the picture is already being scanned, the soundtrack can be extracted from the scan itself rather than run as a separate pass, including scans made on Lasergraphics and Scanity machines. That saves a handling of the film, which for shrunken or fragile prints is worth having.

LtRt, meaning left total and right total, is a two channel fold down that carries more than two channels of information matrixed together. Played back as ordinary stereo it sounds like stereo. Decoded properly it unfolds into the surround mix the film was finished with. So yes, a genuine 5.1 presentation can be recovered from an LtRt master, which is very different from upmixing a stereo track into invented surround.

No. Digitising captures what is on the carrier, including everything wrong with it. Hiss, crackle, distortion, dropouts, speed error and any Dolby encoding left undecoded are all preserved perfectly by the transfer, because that is what a faithful transfer is for.

The risk is that the file then becomes the reference copy and the original element goes back on a shelf, or is disposed of. Digitising buys time, which matters. It is not the same as making the sound usable, and the two are often costed as though they were.

Yes, in both directions. Dolby A and Dolby SR encoded material is best decoded at the transfer stage, because an encoded track played back undecoded sounds bright and unnaturally wide, and restoration work done in that state is being judged against the wrong reference. It can be decoded later from a flat transfer if that is what you have, and often is.

We can also encode back to Dolby A and Dolby SR, which is what archives need when a restoration is being recorded out to film rather than delivered digitally. The soundtrack negative has to carry the same encoding the print was made with, so the film plays correctly in a cinema fitted with the matching decoder.

In most cases. We transfer from DAT, DA-88, quarter inch tape, compact cassette, vinyl and 78s, 16mm and 35mm magnetic, optical film, and video formats including VHS, Betacam SP, U-matic, DigiBeta, MiniDV, DVCAM and Video8, along with digital files. Obsolete carriers are worth acting on sooner rather than later, because tape binder degrades and playback machines are getting harder to keep running. If you are unsure what you have, photographs of the labels and the reels are usually enough for us to identify it.

Yes. Delivery includes DCP authoring in SMPTE and Interop, 5.1, 7.1 and Dolby Atmos mixes, BWAV masters for archive deposit, and the stereo and mono deliverables required for broadcast and streaming platforms. Restored audio is conformed to the current picture cut before delivery.

Work is scoped in studio days, so the cost follows the condition of the material and what the finished version has to be. A short film is usually a day. A feature is typically three to four days for a straight restoration, and beyond a week where the mix has to be deconstructed into stems, rebuilt and remixed.

Every project is quoted after an assessment rather than from a price list, because a clean digital master and a decaying optical print are different jobs even at the same running time. Send details of the material and what you need at the end of it, and you will get a scope, a timeline and a fixed quote.

A short film is usually a day. A feature typically takes three to four days of studio time for a straight restoration. Taken further, where the mix is fully deconstructed into stems, rebuilt and remixed, a feature can run beyond a week.

The range is that wide because it depends on the condition of the material and on what the finished version has to be. Festival and release deadlines can be worked to, provided the material arrives with enough time for the transfer stage, which is the part that cannot be rushed.

Our studio is in Stratford, London. Physical elements can be shipped in, and digital files can be sent from anywhere, so a good proportion of work comes from outside the UK.

Where material cannot or should not travel, we work with partner facilities in the United States, so elements can be transferred locally and the files sent on to us. We can arrange and manage that for you, briefing the facility on resolution, alignment and documentation so that what comes back is what the restoration actually needs.

That part matters more than it sounds. Some faults in a transfer can be put right afterwards, but detail that was never captured cannot, and for fragile or unique material it's worth talking through the options before anything is put in the post.