Premix Dialogue WAV: Localized
The dubbing partner should create a mono WAV file containing all the premix dialogue units using the offline bounce function in Avid Protools, or their digital audio workstation. All premix dialogue units should be recorded at a well balanced volume and in sync with the provided picture. The premix WAV file should be submitted against the "Premix Dialogue WAV: Localized" source request.
- Before export, please attenuate any group walla tracks by -10db. For QC purposes, principle dialogue should play on top of group tracks to discern the two.
- All recorded audio should be included in your offline bounce. Make sure to select the entire length of the movie file or guide track audio. This will ensure the bounced WAV is the exact length as your picture and the picture used for QC.
- Bear in mind that the bounce to disk feature relies heavily on your outputs. You can select what outputs you want bounced in the “Bounce Source” dropdown. Make sure to select the correct output(s) of the tracks you want in your bounced file.
- Please insert a -6db true peak limiter on all dialog tracks or on bounce bus to control peaks.
Open the "File" menu and select "Bounce to" and "Disk." Click the "Offline" check box (bottom left) if you want to perform the bounce faster than real time. Select the other options as required for the output file format you need, name your file and select the folder destination.
- Click the "Bounce" button in the lower-right corner of the dialog box and your bounce will begin. If the "Offline" box was not checked, the bounce will take as long as playing the project file would. Offline processing will occur in a matter of seconds, with speed depending on your system resources.
- The bounced WAV should then be uploaded against the following source request:
- (Language) - Premix Dialogue WAV: Localized
- Naming convention:[PackageID]_ShowTitle_[framerate]_Mono_[langCode]_PreMixQC_[Version].wav
- Redelivery of the Premix WAV will be required once fixes are made for archival purposes. Please upload a new version to the source request to complete request.
Localized Edit Project Files
The Localized Edit Project Files session is a digital audio workstation session prepared for the alternate language mix at the sample rate, bit rate, frame rate, and the file format of the original version printmaster containing the following items: Raw, untreated, synced, and edited alternate language dialogue tracks from the dub recordings. This session should not be uploaded against the source request until the Premix Dialogue WAV has successfully completed QC and arbitration.
The Localized Edit Project File is mostly necessary for centralized mixing. Your Netflix dubbing contact will communicate whether your title is undergoing centralized mixing at project assignment.
- Ensure the submitted Pro Tools or DAW (digital audio workstation) session is self-contained and not linking to audio files from outside of the Audio Files folder in your deliverable session.
- Important: If Pro Tools is not being used to deliver the recorded dialogue, it is imperative that an AAF export be created and delivered. This will enable Netflix and its partners to import the audio into any DAW for mixing.
- Clean out unused audio from the session and Audio Bin.
- Framerate and start time of the DAW session is to match the provided proxy video
- Record 24bit/48Khz Audio.
Deliverable Session Naming
- [Title]_[PackageID]_[episode#]_[language code]_LocalizedEditProject_[Frame rate]_[version].zip
- [language code]_[character] - all caps or all lowercase
- If dubbing in a group style recording - [language code]_[character(s)]
- Name each track with the character’s name in English and ensure it matches the KNP before the recording so all the regions take the same name. Do not use any symbols or accents (Ò, È or / \ |). Files can be disconnected in a different language system and symbols are not allowed in some file systems.
Pickup Track Naming
- [PU]_[language code]_[character]
- For pickup recordings, place new clips on separate new tracks per character. The version number of the session is to be updated. (v2,v3,etc)
- According to the supplied dubbing script, the characters with the most lines will appear on separate tracks in hierarchical order. Highest line count tracks to lowest in descending order. In the case of multiple characters being recorded at the same time, provide the highest line counts on the uppermost tracks in the session as well.
- Clips on tracks (recordings) will have the character name(s).
- All Voice-over/Narration is to be placed on dedicated tracks, not mixed with dialogue character tracks.
- After recording, an editorial pass should take place. All clips are to be edited and faded in and out to remove extraneous noise possibly captured during recording.
- All dialogue to be compared against picture and confirmed for best possible sync to mouth movements and location of off-screen lines.
- To ensure all on-screen and off-screen dialogue has been recorded, a side-by-side comparison should be executed between the dubbed audio and the OV Dx stem.
- For Walla or Group recordings (background incidental voices and callouts), up to 8 tracks can be incorporated. Provide callouts separately, however general Walla can be comprised of multiple people recording on each channel, stacked up. (Example: 4 tracks of four people to get a crowd of 16) If more tracks are needed for larger crowds, use your best judgment.
- Examine the provided OV, Fully Filled M&E and Optionals to determine what coverage is necessary and if Walla recording is actually needed.
- When working on content with songs and new vocals are to replace the original voices, record the vocals on separate tracks, not on the spoken dialogue tracks of the same character.
- Avoid pitch correcting
- Avoid adding reverbs or any effects
- Match backup vocals and the number of vocal parts close to the original version
Preparation and Delivery
- Edited and cleaned raw dialogue, in sync, including fades
- The recordings will have no processing of any kind applied to the audio.
- If time expanding or contracting a clip, monitor extremely closely for artifacts. Clips will be rejected if induced artifacts from time stretching are detected. It’s possible that a clip will be amplified significantly during a mix, revealing induced anomalies.
- The session will contain audio tracks only.
- Mono tracks for dialogue recordings.
- Avoid EQ, Compression, Noise reduction, DeEssing, Futzing, Filtering, Reverb, Deverbing, or adding anything to the raw audio being captured or delivered.
- If working with optional tracks, keep tracks in original configuration (5.1, 5.0, LCR, Mono) and deliver with recorded raw tracks.
Do not include:
- Automation - All automated functions to control tracks/clips are not to be included. Clip Gain is the only exception.
- Master Tracks - The recording session should not contain any Master Tracks that control other tracks.
- Auxiliary Tracks - The Recording Session should not contain any Auxiliary buses or added signal routing and processing.
- Groups - clips or tracks should remain independent and not grouped to each other.
- Unused, empty or visual divider tracks of any kind. Remove all tracks that do not contain audio.
- Alternate Playlists - Tracks shall contain one playlist and nothing hidden.
- Inserts - Any plug-in on any track is to be removed.
- Sends - The recording session should nott contain any internal signal flow other than direct monitoring paths of the tracks.
- VCAs - The recording session should not have any type of track that controls other tracks.
- Source material audio - Printmasters, Video Files, DME Stems, Fully-filled M&E (if applicable)
- Using clip gain on clips is accepted, however please do not over gain tracks that have not been recorded correctly. If recording at a nominal level, less clip gaining will be needed.
- Capture recordings at a consistent average level. Attempt to match gain between characters and the day to day recordings of the same character. Keep the same microphones, preamp settings and microphone placements for all involved.
- To help create consistent recordings, document preamp settings, microphone type and mic placement for each character to help achieve repeatable quality for each part. Use this information for reference over the entire recording project.
- Mapping the multi-track Optional stems is welcome and sometimes required when provided to the recording partner (try to give the mixer a map of how you are intending to use the Optionals). To accomplish this, mute the section that is not intended to be incorporated into the dub.
- Please use best judgment for transitions from Options to new recordings. When needed, record efforts or dialogue that may be difficult to match.
- If there's a character who will not be dubbed, and the mixer will need to use the original dialogue, then please ensure that the original dialogue is either:
Available on the Options stem.
- Edited onto separate tracks for mixing (please do not add OV dialogue directly to a newly recorded studio track, even if it's the same character).
- If the new dub relies on small portions of the Original Version Dialogue, it is preferable to take these instances from the 5.1or 5.0 OV Dialogue stem, not from the 2.0 or mono Dialogue stem.
For reference, please check out this article on the Netflix Partner Help Center: Localized Dialogue Recording Recommendations
It is greatly appreciated that these guidelines be followed, as it allows smoother transitions into the next steps that happen after recording. Please inform your Netflix contact if there are any issues that should be addressed in regard to recording requirements and the Edit Dialogue Session deliverable.