In this final instalment, lets look at the Mastering process for Dolby Atmos during Bombay Velvet. The Mastering and workflow help was given by Bharat Reddy and Dwarak Warrier from Dolby India. Dear friend and the resident Dolby Jedi Masters! This step is very little documented online and I will try to explain it as detailed as I can. It would help to have a look at the workflow diagram and the atmos format I have mentioned.

Mastering

One thing that is different in a Film mastering compared to mastering songs for audio is the meaning of the term itself. In the Mastering process for a CD or iTunes etc, great care is taken to ensure the different songs are similar in levels, the dynamics are clean, the song breaks are ok. There will be a master compressor, Limiter, EQ, etc, and many times, the song mix will sound different after mastering. None of this happens in a film mix master!

The reason I mention this is because I was asked quite a few times as to what mastering plugins do I use, what compression is used during the final master etc. The reason is that Film Sound is spread over a very long period and so the mix itself is done to sound the way it is intended to sound. There is no final overall glazing process that I use. I am not sure if that is the case worldwide but I would definitely think so. Any compressor or limiter would be in chain during mixing itself.

Preparing for Mastering

For the mastering, the sessions are prepared with all the start marks and first frame of action (Commonly called as FFOA which stands for the first frame of the reel where the movie starts. It is 48 frames after the beep) and the Last frame of Action (LFOA). Once these are done, the timecodes are entered into the RMU’s (Rendering and Mastering Unit) interface. The control is via a web interface on the RMU itself. Once done, the playback is started 15 seconds before the beep. the reason is to also have a buffer for the file while doing overlaps. This time, since we were running two systems in sync and didnt have an additional system to record the Atmos Printmaster, the final atmos mix was recorded only on the RMU. Simultaneously, the downmix is recorded onto a separate recorder in 7.1 from which we created the 5.1.

The Mastering Process

The Mastering Process involves the RMU recording the mix that we send to it via the 128 inputs on MADI. The process is done reel wise. Basically we run the mix and the RMU records.

The Transport section that you see is only active during mastering. The RMU requires LTC (Linear Time Code) for sync and mastering. Without that, the mastering wouldn’t trigger. The Web interface on the Mastering unit has the option to show where the files are being recorded to. It creates 48kHz and 24bit .wav files. The first 10 are labelled as the beds and the remaining are the objects. So, what is created after a recording pass is :

Ten mono .wav files, which make up the 9.1 bed
One .prm and one .wav file per object
One dub_out.rpl file

The dub_out.rpl file is basically an xml format file that has the details of all the associated wav files that are recorded. The .prm file contains the panner automation that we make on the plugin. This is also recorded and each object will have its associated prm file.

The Encoding

Once the Mastering is done, it has to be encoded into the DCP. The DCP that is made for atmos has some requirements. The Original fils comes from the DCP package that contains the 5.1 mix. Once that is made, the file issent to the dolby consultant with a KDM. KDM stands for Key Delivery Message. It is a license file that specifies the limitations of an encrypted file like when should it play, which theater it should play etc. The KDM is provided for a particular reason. When the consultant gets the DCP with the 5.1 embedded, they have to unwrap it, and add the atmos tracks into it. At this stage, there is one step that is done. Once the mastering is done, it has to be encoded into an MXF format. It is this MXF that is added into the DCP. But, the DCP is always made as a First half and second half each of which is an individual clip. How is it that the atmos that has been mastered reelwise converted into this?

There is a tool that is used which can match the Beep and also “stitch” the multiple reels into a single first half and second half audio. One of he biggest issue is usually sync. The Atmos encoding method allows for the tracks to be slipped as they have buffers to do so.

At this stage, there is one important thing that needs to be considered. It is called overlaps. An overlap is basically any music or sound that extends beyond a reel and needs to be taken into the next reels beginning. This is usually music or the tails. Now, the issue with the digital format is it is sample based. If there is a cut that is made on a sound that contains low frequency then you will hear a click oat the reel change. To prevent this usually the score is ended before the reel ends or has a highpass that can be run at the very end on the transition.

So, once the reels are stitched, the atmos tracks are added into the DCP. The DCP has a naming convention that is followed.

The DCP created by the Atmos encoder is a SMPTE standard. The usual standard followed by Qube or Scrabble is an interop standard although there is talk that the SMPTE will be the standard for DCPs in the future. You can read more about it here. Once all of this is done, we have an Atmos Package that can then be played back in theaters.

This concludes the entire workflow that was used during the mix in Bombay Velvet. I hope you had a great time reading it as I had mixing it and documenting it. I hope it was useful for all of you. Here’s wishing to great soundtracks and techniques and more importantly learning and sharing. Please do watch it in the theaters and let me know your thoughts.

Till next time, Enjoy!

-FM