LFE 10dB attenuation and boost

[Beber]

Hey, guys.

Is this common knowledge? While discussing multichannel audio editing in DaVinci Resolve with Gemini, this came up:

When editing audio tracks, there's an invisible digital trap regarding the LFE channel. If you do not correct it during your editing phase, your final export will lack punch, and your subwoofer will sound completely flat. Here is why it happens and how to fix it.

1. Why is there a -10dB difference?
- In movie theater / mixing stages: to prevent extreme sub-bass frequencies from clipping and distorting the digital master track, the industry standard (Dolby and DTS) dictates that the LFE channel is recorded on the disc 10dB lower than the other 5 channels.
- On playback: your home theater AV Receiver automatically detects the movie track and boosts the subwoofer output by +10dB to restore the correct theatrical balance. Everything sound perfectly heavy.

2. The trap in video editors (DaVinci Resolve, Premiere, etc.)
When you extract a DTS-HDMA track to uncompressed multichannel PCM (WAV) using tools like Shutter Encoder or eac3to, you get raw digital audio streams. When you import these 6 separated mono tracks into your timeline, your video editing software does not know track 4 is an LFE track. It treats it like a regular audio track. If you edit, crossfade and export your project as-is, that 10dB attenuation becomes permanent in your new file. Your final AV Receiver will not boost it, and your mix will lose its cinematic rumble.

So basically, when editing a track you ripped from a DVD or Blu-ray that you converted to PCM to work on, you must take into account that you'll have to boost the LFE channel by +10dB.

[LucasGodzilla]

2. The trap in video editors (DaVinci Resolve, Premiere, etc.)
When you extract a DTS-HDMA track to uncompressed multichannel PCM (WAV) using tools like Shutter Encoder or eac3to, you get raw digital audio streams. When you import these 6 separated mono tracks into your timeline, your video editing software does not know track 4 is an LFE track. It treats it like a regular audio track. If you edit, crossfade and export your project as-is, that 10dB attenuation becomes permanent in your new file. Your final AV Receiver will not boost it, and your mix will lose its cinematic rumble.

So basically, when editing a track you ripped from a DVD or Blu-ray that you converted to PCM to work on, you must take into account that you'll have to boost the LFE channel by +10dB.

Technically from a metadata level, if you re-export the mix in a format that interprets an LFE channel correctly, then it should be fine when playing the audio back through a proper decoder? This would apply to PCM, DTS-HD, TrueHD, etc (just about any audio format frankly).

However, that said, I personally have been operating on the assumption that if you are working on surround downmixes, then technically that's where the attenuation comes into play.

For example: Since Dolby Atmos drops the LFE entirely in a binaural bounce, I just mix it in as a front-center channel / object and attenuate all the other channels by -10 dB to avoid clipping.

[Beber]

The examples I gave Gemini that made it came up with this were my intentions to edit the DTS HDMA tracks of Sin City to create the 5.1 mix of my extended cut project, and my intent to make a hybrid cut of the CDS mix and the DTS HDMA of T2 Special Edition. So we're dealing with 5.1, and I'll try to work on those in DaVinci Resolve 15 free version which doesn't go beyond 5.1, I believe. Apparently, according to Gemini, my workflow that consists in transcoding these original tracks to PCM so that I can work on them in Resolve requires to boost the LFE channel by +10dB for the reasons explained above. The idea being to export my work in PCM 5.1 and potentially transcode that to FLAC 5.1 to gain space.

[Beber]

Okay, so I just had Gemini elaborate on the subject: this whole thing only applies if your target is PCM, FLAC, AAC and such. However if your target is Dolby or DTS, proprietary theatrical codecs, the encoders know all this and are dealing with it on their own, so you definitely don't need and you definitely should not do it if you go for, say, a DTS-HDMA final rendering, or you'll get a double-boosted LFE output on your AV Receiver. That was an important precision. Lucas was right when mentionning metadata. It's all about flags that dictates the AV Receiver to boost a Dolby or DTS LFE channel, flags that are not part of PCM, FLAC, AAC and such, hence the AV Reveiver not triggering an automatic boost on the LFE and the necessity to apply it on your own in the editing process. So to decide to apply it or not, you need to know what final codec your audio track will be rendered in.