Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[rxp]

I've got my library setup to calc R128. I actually use passthrough so I use the LU adjustment to just offset the volume knob on my reciever. e.g. if it reports -2 I just put the master volume to -2 on my AVR. This works fine on Dolby tracks. I've calibrated my AVR to 83db at 0 volume.

However the calculations seem really off for DTS. I have some very loud DTS tracks but they're reporting similar offsets to the Atmos movies I have which are much lower in volume. Upon playback in passthrough it's very clearly wrong - the DTS is far louder. The same is true decoding to LPCM on a PC or Infuse on the Apple TV. Is there a known issue with calculating DTS R128 figures? The calcs it seems to make for Dolby tracks seems bang on. But for DTS it's really off. For example in a Narcos episode yesterday it calculated -2LU and I had the volume on my AVR to -12 to hit similar volumes (by ear and SPL meter on dialogue)

Also what happens when a file has multiple audio tracks e.g. DTS/AC3 or those with commentaries. What gets analysed? How can I verify that?

I might have figured out (https://hometheaterhifi.com/volume_7_2/feature-article-dialog-normalization-6-2000.html) the DTS issue. Seems that receivers playback DTS tracks at +4db. Is there a custom field I can calculate for when a movie has a DTS track to minus 4 from the Volume R128 field?

Edit: Also Oppenheimer's DTS track was bang on the suggestion so I feel that the Narcos DTS track is just a little off for R128's algorithm

[Hendrik]

There can be no guarantees that the computed loudness values would match an external decoder. We decode the audio and analyze the volume of the decoded audio, at that point it makes no difference what format the audio once was - its now all just PCM. If you play the decoded audio through MC, you would always get a consistent experience.
We do not know what an external decoder might do. They might apply different volume levels to different formats for unknown reasons, or their decoders might process dynamic-range metadata differently.

[rxp]

Thanks - that does make sense.

I'll try with playback on JRiver and report back.

Do you know what the implementation of R128 does with multiple audio tracks?