How to reduce the LOUD (Sub Bass) of AC3 5.1 audio

[DruidCtba]

Guys, here's the thing: I have audio from a DVD show, Yanni-Tribute, in AC3 448 Kbps 5.1 channels, and my home theater system doesn't have a subwoofer, so my ONKYO Receiver transfers all the subbass to the front speakers, which they are LARGE, however, the bass is quite loud, and this bothers me, and I would like to know what I should do, just to reduce the sub-bass of it via software, VST plugin.

I tried two things here, which didn't give the result I expected:

1. I normalized the audio via Sound Forge 16, to 4db, 5db, 6db, and it didn't solve the problem, as the sound was very low, cut the bass a lot and lost the depth of the 5.1 audio (that sub-bass punch), in relation to the original sound.

2. Through a YouTube video, using the IZOTOPE RX10, I followed the YouTuber's guide to remove the popping sound, with the preset De-clip, Threshould -1,-1, Quality HIGH, Gain -2, -3, -5, -6, and again the sound was very low, cutting the bass a lot and losing the depth of the 5.1 audio (that sub-bass punch), in relation to the original sound.

Could anyone help me, telling me how I can reduce it a little, via software/preset/plugin, to get what I need here in this audio?

The audio height can be the same or a little lower, but without losing too much sub-bass, but reducing the sub-bass that are blown out, without losing the depth of the audio as happened with the two interventions I tried above.

Here is a part, from the original audio, where it is clear that the sound is blown out in sub-bass, perhaps in a system with a subwoofer, this does not happen, as in my case:

https://mega.nz/file/R7AT2KQQ#f3P2g1SVU2DYXfzmVv75MrF0XfWTLLe2lyG8wLEwVdE

Here original audio AC3:

https://mega.nz/file/8ipFDATI#QY1kJ4Up5Ufq9t9QdzmwbJSMEmIszrrifJgLmEz4q9g

Att.

Druid®.

[junh1024]

You can try to set the LFE level of your AVR to -10 or so https://support.onkyousa.com/hc/en-us/articles/9510518587540-TX-NR5100-Setup-Menu-Audio-Adjust .

[DruidCtba]

You can try to set the LFE level of your AVR to -10 or so https://support.onkyousa.com/hc/en-us/articles/9510518587540-TX-NR5100-Setup-Menu-Audio-Adjust .

Thank your good friend @junh1024, I will try it ;).

Att.

Druid®.

[Emulgator]

Throwing the Yanni.ac3 onto Audacity and soloing Channel 4:
This mix is not compliant from the beginning. LFE was derived from a already compressed and low-passed mono sum, which asks for collision.
Athough it may seem attractive, LFE had been conceived and was only meant for thunder, rumble etc., definitely not for musical instruments.
The Subwoofer may stand anywhere, and can never be assumed in phase with the main L/R speakers.
For this show LFE could have been used with heavy tympani, a swoosh, whatever, maybe supporting a extension of the base drum.
The same applies to center.
Center was meant for dialogue, and with music maybe for the main singer only, but definitely not for placing all strings there...
Oh well, and the surrounds do differ by spectral response and level...
On mixdown all that together will just tip into the ocean.
It can be saved, but one has to throw away the LFE, if needed, regenerate LFE from... maybe center ?
Match the surrounds, L/R need to be scrutinized too. It is work.

[tebasuna51]

I agree with Emulgator comments, maybe with a subwoofer speaker that sound not so bad, but when mix front and LFE to be played with only front speakers problems may appear.

BTW if you can try with a avisynth script like, make the mix by soft instead let your ONKYO do it:

FFAudioSource("C:\tmp\Yanni.ac3")

front = GetChannel(a, 1, 2)
center = GetChannel(a, 3)
lfe = GetChannel(a, 4, 4)
side = GetChannel(a, 5, 6)
mix = MixAudio(front, lfe, 0.8, 0.3)
lfe = GetChannel(a, 4).Amplify(0)
MergeChannels(mix, center, lfe, side)

Can be played with mpc-hc if your PC have a HDMI conexion to your ONKYO.
Try with different values for the mix (front, lfe, 0.8, 0.3)

[DruidCtba]

I agree with Emulgator comments, maybe with a subwoofer speaker that sound not so bad, but when mix front and LFE to be played with only front speakers problems may appear.

BTW if you can try with a avisynth script like, make the mix by soft instead let your ONKYO do it:



Can be played with mpc-hc if your PC have a HDMI conexion to your ONKYO.
Try with different values for the mix (front, lfe, 0.8, 0.3)

I'll try here, do some tests friend @tebasuna5, thanks for the answer :).

And yes, I also agree with friend @Emulgator, thanks for the prompt response.

I use the DUNE HD SOLO 4K media player and NVIDIA SHIELD TV 4K, both from 2017.

Att.

Druid®.

[hello_hello]

It's late at night here and I can't make much noise so I haven't listened for popping sounds, but it seems to sound okay with the LFE channel removed.
There's stereo and 5ch versions in this zip file. Yanni.zip (https://files.videohelp.com/u/210984/Yanni.zip) (461 MB).

[mannequin80]

It's late at night here and I can't make much noise so I haven't listened for popping sounds, but it seems to sound okay with the LFE channel removed.


sorry for the slight off topic but i don't want to start a new thread since this issue has been discussed before ad nauseam (to which you contributed aplenty hence my question) but it's still difficult to find a good answer:

is something like "dynaudnorm f=75 g=11" still the "best" way to normalize (harmonize) a typical multichannel movie stream (especially the dialogue) and still retain some dynamic range? i've done some brief experiments in the past and 75/11 values seemed the best to keep "volume pumping" at hard to detect levels at least (however, in certain scenarios, when dialogue is on top of the background music and has pauses, these settings may cause very quick fade in/out effect and certain (usually music) parts to be very loud which i couldn't find a way to mitigate). i still don't quite understand why it would work better than let's say 250/5 but like i said, my experiments were really brief.

are some other dynaudnorm parameters (like "b" and "s") worth exploring or they won't make much difference? or, perhaps, there's a better tool these days to achieve "better" results - "natural" sounding dialogue that still retains some dynamic range with the rest of the channels without introducing of "pumping"?

thanks!

[FranceBB]

is something like "dynaudnorm f=75 g=11" still the "best" way to normalize (harmonize) a typical multichannel movie stream (especially the dialogue) and still retain some dynamic range?

For personal / non important stuff, I generally go with

-af loudnorm=I=-24:LRA=12:tp=-2

which shouldn't create too much fluctuation (note that -24 is the Italian AGCOM standard, while the rest of the EU is probably using the EBU R128 standard at -23).
This, however, won't do much in terms of preserving dialogues, which can bring to the annoying situation in which you're watching a movie and you constantly have to raise the volume when people are speaking and lowering it when there are action scenes ('cause you don't want the neighbors knocking at your door).
For professional stuff at work I'm using the Dolby DP600 (paid product) which has the ability to treat the dialogues differently regardless of whether it's a 5.1 stream or a 2.0 stream, however since that's no longer supported we're slowly moving to Emotion Engine (another paid product) which is using the Dolby SDK under the hood to do the very same thing.
This means that in a Dolby 5.1 stream, the dialogues will always be treated differently and the center channel will be prioritized, while for stereo mixes, some frequencies will be attenuated less than others so that dialogues should, in theory, be preserved.
This has been an issue over the last few years in which so many people reached out to me complaining about the very same thing, namely the fact that they always had to raise and lower the volume of the movies, so I thought it was better to tackle this from the very beginning in the Content Acquisition step. I mean, we have to bring everything to -24 to comply with AGCOM (the Italian regulatory body) so why not fixing the "issue" in the LC step anyway? :P

That being said, it would be nice to have something like that in the loudnorm filter as well instead of having to rely on Dolby proprietary stuff... :(

[richardpl]

Dynamic processing within loudnorm is not following any standard, creates clipping audio, and distorted audio and at same time is very inefficient code and wasted CPU power. But as usual people deserve what they wish for.