Beam (Gen 2) disorted / metallic sound w/ TV (eARC) and Night Sound Mode

Just bumping this up:

Got a Beam Gen 2 last week and finding that I am getting the same issue after going through every setting on my LG Oled and working through with Sonos Support. Just received my new unit with RMA and still noticeable.

I find that it's really evident in Crowd and clapping scenes, very noticeable in the world cup games that are now on. When you go close to the speaker its definitely coming from the outermost speakers so must be how the sound is processed to the rear channels.

Have you had any further comms with the support team on a possible fix?

No further comments till now but Support asked me to test with the opical connection instead of HDMI (ARC).

Which I have not done till now. Maybe you can post your results of this test here, too? Mine will probably follow tomorrow.

That said, I can second what you said: The distortion is especially “evident” where there is clapping in the background.
However, silent scenes with a specific / clear and simple audio allow to capture it much more accurately.

When outmost speakers mean the “rear” channel, that could back the theory of bogous processing… as for me it is mostly evident in 2.0 channel audio, which by definition should not have any “rear”.

I've tested the optical adapter before. No change. 

Shnakey, as you seem to experience the exact same Issue that I have, do you have any specific source you can reproduce it? Till now I have only had German Live TV exposing this behaviour.

Ideally we have a Nexflix / Prime movie / specific audio channel that always exhibits this distortion. But I have not been able to identify till now.

I have it on all sources, apps etc. I also experience it with 2.1, 5.1 and Atmos content. It happens anytime there is surround sound effects, crowd noise and the likes.

OK. Is it more prominent for you also when you enable “Night Sound” ?

And just to double check - when you say “all sources”, you mean input anyway from TV (either HDMI or Optical), but all sorts of audio encodings, is that correct?

Haven't noticed night sound making it any worse. I get the metallic sound on all audio coming from the TV through the Beam, regardless of source or encoding. 

Yeah I tried the optical connector and it was the exact same too. 
 

I am getting the issue from all TV inputs via HDMI. Strangely if I put something on my phone and just airplay the audio to the beam it doesn’t have the same robotic reverb sound. 
 

Noticed it across all content from Atmos to stereo. It is really noticeable on the Elton John farewell show on Disney+ because of the crowd noise. Would be a good example to share with the sound engineers.

Hopefully they release a patch to fix it soon, but from seeing how long people have been mentioning it, maybe that’s a bit of wishful thinking. 

Additionally it maybe of help but I have noticed if you use speech enhancement it brings the effect down slightly and also if you use the height setting and go to minus 6 or lower it brings it down. 

Have a play with the height setting while watching a game of football in the world cup and you will notice the difference in the robotic crowd effect.

Hopefully this can help you out 👍
 

I can confirm that playing with the “height audio” level setting changes this bogous metallic effect.
I cannot however make it go away using height audio settings … just change its “presence”.

What is especially astonishing is, that height settings also changes the metallic effect with audio that is only stereo, 2.0 (e.g. Dolby Multichannel PCM 2.0).

Astonishing as 2.0 audio does not in itself carry anything that could be directly interpreted as “height” or even ambient audio.

This again points to some advanced audio processing that is going berserk.

It is really strange that there are so few people realizing this or are complaining. I am on the brink of returning my beam as this effect is extremely annoying.

Benny, did you also have the change to check with non-LG TVs? As mentioned earlier, I had a Panasonic before which also exhibited the issue… but I guess we could use some more data-points on this one.