But this is still strange… I have never used this option and have never had any of these strange noises. What I’m trying to say: The noises shouldn’t be related to this option being diabled as it hasn’t given other users, like me, any issues, right?
Ok 
Still some kind of wrong behaviour here? More so, if it should have SL+SR by what you says is the channel layout in MediaInfo… but why does Kodi interpret it this way? 
Does that debug report support OSMC sending a signal to all 8 speakers even if the signal is for less speakers?
Yes. That’s what the UNKNOWN1,UNKNOWN1 in channels 7 and 8 are I believe.
Phew. I thought my newbie posts were going to humiliate me. 
I still haven’t figured out how to post debug logs but I am happy to continue testing as I think my AV amp is accurate.
Why are 5.1ch PCM sources being sent to the receiver in a 7.1 container?
Seems if we fix that the whole discussion of side vs back isn’t necessary.
What’s the status of fixing that? Seems like it would be pretty easy as I’ve never noticed this in Kodi 17 on other platforms (Linux, Android, Windows). Or maybe I’m mistaken.
That echoes my thoughts. The side / back issue is probably caused by the 7.1 wrapper on a 5.1 signal and therefore confusing the AV.
Before reading the following please bear in mind I think OSMC is incredible and that I have no knowledge of programming since the days of Psion’s! These are just my thoughts based on how my AV is interacting with the output signal from the Vero4K
My thought process is that the choice of speaker set up in Kodi/OSMC is influencing the output signal whereas with passthrough it should not need to be told how many speakers are actually being used.
This data is only needed if the signal needs to be amended by the Vero. Most modern AV will take ANY signal and adjust it at their end to accommodate the speakers attached (the Vero will still need to retain the options of what passthrough signal can be understood by the AV, in it’s suboptions).
My blu-ray player is fairly advanced and only has the option of outputting stereo (down-mixed from original) or the original unaltered signal (passthrough) and my AV decodes every film perfectly.
Therefore I am not sure if OSMC needs to know the speaker set up at all as it would always be handled by the AV if it is anything more than a stereo set up. i.e. perhaps just offer Stereo down-mix or passthrough.
Hopefully that isn’t too much of a ramble…
Taking it into my hands to give a short reply here… I was among some others raising this issue as I’d love the Vero to pass on every audio format “as is”. The problem here is that - as I understood @sam_nazarko, correct me, if I’m not putting it right! - the board vendor AMLogic would have to add driver support for output of other channel formats than the ones used for PCM atm: stereo and 8-channel audio. So, yes, I’d agree that changing this to all possible channel layouts for PCM would solve the side vs. back surround issue, but it’s not easily done.
But it could help, if others post their concern, so this might become an issue big enough for the board vendor to change things.
To my knowledge and experience, other devices can’t output every PCM channel layout either (many Blu-ray players can play stereo and 5.1 surround audio files, but probably no 7.1 PCM from sources other than Blu-rays). The Vero 4k would probably be the first - which would make it even more awesome, if it could be implemented! 
If the Vero outputs everything “as is” in the source format (easy for bitstream as it doesn’t touch the audio anyways, but more difficult for PCM as we learned), this is absolutely true.
The only true alteration of the audio channels - the stereo-downmixing option - isn’t specifically meant for users using an AVR with HDMI input as those can normally just let their AVR do the downmixing, if needed for a physical stereo setup. It’s meant and useful for those using S/PDIF who don’t want the signal to be transcoded and becoming lossy or those using the stereo analogue out, I guess.
So, yes, the Vero normally shouldn’t need to know what the speaker setup is - that’s what the AVR is for -, but it seems to me, that this isn’t really the issue here. Down-mixing to stereo should work, pass-through should work and PCM output (atm for stereo and 8-channel) should work. It should work without noises or other problems for everybody - which is the issue at hand here. If that’s all working, then there is no need to know the physical speaker setup (when connected to an AVR), you’re right. But isn’t that a given? 
Earlier in this thread it was already discussed that the Vero can’t know whether a user is passing audio via S/PDIF or HDMI or analogue - so all settings have to be available all the time. That’s why, I guess, one can set the speaker layout even when using HDMI out connected to an AVR. But e.g. think of a TV with HDMI in. There are some TVs which can just handle 5.1 PCM. So, you have to be able to set the layout and used formats at all times…
Sorry, that this “short” reply has turned into a way too long monologue now! 
Perhaps a bit late with this info but the source I got my original 5.1 flac test file from also has 7.1 and even 6.1 files. GitHub - sfiera/flac-test-files: FLAC test files for multi-channel sound systems
In light of the blank channels that are added to a 5.1 PCM signal, this would be interesting: What does happen to the third surround speaker signal of a 6.1 file? Does the Vero expand that channel to both 7.1 back surround channels or is it just played on one of the back speakers with the other (missing 7.1 channel) being turned into a blank channel?
I’ve just copied those across and:
Center.flac - OK
Quad - OK and place Rear R & L correctly
Surround 5.1 - Problem - transmits Rear R & Rear L out of Side R & Side L. The speech is clear so 5.1 is definitely playing Rear R & L out of the side ones by mistake.
Surround 7.1 - OK but on tries 3&4 the Rear R & L also had white noise mixed in.
Also on 5.1 & 7.1 the Subwoofer symbol lit up appropriately on my AV when the signal was sent but no sound was actually heard from it.
My AV is displaying a 7.1 signal being generated in every case regardless of actual source i.e. the Vero is ‘padding’ the additional channels with something. My guess would be this is the cause of the remaining problems.
That I didn’t notice yet, but I’ll test again as soon as I’m back home at the end of this week.
Just to be on the same page: I don’t own a sub and therefore can’t test the LFE channel - which didn’t show up as a problem before either. Any others who can test this?
@Xanderzdad I don’t know, if you posted this before, but can you provide the test files you’re using? Or are you using @swrobel’s?
The white noise appeared randomly on plays 3 & 4 but not on plays 1, 2, 6 or 7 so perhaps disregard that for now as it could be a connection issue.
With regard to the LFE - my AV shows an active signal at the point a sound should be heard so perhaps it is just too quiet to be audible (the crossover point can affect this and in a real scenario the AV would split the signal between LFE and other channels depending on the frequency so possibly not an actual problem).
I used @swrobel’s files as linked to above.
I will also test some DTS and other test files later today (downloaded from [http://www.demo-world.eu/2d-demo-trailers-hd/]).
Ok… I’ll double-check anyways, just to be sure. I’ll report next weekend.
I’m homing in on two issues.
Issue 1 is vero ‘padding’ the unused channels. This is because the hardware has only two states: 2-ch or 8-ch. The metadata (‘audio infoframe’) in the hdmi (EAC) signal should report the actual number of channels to the AVR. I suspect that kodi/alsa/drivers is taking the 8 channels it gets from vero’s hardware and using that to construct the audio infoframe. I want to see if that’s the case and whether changing the audio infoframe fixes it. But that’s a lower priority than issue 2.
Issue 2 is spurious noises. I have noticed that when a track finishes by itself, it can leave the audio sink in a different state to that if the track is cancelled half-way through. On my system (sans AVR) that leads to corruption of the system (gui) sounds until the sound sink is re-set. Gui sounds (stereo) are being sent to an 8-ch sink. It could be a similar thing causing the ‘white noise mixed in’.
As for whether kodi/vero should care about the speaker layout, I think kodi reads the AVR capabilities from the EDID information. But that’s not much use - a 6-channel AVR will still report 8-channel support - and kodi doesn’t seem to take any notice of it. If I set channels to 2.0 in kodi, it will downmix. If I set 5.1 and send sound to my stereo-only TV, it will not downmix.
I’m pleased that I appear to have helped and agree with all your comments. I do think that both issues are equally important though and may even be linked.
Could padding a channel leave it in a state of limbo? Could this also cause the audio sink to be confused?
With regard to AV capabilities - I think Kodi should default to stereo/stereo down-mix and users with AV equipment can then choose passthrough to open up all the relevant channels to their AV.
The issue of selecting 2.0, 2.1, 5.1 etc is, I think, irrelevant to users as they would always have the ‘receiving’ equipment handle that side of things.
This is how Blu-ray and other sources tend to work.
Combining these to would give me the impression that it would be wise to have the audio menu this way:
Output selection (not in the way of that the others are deactivated - I understood that this is not possible and all outputs are always used simultaneously; it’s just meant as a set of standard settings profiles that give the best/most commonly normal settings for each output option one wants to use):
- HDMI → automatic selection of 7.1 channels → automatic selection of passthrough for all formats without any re-encoding
- S/PDIF → automatic selection of 2.0 channels (with a warning, if it’s set to anything higher than that) → automatic selection of DTS/Dolby passthrough as well as re-encoding of surround PCM
- Analogue → automatic selection of 2.0 channels → automatic selection of stereo down-mixing → automatic deactivation of any passthrough and re-encoding
I’m concluding this for these reasons:
- As @grahamh pointed out, most HDMI equipment should able to handle up to 7.1 channels even if the phyiscal loudspeaker setup is smaller - so why not simply set 7.1 channels as the standard layout for a HDMI profile?
- Some of the confusion about 5.1/7.1 PCM setting combined with S/PDIF (and therefore loss of audio information from at least 3 speakers) lead to quite some talk here and could be avoided in the future, if a S/PDIF settings profile would post a warning in case one sets the PCM channels to more than 2.0.
- Having standard profiles for each output option could reduce the amount of possible settings combinations that are used simply by giving users a guide to which settings are useful or probably not in the selected usecase scenario (some extensive talk here went into figuring out which settings the involved users had selected).
- If more experienced users want to use two output options simultaneously and therefore need to alter the settings of a profile they still can do that (or even create a new, custom profile?), but users with a “normal” use scenario don’t have to worry or can’t make too many mistakes.
Obviously, this doesn’t solve noise issues or possible loss of audio information (LFE channel?), but it would make the whole audio settings menu easier to understand and less likely to be the reason for settings being selected that don’t give the desired outcome.
Lets squash the bugs, first.
I agree and if my testing next weekend can help with that, even better 
But I wanted to put my ideas on the table as I think the settings layout for audio is not strictly a bug itself, but at least has the potential to let users think there is a bug just because the menu layout makes it quite likely for inexperienced users to combine settings in an unfavorable way…