The more data points in various sources usually helps the development team pinpoint the issues. So I would submit diagnostic codes and files.
Please look into this.
I’ve rerun the original test with:
- tvOS 16.4.1 (20L498) or a ATV 4K A2169
- Infuse Pro 7.5.2 (4410)
I report no difference.
The following is a screenshot of an Audacity capture of my original Interstellar clip on the latest Infuse version vs the FFmpeg extract.
The setup was:
- Apple TV → HDfury DIVA → Monitor
- HDfury SPDIF output → USB capture device with SPDIF input → Audacity
I’ve begun to use VLC to play movies because the audio through Infuse on a lot of my Blu-Ray transfers is distorted.
@james, any updates that you can provide?
Edit: modified the screenshot to include a FFmpeg transcoded copy.
I mean it shouldn’t be so damn hard to setup a new audio engine or at least try to fix the current algorithm. Even if the distortion problem is fixed it still doesn’t sound very good regarding soundstage and dynamic range. The center channel is waay to loud as already mentioned. Mrmc does dm really well so I wish infuse could implement sth similar like this. In mrmc you can also choose how many channels you want to downmix e.g. 7.1 to stereo or 7.1 to 2.1, 3.1, etc…
I wonder if it is possible to use the official dolby and dts decoders for the downmixing? Is this doable or is it limited by the apple tv’s hardware capabilities?
Using Bluetooth stereo headphones, I just tested and can reproduce the same thing with my own copy of The Dark Knight Rises and also with many other movies. Knowing that it’s there now, the audio distortion is very easy to notice.
As a follow-up to my previous post, I did more testing. And I must say this issue is a real showstopper for me.
If I’m being honest, these days I end up watching content on my Apple TV wearing stereo headphones at least 80% of the time. So good stereo sound is a critical feature for me. But here, the very regular crackling/distortion and the general quality of the stereo downmix is seriously impacting the experience.
In the last few days, I was testing Infuse to see if it could become my main player for the content on my Plex server (mostly for my Apple TV, but also for my iPad and Macbook Pro). And I must say I am generally very impressed. Infuse seems to be very stable and gives a noticeably better picture than the ATV Plex player in a number of contexts (1080p upscaling, HDR/DV). Even if Infuse is missing some features that are important to me personally (displaying local extras for movies and TV shows, managed user accounts for family members), I was seriously considering completely replacing the regular Plex player with Infuse, it’s generally very impressively good.
But this audio issue completely stopped me in my tracks. As it is right now, Infuse just doesn’t perform well at all while wearing headphones. The quality of the stereo downmix is very distracting. I unfortunately just can’t consider using Infuse because of that.
But I really hope this issue is being investigated by the development team and will be fixed soon. I think I would run back to Infuse as soon as it is.
I have a 2.3 setup and I’ve resorted to mixing all the audio in my Plex collection ahead of time because Infuse butchers the downmix which is kinda baffling to me as downmix matrix coefficients are standardized across all audio formats with the occasional offset from proprietary format metadata like dialogue normalization and center/surround mix level.
I would let my TV do the downmix instead since it could actually downmix properly but it does not mix in LFE (which is standard so I can’t fault them too much) which prevents my setup from reaching its fullest potential of reference level bass. This is really one of those places where users need to be provided an option on how much LFE to mix in for people with more expensive setups.
We have been able to reproduce this issue here, and are looking into a solution.
Thanks for your patience, everyone.
Might be worth to engage with some users that maybe have expertise in the field of audio processing, since downmixing comes in a lot of flavors -just as someone mentioned above (in reference to the LFE), but much more than that- and maybe even include a small upgrade on top of the fix.
The Batman, entering and first fight in the club, horrible distortion on all audio tracks.
7.1/5.1 → 2.0 Downmix.
From the apps I’ve tried on the Apple TV, MrMC seems to do the downmix well; it has a bunch of audio options for customization also. Of course, Infuse tends to simplify these kind of options (and that’s fine for most users), but I’d suggest to at least add a toggle similar to “Maintain original volume on downmix”, defaulted to on.
Can MrMC handle dolby vision content?
From what I gathered online, it seems capable of playing Dolby Vision, but can’t test it myself, my TV is a Panasonic OLED with HDR10+ (and it plays that) but no Dolby Vision.
You can try some of your files with the free version of the app.
I’m not advertising the app, btw, I much prefer Infuse’s design and the video decoder seems more robust (I’ve got the green/purple color bug playing some DV files on MrMC), just giving them a reference point for audio reproduction.
Are there any solution provided in the 7.5.3 and 7.5.4 updates?
It is planned for 7.5.5
Erratum: it is planned for 7.5.6.
Actually noticed this for the first time on an episode of the Mandalorian last night.
I did have the AppleTV audio boost set to 200% as usual, so resetting that to 100% removed the clipping and played back without issue thereafter.
(I’m using an older version of Infuse on that AppleTV because I haven’t the internet to update them currently.)
We have a number of changes in this area available today for beta testing.
Please drop me a PM if you are interested in testing.
Hey guys, I am testing 7.5.7 beta 2 and the team added an audio normalization option that seems to do the trick.
As I said in the beta topic, the volume lowering is very aggressive, even more so than MrMC. That should correct the problem even for the more picky systems.
Heyy everyone! I was able to test the new downmixing settings in 7.5.7 and it sounds awesome so far!! The clipping is completely gone and the dynamic range is restored. Great Job👏 One question: Is the LFE channel mixed in at 100% when enabled in settings? Bc I noticed the low end doesn’t have as much “punch” as the mrmc downmix when comparing the two.