I'm using my 976 for mostly audio, but occasional video watching too. My source is an HTPC, using HDMI out into the 976, which connects to a very modest 40" video display. The 976 replaced a dead Emotiva UMC-200, which did not do this with an otherwise identical setup.

What happens, briefly, is that at the beginning of each track, the 976 introduces an audio "fade-in" which sometimes blots out the very beginning of the track. It's the kind of thing that, once one notices it, becomes a distraction.

Here's the best example I have found so far. In the Hendrix Band of Gypsys album, the song Changes begins with Hendrix saying, "Buddy Miles' gonna do a thing he wrote called...". When played back through the 976, it sounds like "gonna do a thing he wrote called...", that is, with, "Buddy Miles" cut off at the beginning. It's not the playback software doing this, as the UMC-200 I had previously didn't do it, and neither does a USB DAC hooked up to the system using the same playback software. Also, I have the 976 HDMI interface set to the default Windows sound driver. In this configuration, the sample rate and bit depth are set in the Windows Control Panel and never change. So the 976 is seeing a constant sample rate and bit depth throughout each listening session.

The HTPC is using an AMD A8-5500 APU with 8 GB of RAM, and all the latest AMD video, HDMI and chipset drivers are installed.

Another thing that concerns me in this regard is the possibility of misbehavior when doing REW measurements. If the fade-in occurs at the beginning of the REW sweep, there's the potential for the frequency response computed by REW to be different than it actually is at the low-frequency end. That is, the fade-in could appear to REW to be a roll-off of low frequencies that's not really there, since the REW sweep starts out at a somewhat lower frequency than the lowest frequency to be measured.

Search the threads for this topic. I have seen it before and it has to do with the silence between the tracks and the HDMI handshake. The silence tricks the processor into thinking the music is over and the has to reacquire the signal when the next track comes on resulting in a few seconds loss of data.

Thanks. When I first noticed this, I decided to check the behavior with gapless playback of consecutive tracks. It works perfectly. Then I went back and re-ran the Hendrix test track I mentioned earlier, this time including the prior track in the playlist. There is apparently silence between these tracks, so the loss of the early information in the second track still happened.

I quickly discovered this issue and sent a bug repport. I was more or less told "that's not a bug, it's a feature"... Then, I was probably the first person to discuss the issue here, and most people had a meh reaction, as if it's a trivial situation.

So I have not much hope for a solution!

Thanks Eric. I've removed it from my system. I'll report it elsewhere.

Is this behavior only present with an HDMI connection? For example, if you use one of the optical inputs does it happen?

I chimed in on this topic in the original thread. I had tried a variety of discs and files, to see for myself what others were seeing. For me, it happened on some tracks and not on others. Some of the discs I tried were ones that I ripped myself. These did not produce the fade in.
Some file tracks I tried, from a flash drive plugged into my Oppo, exhibited a fade in, but not others on the same playlist. Whenever it occurred, it was a few mS at best, and resulted in not much more than part of the first note.
None of the tracks ever missed an entire first word or two. In my experience, this "fade in" was too quick to miss more than a part of this first note, and this was about 5% of the tracks.

The common thing I noticed was this happened whenever there was a loss of audio in between tracks and my 976
had to reacquire the audio signal.

I admit, I mostly use my 976 for movies and TV.
If I were predominantly using for music, and I turned off the lights, closed my eyes and focused on specific music experiences, I'd probably connect to one of the pure analog inputs that would eliminate this very small % of fade-ins.

Does anyone know if the lip sync delay is in play when just an audio source is being used. I ask because, if it is, the delay could affect the ability for the processor to sense the signal since it is being delayed by some amount. Just a shot in the dark.

All of the files in my collection are FLAC, ripped from CDs using EAC, with read offset determined using the AccurateRip plugin.

I brought that Hendrix track into Audacity and determined that roughly the first 600 msec was being removed from the beginning of the song.

Some albums don't do it at all, including gapless playback, which works perfectly. I didn't try to figure out percentages of albums that show a problem, but it's a lot more than 5 percent.

I didn't try the optical input.

600 mSec is a long time for what we're talking about. I've not had that experience in any track or file I've tried. Any delay should be consistent with the time it takes to acquire the signal. That time should be the same. But, my experience has only been noticeable 5% of the time, and that was when I was looking for it.