If you Google lip-sync error correction you will find it is "always" present because there is nothing within the signals to define when they were "ever" in sync. We usually ignore it by subconsciously looking away from the faces and lip movement because it is impossible in the "real" world and our brains can not reconcile the contradiction of reality we would otherwise experience. This has a negative impact on our perception of the characters even when we don't consciously notice the problem.
If this interests you Google "Reeves and Voelker Audio Asynchrony" and I'm sure you will find many links to a study done at Stanford over 20 years ago which documented lip-sync error makes us feel the characters are more agitated, less truthful, less successful, etc. The same way we feel about people who don't make eye contact with us but ironically "we" viewers are the ones looking away from the characters' lips and faces.
The Stanford study reported most people don't consciously notice lip-sync error under 42 ms but the negative impact is still there.
But once we do notice the problem and it enters our conscious mind we start seeing tiny lip-sync error orders of magnitude less than we were previously able to ignore. Years ago my TV had about 64 ms video delay and I had never consciously noticed it but after encountering larger lip-sync error my defense mechanism of looking askance went away and it began to bother me on every programme I watched.
I bought a Felston DD340 (this was over 10 years ago and I moved up to their DD540 and more recently to their DD740) which allows me to tweak the audio delay in 1 ms steps for perfect sync. Several other companies make similar dedicated lip-sync correction units (Alchemy2, Primare's delay box, etc.). All of them give the user a remote control to adjust audio delay in 1 ms steps while watching talking heads without plastering menus all over the screen.
Some TV's (not Sony) give the user control over audio delay but it is usually in such huge increments (like 10 or 20 ms) you really can't ever get it to seem perfect. You really need 1 ms steps. My DD740 even offers a fine mode with 1/3 ms steps but I never use that mode.
In any event my experience is that there is no way to achieve perfect lip-sync without one of these devices. I say that because it varies so much from programme to programme a fixed delay won't really help. In my case I sometimes have to reduce the audio delay below my TV's video delay (about 60 ms as I don't have a 4K set which would be greater) which means I am using some of the screen's video delay to offset arriving udio delay - which is rare but does happen occassionally. But usually I have to add 70 to 170 ms extra audio delay to offset the varying video delay from programme to programme. I even had to add over 300 ms one time.
One thing to note: Sony adds audio delay equal to the screen's video delay to both the internal speakers as well as to the optical audio output from the TV so if you use the optical output to your AV receiver or soundbar you will not have the negative adjustment range of reducing audio delay below your screen delay using that video delay to cancel arriving audio delay. Again, rare, but it does happen. Sony assumes the arriving signals are in sync but that is rarely the case.
But some spaeker systems add audio delay which can push lip-sync error above your threshold of recognition. Sonos for example adds about 30 ms which normally isn't recognizable but may be on top of the audio delay Sony adds. To correct that just take each source's audio directly to your sound system and bypass the TV. That way in conjunction with a variable audio delay you have a little variable video delay equal to your TV's video delay. In my case with a screen delay of 60 ms I have an adjustment range of -60 to +620 ms (since my DD740 offers 680 ms audio delay).
Devialet speakers add about 150 ms audio delay which can't be corrected by adding more audio delay so I wouldn't use them for AV application unless perhaps with a video projector which I am told often have over 300 ms video delay.
The insidious thing about lip-sync error is that it normally masks itself (our defense mechanism of looking away) but when noticed that subconscious defense vanishes and we start consciuously focusing on the lips as never before. We aren't even aware of this change in our own sensitivity and this means we ususlly still see the problem even after correcting the error below the level of error we were previously oblivious to.
