Audiovisual Sync Thresholds

[Curvature]

https://www.researchgate.net/publication/313774497_Monitoring_of_audio_visual_quality_by_key_indicators_Detection_of_selected_audio_and_audiovisual_artefacts

I use a computer connected to my TV for all video and music. There is a slight delay with movies and the like, but it's short enough that if decide to ignore it it stops being annoying.

That said my audio pipeline is very simple. Some PEQ through EQ APO/Peace and out to an RME DAC.

When I looked it up I was surprised that the thresholds for desync were so generous. Quote explaining the chart above:

"Figure 10 is a graphical representation of different audio/video delay and lip sync thresholds of detectability as identified by several standard bodies and independent studies. The thresholds used for the lip sync artefact in MOAVI are set to 100 ms when the audio is delayed with respect to video and 140 ms when video is delayed versus the audio. These thresholds are based on research work by Steinmetz on human perception of jitter and media synchronization."

Steinmetz article mentioned: https://ieeexplore.ieee.org/document/481694

 

[kemmler3D]

View attachment 460119

https://www.researchgate.net/publication/313774497_Monitoring_of_audio_visual_quality_by_key_indicators_Detection_of_selected_audio_and_audiovisual_artefacts

I use a computer connected to my TV for all video and music. There is a slight delay with movies and the like, but it's short enough that if decide to ignore it it stops being annoying.

That said my audio pipeline is very simple. Some PEQ through EQ APO/Peace and out to an RME DAC.

When I looked it up I was surprised that the thresholds for desync were so generous. Quote explaining the paper above:

"Figure 10 is a graphical representation of different audio/video delay and lip sync thresholds of detectability as identified by several standard bodies and independent studies. The thresholds used for the lip sync artefact in MOAVI are set to 100 ms when the audio is delayed with respect to video and 140 ms when video is delayed versus the audio. These thresholds are based on research work by Steinmetz on human perception of jitter and media synchronization."

Steinmetz article mentioned: https://ieeexplore.ieee.org/document/481694

I looked into this while working on the firmware for some Bluetooth headphones. (I was the product guy, not engineering) And I had the same reaction, it's surprisingly loose. For me personally the delay wasn't blatantly objectionable until it hit 200ms.

Maybe this makes plenty of sense, though. Sound only travels a bit further than 1 foot per ms, roughly. People often need to talk to each other from pretty far away. So if the brain was getting confused when someone is speaking from 50 feet it wouldn't be practical. You know, evolution etc etc.

 

[Palladium]

I use LDAC for watching FPS games, the syncing of the gunshot sounds with the muzzle flashes are good enough that I couldn't tell a difference vs wired.

IIRC that's additional 100ms latency over wired.

 

[Cbdb2]

Dialogue/SFX editors can usually spot audio thats a frame out, 30ms. I remember helping to hand sync all the dialogue in a movie once that lost its time stamp. You had to find the right take for each shot then sync it by ear. Great training for recording ADR.

 

[kemmler3D]

Dialogue/SFX editors can usually spot audio thats a frame out, 30ms. I remember helping to hand sync all the dialogue in a movie once that lost its time stamp. You had to find the right take for each shot then sync it by ear. Great training for recording ADR.

I think that's definitely true and I think a lot of people can identify the gap if they sit there scrubbing back and forth and paying close attention. I would say this is the difference between detectable delay and objectionable delay, the range between being able to notice it and not being able to ignore it.

 

[Dougey_Jones]

What's funny is, people download/pirate content for all sorts of reasons, and the reason I do it is because if I have the content on my NAS, I can play it with KODI and I can adjust the audio delay in 5ms increments until it's correct. I'd say 90% of the stuff I download is available to me on the various streaming subscriptions that I pay for, my OCD just makes it so that I can't enjoy anything if there are sync issues.