Why Should I Care about Synchronized A/V?
As A/V projects become more extravagant, the ability to synchronize multiple channels of audio and video is becoming critical to the industry. For companies that are pushing the envelope of technology, it’s not enough to just throw an LCD panel on a wall and call it a day. Today’s modern A/V projects can involve entire areas covered in flat-panels, LED surfaces, or projection. With so many display devices side-by-side, it’s critical for the guest experience to ensure that the content being shown on these displays is high-quality and coherent across the display devices.
There are video wall solutions on market that can take a single source of video and scale it across multiple displays to form a larger coherent image. However, the problem with this approach is that display resolution is wasted by this scaling. To illustrate, imagine that you have a 2×2 configuration of 1080p (1920 x 1080 pixels) displays:
Combined, these displays have a total resolution of 3840 x 2160 pixels. However, if a 1080p (1920 x 1080) source video is used with a video wall controller, 75% of the available resolution is wasted when this material is scaled and stretched across all 4 displays. Making the 1080p source content bigger does not increase its resolution, so the more it’s stretched…the more display resolution is wasted. Imagine how much worse this problem would get if you had a 3×3 wall…9×9 wall…etc.
This is where the concept of synchronized video can help. Why sacrifice resolution by enlarging a single source of video when you can feed each display their own video source and synchronize them? Not only does this method spare you the expense of buying a video wall controller, it takes full advantage of the display resolution and provides a crisp coherent video presentation that is scalable.
The Key Concepts of Synchronization
Now that we’ve discussed the benefits, let’s talk about the important technical concepts involved synchronizing multiple channels of video and audio.
Concept #1 – Starting Playback Simultaneously
The most obvious and important aspect of synchronizing two or more devices is ensuring that they start at exactly the same time. That may sound simple, but many things are involved in achieving this. First, the playback devices must have a consistent reaction time when they are told to play. If their reaction time is unpredictable, it is impossible for any control system to issue commands that would result in a simultaneous start.
Even when you have video players that are capable of consistent reaction time, the timing in which these devices receive commands is very important. For best results, these commands need to be transmitted as simultaneously as possible. If using serial communications, this means the control system must be capable of transmitting commands simultaneously from multiple serial ports. When using Ethernet, individual commands need to be sent very quickly or a single broadcast packet must be transmitted.
The following graphic demonstrates why this is important:
Notice how the process of transmitting the play commands is not instantaneous. Each player must have adequate time to receive and process this command. Once this processing is complete, playback begins on the next Video Sync transition (i.e. the next frame).
If the playback commands are not received within the same frame by all players, they will not start at the same time. To further complicate matters, if the playback commands are transmitted too close to the ‘next’ frame edge, you drastically reduce the time each player has to receive and process the command. Some players may process the command slightly faster and start immediately, while other players might process the command slower and not start until the next frame edge.
To summarize, the only way to ensure a perfectly synchronous start is to use a control system that is aware of video sync timing. This control system would transmit the play commands immediately after a frame edge, allowing for an entire frame of time to transmit and process the playback command.
Concept #2 – Preventing “Drift”
Although Video players may be configured for the same video output format (i.e. 1080p @ 59.94Hz), it’s nearly impossible for separate video players to operate at the exact same frequency due to variations in electronic components, temperature, etc. For example, one video player might be operating at 59.940001Hz while another might be operating at 59.939999Hz. This might seem like an insignificant difference, but this discrepancy can cause the players to ‘drift’ from one another over extended periods of time.
The proper solution to this drifting problem is a feature called Genlock. Rather than multiple video players generating their own individual clocks, the video players slave to a single external reference clock. This external clock source is a commonly available device called a Video Sync Generator. Video players that have Genlock capability are designed to lock their A/V clocks directly to the output of the Sync Generator, ensuring that there are no discrepancies from player to player.
Aside from preventing ‘drift’ between players, operating in Genlock mode also ensures that the vertical sync (VSYNC) timing is aligned between players. What this means is that the players will decode and display video frames at exactly the same time.
What is “Network Sync” and how does it work?
One important thing to keep in mind here is that perfectly synchronized audio and video require high-frequency clocks to be in perfect sync with one another. To date, there are simply no players that can synchronize their A/V clocks with that kind of precision over an IP-based Ethernet connection.
Instead, what these “network synchronized” players do is periodically monitor each other’s playback position…usually in Hours, Minutes, Seconds, and Frames (HH:MM:SS.FF). Since these players do not share a common clock, they will inevitably drift from one another over time. When the drift becomes severe enough (usually a few frames or more), “network synchronized” players react by skipping or repeating frames to compensate for the drift.
Just to illustrate by example, let’s say we have two players that start playing at the same time; 00:00:00.00. Approximately 2 minutes later, Player #1’s current position is 00:02:00.01 and Player #2’s current position is 00:01:59.27. The players are about 4 frames out of sync, with Player #2 being 4 frames behind Player #1. In this situation, a “Network Synchronized” solution might force Player #2 to skip a few frames to catch up to Player #1.
The drawbacks of “Network Sync”
Unfortunately, there are some major drawbacks to this method of synchronization:
- Skipping or repeating frames causes hesitation and stuttering in playback that can be quite noticeable to the human eye.
- Even in the best of circumstances, synchronization can only be maintained within a range of 3-5 frames.
- Network traffic can interfere with the precision of this feature.
For many applications like 3D theaters, edge-blended projection, and seamless video walls…these drawbacks are deal-killers. Even a single frame of drift in these applications is blatantly obvious. Even for other applications like standard video walls, shows with multiple displays, museum exhibits, etc. these drawbacks are certainly not ideal for a great guest experience.
With so many drawbacks, why do manufacturers choose to implement “Network Sync”?
For the manufacturers that implement synchronization in their A/V products, many choose to implement “Network Sync” rather than proper genlock sync for a few key reasons:
- Ethernet is everywhere, and most A/V devices are connected via Ethernet anyway.
- Designing products that use Genlock to properly synchronize multiple A/V devices is difficult and requires sophisticated hardware designs and serious engineering resources.
- Cost savings. The electronics and effort required to implement true genlock can be costly to a manufacturer.