Advanced Digital Video (Part 4)

Welcome back! Let’s wrap up with video codecs.

We found that there are a lot of codecs out there that can deliver nice looking video

but in doing so they're using up loads of unneeded bandwidth.

That’s why we created our own codec.

Q-SYS Shift is true to its name, in that it changes based off the content that’s being transported.

For example, content with relatively little motion or changes occurring can be transmitted with very little bandwidth.

However, when full motion video begins the shift codec can adjust and increase the bandwidth as needed

to send full motion video without any compression artifacts.

We started with a DCT (or Discrete Cosine Transform) as our base compression algorithm,

which is also used by the most popular codecs out there today,

along with both temporal and spacial compression to give you a codec that can adjust to any situation.

Q-SYS Shift also has a few other features like RFI and RTSP. Let's take a look at all of these in detail.

DCT is less intense with computations

and widely accepted when compared to DWT-based codecs which use discrete wavelet transport.

Spatial compression reduces the amount of data within a frame.

Think of it like a JPEG. You can take a RAW image from a camera

and the algorithm will reduce it without any negligible compression artifacts.

We do this with every frame in video.

We know that video is really just a sequence of still frames played at a rate

so our eyes cannot detect the change in frames.

The change between frames is where the temporal compression is used.

For example, the image below shows a few frames of a boomerang flying through the air.

Using temporal compression, we only send the data that is changing between frames.

Note that the background does not change so that data does not need to be resent.

Only the data about the boomerang.

What we get as a result of these two compression techniques is a native Q-SYS device

well suited for the conference room environment,

although the NV series is capable of being in any environment.

For example, a user has a slide deck or a spreadsheet which doesn’t change very often,

you can see significantly less bandwidth usage in that kind of content.

Compared to a JPEG 2000 codec, Q-SYS Shift has far less network bandwidth usage with static content.

However, Q-SYS Shift can easily be deployed where full motion video is required,

the codec will just shift or adjust based on the incoming content

and use the right amount of bandwidth that is required for full motion video.

RFI stands for Reference Frame Interval.

RFI can be a useful setting within Q-SYS that can be used to help with network recovery

with increased bandwidth or reduce bandwidth with a reduced network recovery.

In general the codec uses I-Frames, B-frames and P-Frames.

The Shift codec doesn’t exactly use these but the idea is the same.

The I-frame contains the full frame containing information for the full image.

The P-frame only encodes the changes between the current frame and the previous frame.

Like the P-Frame, the B-frame encodes the changes between the current frame and the previous frame,

but also encodes the changes between the current frame and the future frame.

So, how does this apply to RFI?

The RFI decides more or less how frequently we send a full frame of data.

If there are network errors, an I-frame will clean those up.

However, I-frames require more bandwidth than a P-frame or a B-frame.

So, how frequently should an I-frame be inserted? The lower the RFI, the more I-frames get inserted,

which means you get better error recovery at the expense of increased bandwidth.

The higher the RFI, the worse error recovery but and a decrease in bandwidth usage.

What does that look like? Well, here's a real-world test using a computer desktop, a static image, at 4K.

With an RFI of 200 we see a bandwidth usage of around 7 megabits per second.

That same content with an RFI of 30 will yield a bandwidth of approximately 40 Mbps.

If you open the NV Series component in Q-SYS Designer Software,

you’ll will find the RFI setting, with a range between 5 and 255 and a default setting of 30.

You will also note that you can set the bitrate here.

This is the maximum allowable bitrate that the algorithm will try and use.

This is equally applied across all HDMI inputs.

The NV Series is capable of encoding either a single 4K source or 3 simultaneous 1080p sources.

If you set the encoder to a bitrate of 600 Mbps and you set it to 4K60 max mode

then you can transmit a single 4K stream with a maximum bitrate of 600 Mbps.

Now, if the encoder is set to 1080p60 max mode

then you are able to share that 600 Mbps across all three inputs.

300 Mbps for 2 streams, 200 Mbsp for 3 streams. Math!

One last thing. If an encoder in 4K60 mode attempts to encode more than one 4K60 stream

then a hot plug event will be triggered and this will change all input EDIDs to 1080p,

thus allowing multiple streams at a resolution of 1080p or less.

So, if a 4K max mode is required

it is highly recommended that the encoder only wire a single AV stream to any decoder,

try and not wire multiple 4K sources on a single encoder.

If an encoder is in the 1080p60 max mode then there are no worries for a hot plug event,

you can stream all 3 inputs simultaneously without worry.

Real time streaming protocol or (RTSP) is another useful tool in establishing

and controlling the media during a streaming session.

The decoder requests the media from the source device and the source will send data as unicast or multicast.

Also known as RTSP pull streaming.

This means that video data is not broadcast over the network for no reason.

Encoders do not send video into the network until the decoder requests the video stream.

If all decoders are set to a graphic image then no video data is sent onto the network.

And that’s it for Digital Video. Phew, you made it! Congratulations!

Thanks for watching and we’ll see you next time.