Lots of confusion out there: High Quality Video, HD Video, HD Voice, High Definition Audio? What do these terms mean? And what does High Quality Video mean when it comes to Skype Video calling? “Well, my configuration should work”, comes from both other vendors and some end users. Did a bit of research to try to find a baseline against which all these claims could be measured.
HD Video generally refers to video with a minimum 1280 x 720 resolution (720p). Common to all High ?? Video descriptions is the transmission of video at 25 – 30 frames per second (fps). “HD VoIP” was the subject of an excellent session at Fall VON and generally refers to audio transmission of voice covering at least 8kHz audio — primary range of the human ear (sometimes called wide band audio); Intel appears to have taken over the term High Definition Audio to describe some of their audio-related technology.
The aim of this post is to provide some background that allow us to move forward in testing High Quality Video. Without a standard to be achieved we are seeing lots of claims about various webcams and the ability to transmit video. In this post I will set some basic benchmarks to help identify High Quality Video when it happens. In future posts I will cover some of my experiences attempting to have Skype-based conversations using High Quality Video.
My first question to Jonathan Christensen, Skype’s GM for Audio and Voice, was “What defines High Quality Video?” Summing up his response it’s the sustained transmission of 640 x 480 (VGA) video resolution at 24 – 30 fps over a minimum 384 kbps connection. And the key word here is “sustained” as in “carrying on a reasonable length conversation” user experience. Previously one could get burst of VGA quality but sustaining it was not practical due to non-optimization of all the technology pieces that need to come together.
Skype initially considered a higher resolution involving the 1280 x 720 “true HD” Video spec but soon saw that as unrealistic within today’s network conditions and video technology limitations. According to Jonathan, his Skype Audio/Video engineering team did a significant amount of research looking at factors such as webcam technology, codecs, webcam drivers, end user configurations, network conditions (including user end point configurations), image sharpness, CPU utilization, and surrounding light sensitivity in coming up with a video calling product specification. And they had to settle on the type of video content that would be most commonly used: “talking heads” in a home or office desktop environment. (No Discovery Channel panoramic scenery video or high speed action sports such as hockey or football here.)
Basic requirements: many have heard about the Sender (Capture) requirements:
- Intel Dual Core (or CoreDuo) processor
- Logitech QuickCams with Carl-Zeiss optics (9000 Pro, Pro for Notebooks)
- QuickCam software drivers version 11.5
- Skype 3.6
- Internet Connection: (an industry standard minimum requirement for video transmission)
At the Recipient (Rendering) end these requirements are relaxed to:
- Fast processor (but not necessarily Dual Core)
- Skype 3.6
- Internet Connection: minimum 384 kbps upload and download speed
So how do you identify High Quality Video is being captured and transmitted by the sender and received. and rendered by the recipient? Simply put, Skype injects a logo into both the active call window’s video indicator and the recipient’s video picture, whether within the active call tab, a 640 x 480 Window or full screen. This logo only appears in situations where either the Sender is sending 640 x 480 @ 24-30 fps (above) or the recipient is receiving 640 x 480 @ 24-30 fps (left). Drop below 24 fps and the logo goes away; come back above an it appears. To add to the confusion, the Sender can be sending High Quality Video, yet the recipient does not necessarily see the logo unless certain conditions are met.
A few other notes:
- In my testing I have come across a couple of situations where High Quality Video was being identifiably received under the “Recipient” conditions (for instance an Inspiron 6000 with a single 1.6 GHZ Pentium M processor, shown above, and Phil’s 64-bit AMD processor PC, both of which had Skype 3.6 installed).
- Note that the Recipient can launch not only a Chat session from the video within the Active Call window but can also “capture” the picture as a still photo (camera icon).
- If a recipient is using an earlier version of Skype (3.5 or earlier), they may be seeing High Quality Video but there is no indication via the logo. The only way to check is via “displaying the technical call info”1.
- It can take about 45 to 120 seconds for a High Quality Video connection to be established.. Using Skype’s “technical call info” feature1, you can actually observe the Send fps rate ramp up; High Quality Video identification only occurs above 24 fps (and a 640 x 480 frame size). So there is obviously some initial handshaking/negotiation activity generated on launching a video call.
- It is worthwhile checking your own Internet connection speed and connection conditions2. I have found using the Visual Ware VoIP Quality and Speed Test provides a relatively good measure (requires Java on your PC). However, other measures may be required on other continents as distance from the measuring server can play a role in the resulting tests. My cable connection appears to be running between 6 and 6.5 Mbps for download and consistently around 800 kbps for upload. QoS is usually above 90% with low jitter (<10 ms) — more than sufficient for High Quality Video requirements.
- The best SightSpeed can support for “Send” under my network conditions is 320 x 240 @ 30 fps; if my upload speed could be raised to 1.5 Mbps, it could then support 640 x 480 at this frame rate.
But the real challenge is the overall end user configuration including network conditions. And this is where Skype still needs to some testing and more experience to achieve a consistent and sustainable3 High Quality Video user experience. I have attempted Skype High Quality Video calls in several situations with significantly inconsistent results and continue to expand on that experience base.
A second question: In launching High Quality Video as a marketing tool “why did Skype let the situation get out of their control in terms of setting user expectations?” (Note: broken link to original post; no other reference) Can’t hit the target when there is no stated target to aim for. It’s insufficient to state what the requirements are; independently measurable benchmarks are also required:
- Resolution: 640 x 480 (VGA)
- Transmission rate: 24 -30 fps
- Time of call: 30 minutes meeting these requirements
- Time to reach transmission rate at stated resolution: < 2 minutes
And how does it deal with other considerations?
- Low light environment
- CPU utilization management
- “Intelligently” adaptation to network traffic conditions
Both subjects of upcoming posts.
1 Tools | Options | Advanced | Connection –> “Display technical call info ….”; then run your cursor over the active call window to bring up the technical info for your call.
2 In one instance the other party found his speed to be about 400 kbps download; he thought he had subscribed for a 3 Mbps service 3 years ago so checked with his service provider who confirmed he was only getting a 500 kbps service. At this point he is seeking compensation from his service provider!
3 Sustainable means to have a 30 minute or longer video call while maintaining the High Quality Video status as indicated by the presence of the logo at both ends.