Opinion: the competing demands of AVoIP10 February 2017
Myles Carter of Matrox reveals how customers can strike the right balance between performance and reach in AV over IP.
While there is a lot of market attention focused on the topic of AV over IP, there are still a large number of AV customers that find the information overwhelming, questionably accurate, threatening, contradictory, or simply lacking relevance for their unique needs.
In AV, some products offer ‘extension and switching’, while other products are focused on campus-wide and multi-site deployments and must – by definition – live on low bandwidth. How do you compare two products when they can use the same, similar, or entirely different core technologies?
Let’s look at some some defining characteristics of AV over IP.
Reach vs performance
When assessing a particular set of AV needs and selecting a solution, it’s important to think about two elements:
- Reach: How widely the technology touches intended audiences, reach incorporates the number of devices that can be simultaneously extended, types of compatible devices, distances that can be covered, ability to provide AV capabilities on wireless devices, and interoperability and open standards that reduce vendor lock-in.
- Performance: The fidelity at which the original source AV material is reflected to the viewer, performance includes resolution, latency, image quality, real-time processing (including anything like scaling), compositing, cropping, special effects (like transparency), and more.
Looking at reach, no AV technology other than IP has the infrastructure in place to handle objectives such as streaming wirelessly on premises, nor can anything else be deployed to provide AV in campus and multi-site projects where the AV can reach significant distances. Similarly, with regard to performance, we can divide the market into solutions that can and cannot deliver 4K and full HD content. However, unlike reach, where only IP-based technologies are in contention, the same is not true for performance: many competing technologies can deliver 4K and full HD video with synchronised audio, including those designed to work on IP.
This is where compression comes into play.
As certain uncompressed and lightly compressed technologies understandably cannot break the reach barrier, there are some combinations of resolution, quality, low latency, and processing performance that simply favour uncompressed core technologies. Similarly there are a few use cases where light compression technologies offer greater ‘reach’ than pure uncompressed video.
High-efficiency codecs, like H.264 and H.265, have been implemented in certain AV products that really strike a balance between the elements of reach and performance.
H.264 is excellent for mitigating heavy transmission bitrates for audio-video sources while still preserving overall quality. The simple fact that H.264 is mature and is widely available on all popular browsers, computers and handheld devices makes it the de-facto ‘maximum reach’ solution in the market. More content has been developed and archived using this codec than anything before it in computing history. Free, open software is available for encoding and decoding. In short, there has never been a more ubiquitous format of massive multi-user standardised audio/video.
H.264 can perform similarly well on the performance dimension when implemented with that intention. “Mass market” H.264 is typically designed for video content and the majority of H.264 streaming video is in 4:2:0—perfectly adequate sampling for motion video. However, advanced H.264 products have been developed that can deliver all of the performance characteristics defined above. This includes the transmission of high-fidelity content like YUV in 4:4:4 (no colour space compression for maximum pixel fidelity).
Almost every single frontier of performance has been reached in H.264 today, all achievable at 10% of the bitrate of light compression technologies. Perhaps latency is a feature where advanced compression faces a practical limit, with the best achievable latency being one frame (17ms at 60Hz, 34ms at 30Hz). Matrox, for example, can offer glass-to-glass latency of approximately 50ms with its performance-oriented H.264 products.
Comparing them with traditional AV offerings, today’s H.264-capable products can also handle real-time best-of-breed scaling and offer cropping at the source and destination. Even advanced processing including multi-source compositing can now be handled in real time, directly within the encoder product connected to the sources or at the advanced H.264 decoder destination.
For the next several years, H.264 will be the gold standard, yet there is no doubt that continuous advancement is ushering in next-generation codecs such as H.265, which will continue to grow in terms of both reach and performance. In terms of striking the best balance between getting signals to eyes and representing top quality, H.264 and its high-efficiency compression technologies in general are not only strong contenders, but are the de-facto standard for AV-over-IP transmission.
Myles Carter works in media relations at Matrox.