The Streaming Toolbox: Vegas Stream, axle ai, and Lumen Orchestrator
All three of the products in this installment of Streaming Toolbox are perfectly suited for today's work-from-home streamer. The first produces a professional-looking live stream for your virtual events. The second uses cloud-based content management for search and collaboration. The last one coordinates multi-CDN edge delivery to ensure each viewer is getting the best viewing experience.
Vegas Stream is a brand-spanking-new product from 20-year-old Vegas Creative Software. PC desktop users (sorry, it's not compatible with Macs) can enhance live streaming with many of the features found in higher-end professional production tools. Vegas Stream can push 1080p content out to any Real-Time Messaging Protocol (RTMP) environment, including YouTube, Vimeo, Kultura, Brightcove, or Qumu.
There is a growing demand for quality live-stream content across multiple markets, but live streams by themselves are too often just not that interesting, says Sumit Rai, chief product officer for Magix Software, owner of Vegas Creative Software. Vegas Stream was developed because its customers were producing more content for broadcast via the internet and were looking for additional functionality.
Stepping Up Live Streaming
Vegas Stream supports live-events production for up to 10 video sources and 16 audio tracks. "You can bring in graphics and switch cameras. We support Blackmagic Decklink inputs, [so] you could theoretically connect eight high-end, full-frame HD video feeds," says Rai. The product demo showed some great features for producing very polished-looking live streams, incorporating content from NDI cameras as well as webcam or screen output from a Microsoft Teams or Skype call.
The differentiator Vegas Stream pitches is a built-in postproduction workflow and a very powerful graphics engine. It ships with five customizable templates with lower thirds, branding, overlays, integration with social media feeds, a Microsoft directory, graphics, live spreadsheet data, picture-in-picture, two and four box transitions, and animation support.
"Not only could you broadcast to the internet, but the product has the ability to send out a live NDI feed, so it could be picked up by monitors around a conference," says Rai.
Precise Audio Control
"We can bring in up to 16 audio sources as well. If I were doing a podcast with multiple guests, I could bring all of their microphones in. Another cool thing is the ability to follow your video switching," Rai says. "Let's say that I had a professional microphone and I wanted to make sure that when I switched to a different camera angle, it always stuck with that really high-quality mic. I can make sure that it follows along with that source. Or, you could have a different mic assigned to each camera."
Content is encoded to H.264 or MPEG-2 and is stored locally. It can be recorded to separate tracks for use in a nonlinear editor. "We see ourselves as a really good stepping-stone to the highest end, but we're definitely not at the lowest end," Rai notes.
This release supports streaming to a single location, but look to later versions to have better configurations. Vegas Stream does not do subtitling or closed captions and has no cloud support, although the company is working on it. According to Rai, "We are moving into the cloud soon, so content recorded via a phone can immediately edit from it and be able to archive your projects," he says.
Pricing: The software bundle for Vegas Stream Pro 365 is $35.99 per month, or $31.99 with an annual commitment, and this includes Vegas Stream, Vegas Pro, and Sound Forge Pro. vegascreativesoftware.com/us/vegas-stream
axle ai, a cloud-based collaboration and asset management tool rolled into one, is used by at least 650 companies worldwide on desktop, tablet, or mobile devices.
When the founders started the company 9 years ago, they wanted to create a lightweight tool for non-broadcast users to catalog and repurpose their video. "You need a strong search capability to avoid just fumbling around all day [trying to find something]," says Sam Bogoch, CEO. "We thought it would be a few thousand clips. … [N]ow we routinely have customers with hundreds and hundreds of terabytes and even petabytes of material."
This year, the product has helped many companies get back to work. "One of our customers, in 24 hours, switched 40 people to working at home," Bogoch says. "Because this is in a browser and it's making H.264 proxies of all your high-res and 4K media, you essentially have the potential to work from wherever you are.
"Over time it's evolved to include a lot of AI functionality," Bogoch continues. This includes multiple language audio transcription, facial recognition, and the ability to train for unknown faces. "I can search across all the videos for a particular phrase, like ‘goal,'" says Bogoch. axle ai has created custom training dictionaries for larger customers, but they are not yet in the base transcription tool.
"You can quickly see what's going on [via scrubbing the thumbnail], but then you can drill down and click on the individual asset and see the contents," says Bogoch. Users can annotate or even grab a bunch of clips and push them to their editor of choice. Technical metadata gets extracted from the video clips. You can also create unlimited numbers of custom metadata fields. "We have a workflow engine called Connectr that lets you build custom workflows so that you can push the metadata from within axle onto the next steps of the workflow," says Bogoch.
Note that I did not demo the workflow engine. axle ai is compatible with live workflows and can even trigger ingests of live material into the system.
Bogoch says that most media asset management systems "want to take stuff in and stash it somewhere and keep it in some canonical format." axle ai doesn't alter content but creates proxies of content in your folder structure, keeping the original files intact. Its proxy is linked to the high-res source file, and it has started making a third-tier mezzanine file, which is a 720p proxy with all the audio tracks intact. There is a separate module for handling high-end transcoding.
axle ai has plugins with modified functionality for Adobe's Premiere, After Effects, Photoshop, Final Cut, and Avid solutions. "Whichever editor they're using, they can be remotely accessing their media instead of handing each other hard drives or FedExing things. It makes a huge difference on the collaboration," Bogoch says. Customers generally use axle ai in a mix of cloud (Microsoft, Amazon, Google, Wasabi, and Backblaze) and on-prem. The product was very easy to use, there's great product content on their site, and the plugin functionality is a real plus. Bogoch says setup usually takes an afternoon.
Pricing: Fees start at $395 per month for a five-user license, and there is module-based pricing for add-ons such as integration with various archive systems. The transcription add-on is priced from $1 per hour of transcribed material in English, Spanish, or French. Servers can be MacOS or Linux, and client workstations can be Mac or Windows. Adobe Creative Cloud plugins are available at no cost for both platforms. axle.ai
Getting data from all the different sources in the delivery chain that can impact the end user is a real challenge and quite expensive. Streamroot, recently acquired by Lumen, saw that back-end CDN load balancers couldn't switch midstream and developed Orchestrator, a device-side routing application. "We thought this device-side intelligence could be really interesting and very complimentary to some of the solutions that are out there already that don't have that perspective," says Erica Beavers, senior product manager for go-to-market and marketing strategy at Lumen.
"Most people use standard monitoring from companies like Catchpoint or Dynatrace," says lead product marketing manager Craig Lowell. The data collection location is determined by each monitoring vendor. "If the local ISP goes down, that may actually get missed by a standard load balancer unless they have monitoring set up at the last mile as well," says Lowell.
Orchestrator calculates its own QoS score based on error rate, time to first byte, and bandwidth to work with client-designated business rules. For those who prioritize on quality, all CDNs would be rated relatively equally. "That would allow us to switch more liberally if we see even a slight decrease in bandwidth. The midstream switches are completely seamless to the end user, so they wouldn't need to refresh the page," says Beavers.
Orchestrator is pre-integrated with most open source and proprietary video players. "We act as a proxy—the player requests the segment from us, and we direct it to the CDN that is performing the best at that time," Beavers explains. Orchestrator evaluates each segment request to decide if a switch should be made to another CDN.
"One of the issues that we've seen with other CDN load balancers is they can create a single point of failure," says Beavers. "If Orchestrator is not working, we're not adding a single point of failure. It's going to pass through us completely and just go to the CDN that is the first one listed in the manifest."
Orchestrator's dashboard can filter by different parameters to see how much traffic was delivered via each CDN, via live or video on demand, by platform type, by average burst download speeds, and by concurrent viewers. When Orchestrator detects a point of presence (POP) going down, it automatically switches to a lower-quality video chunk. There's a very brief drop in video quality, says Lowell. Although he pointed it out to me in the demo, I didn't notice it.
No public pricing; streamroot.io/cdn-orchestrator
[This article appears in the June 2021 issue of Streaming Media Magazine]
Vistex's GTMS manages content licensing, Agora provides transport protocols and SDKs for interactive media app developers, and Transmit offers a new spin on advertising
The latest installment of the Streaming Toolbox looks at a remote production tool for creating interactive content, a scheduling tool for delivering live and live linear programming, and a collaborative quality-control (QC) tool for source file analysis.
Automation, machine learning, and artificial intelligence are vital components of any successful OTT service. Here's how to leverage them to improve processes, generate more revenue, and bring a better experience to consumers.
This issue, I'm focusing on products from the world of advertising. Push Live is used to deliver live broadcasts to multiple destinations while providing customized branding for monetization on different platforms or in different regions. MediaMelon's SmartSight for Ads delivers detailed analytics and QoE for ad-supported content. While both of these products respond to existing market needs, the last product—Ryff—sets out to create a whole new market altogether.
In this installment, we look at products for interactive live streams, ad inventory pricing, and CDN balancing
This installment of The Streaming Toolbox is all about remote control, for mobile devices, advertising, and network operations
We look at a B2B intelligence layer for content transport, a company that enables customer relationship management decisioning, and a service that facilitates offline viewing.
Companies and Suppliers Mentioned