Speaker 1 00:00:08 This is the workflow show media production, technology stories, discussions about development, deployment, and maintenance of secure media solutions and some workflow therapy. When you know where you want to go, but are stuck spinning your wheels. I'm Jason Whetstone, senior workflow, engineer and developer for Chessa and I'm Ben Kilburg senior solutions architect at Chesapeake systems today on the show, we'll be chatting with David <inaudible> product manager for move. It GmbH move it as a Germany based systems integrator that Chesa became aware of through the Adobe video service partner program move. It has created some unique product offerings in the industry that focus on Adobe specific workflow orchestration. Today, we'll be talking about helmet, a suite of workflow orchestration tools for the Adobe ecosystem that has the ability to perform automated tasks at the workstation level. Many of them ma'am or workflow orchestration solutions that we've talked about in past episodes are fantastic at orchestration of automated tasks, but also might require a certain centralized processes or workflow changes.
Speaker 1 00:01:13 If you hope to be able to use some of the coolest features of those platforms since move, it focuses on automation of processes within the Adobe ecosystem and can interact with Adobe solutions on the edit workstation itself. Certain automations are possible using helmet that are just not possible with other centralized workflow orchestration platforms, a quick reminder to subscribe to the workflow show. So you're up to date on the latest workflow therapies and send all of your suggestions and feedback on Twitter and LinkedIn to add the workflow show or email workflow [email protected]
And now on to our discussion with David Merzenich, joining us today on the workflow show is David <inaudible> from move it. He is the product manager there. David, thank you for joining us today. Thanks for inviting me. So we brought you on the show today to talk about workflow automation and orchestration in a sort of premiere centric way or an Adobe centric way.
Speaker 1 00:02:09 Mainly I, you know, I, I think I'd like to just talk about how in the industry today and as integrators we at Chessa are very much accustomed to deploying media asset management platforms and solutions that are focused around sort of the enterprise service bus. It's a big sort of monolithic ma'am or workflow orchestration system. And it's great, you know, people get the system into the organization and I think sometimes they very quickly realized that they need to change their workflow in order to fit this new system. And we see that a lot and we see some pain and frustration and just general asked among the creative you end users of some of these solutions that I need to use these solutions to get their work done. And that's why we brought you on the show today because it strikes me that you, as a systems integrator are much more focused on the sort of end use case of in the production. How do I manage my assets within the context of a project or a production? Am I on the right track there?
Speaker 2 00:03:07 Absolutely. So we, as a system integrator, we started to back in 2012 when move-in was founded to glue things together in a specific way. So we declared that we want to bring it and video together in a modern and better way. What we saw at that time was the fact that whenever someone did a system integration, there was an API on that side. And now there was an API on that side and it was just a one way ticket, right? There was an integration been done by someone and you could just do a specific thing. And what we wanted to do was to change the way workflows can be designed. And we thought of how can we, as you described, get rid of that monolithic block and give the end user the best of breed approach, right? To say like, I need a play out system.
Speaker 2 00:03:56 Well, that's the one I need because it brings some functionalities. I need a storage, well, take that one because it fulfills my needs. And so on and on, and at the very end, you need to bring them together. And back in the days, it was like an integration between system a and B. Then there was some integration between system BNC and so on and on and on, and what we wanted to do, we wanted to enable the users themselves and the admins, especially, and the technicians to do those integrations on their own. And I guess we'll talk about those things in detail later on, but that's what we did with Homewood four and the workflow orchestration we call streets.
Speaker 1 00:04:32 Great. Gotcha. So how does that differ from say something like, oh, I don't know, Telestream vantage or a spare orchestrator or something like that. I guess what I'd like to do is just describe sort of the difference between a workflow orchestration platform like that and yours. What does that look like to the end user? What are some differences there?
Speaker 2 00:04:50 Okay. There is one main difference. I think it will be worth to take a look at it and maybe we can do this later if you want me to show something, but I can describe it. So what we see when it comes to take or use workflow engines is that it's just another monolithic block. So you put something into a system, it does something with what you put in there, like files or metadata, whatever, and it spits out some stuff you can use again. So it's somewhere in the middle of the production where you can put some like an acid into a vantage, do a transcode and receive a new transcoded file. So that is not really a part of the workflow. This is not really a workflow. In in fact, it's, it's a functionality. So what we wanted to do was we wanted to break down a production into the necessary pieces we find, and we need, everyone needs to plan a production.
Speaker 2 00:05:45 Everyone needs to start the production with creating projects or something like this. You need to shoot videos, you need to ingest them. You need to add it. You need to export board, you need to distribute and maybe you need to archive and delete those things that are not really functions. Those things are not really workflows. Those things are just the necessary steps in the production. And then we thought off, is there any way to connect those steps and combine those steps in an intelligent way so that it ends up in a workflow, right? Because when you asked the first customer and ask, what is your engine is looking like, you'll get a different answer whenever you ask a customer. Although they're trying to ingest files into a storage, for example, or into them. In fact, it's always the same things, but when it comes to talk about details and workflow, you will always find that they have a unique way of doing this.
Speaker 2 00:06:37 What we do now with them with four is whenever someone is doing one of those steps, for example, just creating a project, we add a workflow to that step directly at that moment, which means someone is pushing a button and then you can attach a workflow to it. Then that workflow runs. And the difference to answer your question between what we find in the market, when it comes to talking about workflow engines is that we distribute the workflows to the local machine or to any machine. So we have a client connected on the workstation that is receiving the information from the server that something has to be done. Here's an example you log into Allenwood for, to create a new project. The first login or you log in is an event for us. That means, I know that for example, David is locked in or is logging in and he's connected to a workstation, which is a Mac.
Speaker 2 00:07:29 The workstation has a specific IP address. It's Tuesday or Wednesday, it's nine o'clock in the morning. We have a specific date and so on and on, and we can take those informations. And based on those informations, we can run a workflow that is attached to the button, lock in to do something. And we distribute that workflow to the server. And the server recognizes that David, for example, in workstation XYZ, that has locked into his client. And therefore we distribute that workflow back to the client and the client is executing everything that needs to be done. And a simple example will be to Mount a volume, right? Because my local workstation has a proprietary client attached that can Mount a specific volume of on a specific storage. And we talk to that API through the client. So in fact, the user doesn't need to do anything they use, it just needs to login and the rest has been done through the workflow. And that is the difference. So we distribute workflows to dedicated machines and that enables us to do a lot of things.
Speaker 1 00:08:26 So it sounds like there is an agent running on the client workstations that sort of receives commands from the server. So it's almost like clients for all intents and purposes. Client says, Hey, uh, I need to do this thing. Client reaches out to the server and says, Hey, client has requested this thing we need to do. How do I do that thing? And then it goes and tells the agent on the client workstation, how to do that thing and what to do, does that, do I have it basically, right,
Speaker 2 00:08:49 Exactly. And that leads to the point that the client itself is the workflow engine. Yeah. The server is not doing anything at all. The server is not executing a workflow. The server is just distributing the workflow to a dedicated client and the client is executing the workflows. So we have a distributed way of doing this. In fact, a distributed workflow engine that really serves as an enterprise service bus for productions. And that is quite unique. And we can do nearly everything the workstation can do on its own. Or let's say things you could do manually on the workstation can be automated. Right.
Speaker 1 00:09:26 So give us an example of how that sort of relationship it sounds like the server then really is just a sort of a message broker. Yeah. So then give us an example of something that you could do in that sort of an infrastructure that sort of a solution that you couldn't do when you've got, say the workflow engine running on a monolithic server, you know, in your infrastruct.
Speaker 2 00:09:44 Yeah. So he comes to really easy and nice example. So let's imagine you would like to create a project on the storage and while doing this, and it doesn't need to be a premiere project, but let's say it is a premiere pro project. You would like to quit that project on the storage. And while doing this, you would like to store it in a specific structure. You would like to create the full structure for that. You would like to set ACL's on folders because your storage is providing those functionalities. You would like to Mount the volume first to create the folder structure and place the project inside of the folder structure. You would like to use a template as stored locally. And then you would like to even tell another system in your environment that you have done that like a planning tool, right? This is nearly impossible with a monolithic block or with a workflow engine that sits somewhere in the environment that has no access to your local machine.
Speaker 2 00:10:36 With the solution we provide. It is because your local client, the agent that is running on your system is able to access your local machine. So your local drives and the attached shed storage is another example is for example, if you would like to do an export from premier pro and you would like to store that export on your local machine, on your desktop without wanting to do the export on your local machine. So you want to place the file locally, but the expert should be done on a render nodes on a distributed render node, somewhere in the environment. What we can do is we can send the sequence export to their inner note. The render own has an agent as well attached. And tell that agent, please render with AME on your local machine. So it's rendered on a disk on a dedicated render node.
Speaker 2 00:11:22 And after that, we can tell the system, please copy that file back to a location that is accessible by my workstation, by my local one and kick off another workflow that is executed on my machine. So we can even distribute a workflow to another one, to another one, to another one to nest workflows. And it would end up with my local machine, just copying the export from the central location to my local drive, but the export has been done on a render farm. So that's nearly impossible with any other system in the world and things like this.
Speaker 1 00:11:58 I mean, you know, just a couple of things that you mentioned right there, light bulb went off over my head. You can tell a workstation to Mount a volume, just that, right. There is something that you can't really do with a lot of these platforms that we are accustomed to or familiar with that are running, you know, in the cloud or on some sort of an infrastructure on some sort of a server in your infrastructure. That's extraordinarily valuable. The fact that you can render using a render farm, you know, that you've set up with AME. That's really pretty cool too. So I would say that, you know, that's definitely something that I have personally seen. Ben, I think you too, and the industry in terms of needing a, you know, a gap that needs to be filled, it sounds like this sort of an approach really fills in that those gaps.
Speaker 3 00:12:39 Yeah. I think specifically the ability, and we've been starting to see more and more of this now that everybody is kind of dealing with pandemic centric workflow where the idea of proxy local and higher as in the cloud or private cloud or what we might call the office as a private cloud, right. Being able to keep those weightier codecs somewhere on a system that can handle them, you know, digging into the backend of helmet. I was impressed and it reminded me that the days back with final cut server, but also tell us streams episode where we could actually do split and stitch workflows, as well as the good old days of compressor, where we could actually set up, you know, a bunch of machines to be able to handle a lot of volume. We were sending at them and you guys have really done the work to be able to do that and dig into the backend of what Adobe has to offer and really leverage everything that people are used to on the front end and that we might be doing in our normal day-to-day work on workstations, but then automating that and taking that off of the workstation and being able to do it in the background, which is super useful,
Speaker 2 00:13:57 Perfect description, because that is exactly what we want it to do. The user should focus on what the user should do. Use premiere pro use the tool as it is, click on the buttons that are there, but we do enable the administrators and the technicians to, and the hands or make everything in the background. More advanced here is a nice example. Whenever you import a clip into premier pro premier pro is just setting a heart link on the storage. If you want the files to be copied to a central location, to a shared storage, which is something you would definitely want in a connected environment at a broadcaster. For example, you have to set up premiere pro to do this, which is okay, which is possible. You can even create those import workflow as it can even create proxies and stuff like this, that's fine, but it is missing a specific piece.
Speaker 2 00:14:52 And that is the option to change this per project type or vertical in your environment. So what we do is we just listen to premier pro's own callback functions. And whenever you import a file into your project, premier says, new file, new item imported. And we take that information and we just grabbed the information path metadata and something like this. And then we do the exact same thing as described before we distribute that information to the stream engine and tell the engine what to do with the file that leads to uneasy fact, you can right away start to add it with the local clips, for example, and in the background, a process starts at, for example, copies the file through the central location. It may be checks the fall of their, if they are in house codec, or if they follow a specific naming convention or whatever it needs to have things set up.
Speaker 2 00:15:47 And then we do a transcode and we create a proxy and we add those things to our own database. We like to say in it's our into play from, from your probe. And then we link back through the panel extension to premier in the background and the user doesn't see anything at all. There is just a notification in the dashboard that something's going on and you can track it, but you can right away start to work. And while working the clips will be reeling to the clips on the new location seamlessly. So you won't be interrupted, which is a super nice fact. So that leads to you can even enable any user to import clips into your premium pro project without having the need to go through a centralized ingest. And that's what we usually find in broadcast environments. There is a map system that has to be a centralized engine because the man only supports a XD cam.
Speaker 2 00:16:32 The proxy has to be in a specific format to follow a naming convention. They have to be metadata attached and all the, all those things. Right. And if you start to tell any of those users to do this in a specific way and to follow rules that are written down in a handout, you'll go nuts. Right? So what we do is we just automate that. In fact, it's, it's exactly what a user would do. The user would copy it to a local, uh, to the central location, create the folder structure, rename the file, check. If it has to be transcoded, do a transcode and then release the files manually in premier pro by saying, this is now my file. And this is just automated them. And let me add one thing. This is exactly what we tried to do with everything we find in the environment.
Speaker 2 00:17:14 The premiere pro is just the entry for what we do. So how mode is a project management and control permit pro projects? How would I, oh, is a render farm controller or a job controller cosmos the database and HK is the housekeeper. But the stream engine underneath is able to communicate with any tool you can find around if we create the dedicated connection or connect it to that party system that say, let me give you another example. I play out system. Usually you play out system is connected to the newsroom system or to the map, right? Someone is creating a story and placeholders and all that stuff. And through Musk connection, you will find those places. So those being created in playout systems. So now there's the rundown and everything's in place. Good. And now you need to send the master files to the play out and attach them to the place holders, which is in 99% of the cases, a ma'am driven workflow, where you have to import the master file back to the man.
Speaker 2 00:18:07 The man recognizes the file. It knows the idea. And then it sends it back to the playout system. This is not seamlessly possible because you have to do at least an import into the man or a watch folder workflow, and you have to type ID or rename the file with a specific ID or create an XML besides that. So the map can't find the clip and so on and on, and on what we can do is we can just receive the placeholder ID from the newsroom system, through our API. Whenever you create an item in the story and create a project out of it. And now we know that this project belongs to the place holder, just because the project holds an ID inside of a metadata field, for example. And then if you do an export, we can take that ID and do whatever's necessary to bring that file back to the play out system, by just calling the playout system directly through an API call and saying like attached to placeholder with IDX.
Speaker 2 00:19:05 Why is that? Because we know that ID and this can be done directly from the panel extension. And again, we don't have to tell the user how to do that. We just tell the user, if you want to export your file, click on export master and that's it. And due to the fact that your project, maybe as a news project or a sports project, or it has been created by Peter, or it has a length of more than 10 minutes, whatever information you can find him use, this leads to a different distribution channel if you want. And that's the funny part, right?
Speaker 1 00:19:37 No qualifications, like sort of qualifications about metadata and the project can drive a particular version of a workflow or something like that.
Speaker 2 00:19:46 Right? Absolutely
Speaker 1 00:19:47 Completely seamless to the user, obviously from the user exports for news might be different than exports for documentary pieces or
Speaker 2 00:19:54 Yes, but the button is still the same, right? For every user, it's just one button it's saying export master. And due to the fact that your project is a different one, the stream will behave in a different way. So we use that as conditions is that the newest project go that way, is it a sports project? Or that way? One thing I would like to add at that point is we've seen those things in the it industry back in the nineties, right? They used to create huge enterprise service bus systems to control data. And in fact, that's what we do. We collect data as a payload, we manipulate the payload and then we send the payload to another system. And in case of premiere pro and in case of the user and the assets, we have a lot of information. We know the project type, project, name, size creation, date, user that has created the project. We know where it lives on the, we know assets that have been used or imported into the project. We know the user and the user object. We know the workstation that has done those things and all of those things can be a condition, right. And even attached metadata on top can be a condition to change the way the stream works.
Speaker 3 00:21:03 I'm going to get cheesy here for a second, but I think that there's a marketing phrase that's been bouncing around in my head for the last few minutes, just listening to you too, David. And I really think that this is a man's best friend, right? Just like a dog is a man's best friend or a person's best friend these days. It's funny you say that colder. I was actually thinking that it was a Adobe editor's best friend. Well, yeah, I mean, because as, as editors and as creatives, we just want to do the work we're assigned as quickly and as creatively and with as much fun as possible. We don't want to have the task of logging it in the ma'am or doing any of those granular steps. We don't want anybody to tell us that we need to do more steps of work. We just want to meet our deadline and get things done. Right. But if on the backend we have something like helmet that can, you know, take the information about the project. Briefly spin that up in the ma'am, maybe create a new collection or catalog based upon the project itself, you know, using cosmos, imagine the background to communicate some of that information. Then really it takes all of that work off of the end user. It's like magic
Speaker 2 00:22:15 And vice versa. And that's the key. So our API enabled any man, if the man comes along with a workflow engine to do the exact same things, let's try to understand what that means. As you said, in the very beginning, most of the, there are missing that premiere pro project portion. They have some of those functionalities and some of them are quite good at it. Some of them are not that good at it, but they're always missing one thing. And that is, if someone changes the project in premiere, they can't keep track of that information. And that leads to databases being inconsistent. So let me tell you how we do this. Well, we do is whenever we create projects, safe projects or interact with projects, we index the projects, file the project file, and we store those informations in our database. So we have a snapshot.
Speaker 2 00:23:11 If a new asset has been added, we can trigger workflows. And if an asset has been removed because we have our own index and we can compare the index with the new index file. So let's imagine you bring in clips through a panel extension provided by the manufacturer, which is great. You can browse the ma'am, you can find your clips, you bring them into premiere, then you do your edits. And the mom now knows that those files have been imported into the project, which is great. And now you remove half of the files and close the project. The man doesn't know what we do is we have the index. We know that there are 10 files, for example, and we have removed file five. We closed the project index, the project, and we understand, oh, five files have been removed from the project and we can right away tell the man by executing a stream per file, that those files did exist in the project and now have been removed, which is great.
Speaker 2 00:24:02 So now we can keep both database consistent. And if the man has a workflow engine itself, it can work the way around. So for example, let's say a journalist starts and creates a project in the man or in the Pam system or collection, as you said, and adds asset to the collection and changes metadata and all that stuff. If there is an option to send those informations to Cosmo, we can create a project in humbled for, through the API, a premiere pro project. In part, all of those files into cosmos. As, as soon as you open up the project in premiere through hundred four, it automatically imports the files through the panel extension as before. So we can have a real combined workflow between those tools and it's no longer a one way ticket. And that's really one of the cool things about it when it comes to talk about premium. And at that point, because that enables manufacturers to really, really keep track of the assets, because that is what they should do, right? They always know if an asset has been used or isn't used, or is used in other projects, we can keep track of those informations. We can keep them in forum and therefore all the housekeeping tasks and all the archiving tasks can perfectly
Speaker 1 00:25:11 Right. Let's back up for a second. What I would like to do this is something I like to do sometimes is to, is to help people understand why this is a challenge in the industry, what your product kind of solves and what your solution kind of solves is in order to have this kind of a tight sink and round tripping functionality with premiere, you really need to have something on the workstation itself that communicates with premieres API that also can communicate in some way, shape or form with the ma'am itself, with the workflow engine that the man has partnered up with. So this is, I think one of the reasons why helmet four is so cool is because of that agent, that's running in the background that we mentioned earlier, that's the one telling the helmet for server, Hey, we need to do this thing.
Speaker 1 00:25:51 You know, broker, do the work of telling whatever it is. Even if it's me, me, the agent running on this user's workstation. Tell me what to do. This is what the users requested. And that is something that we don't often have. And a lot of these man platforms as great as they might be at what they do. What I tend to find is that, like Ben said, a lot of these platforms have, uh, Adobe extensions and panels. And a lot of times they look just like the web UI of the platform itself, which is great that that's, you know, people are already familiar with that. They're already into it and using it. But because at the end of the day, what we're looking at as a website, you know, the ma'am is a website. That's what we're looking at. And it's communicating with another server somewhere. It's not really aware of what you are doing on your workstation and without doing like specific development into a premiere extension or a panel yourself, it's not, you know, really there just isn't that piece there. That's doing that. You know, that's, that's putting those requests out there. So that's really one of the big gaps that, that the solution fills in is that sort of agent message broker situation,
Speaker 2 00:26:55 Right? And in case of premiere and after facts, we have one thing on top, which is a hidden panel extension. So you can run panel extensions in the background. So there are like always, let's say, and that panel extension can receive calls. So we have two layers in preview. One is the visible panel to export clips, interact with the Cosmo database and the dashboard that shows the jobs. And one is the behind the scenes panel extension that runs in the background. You'd never see that one as soon as premiere starts, that one is as enabled. And this one enables us to execute, extend script calls in premium. So use as API. Let me give you an example of a real life example of a customer. We have a customer in Belgium, they're producing the biggest soap show in Belgium. The show that runs since ages and what they wanted to do is they wanted to speed up the engine process because they shoot three <inaudible>. So three streams and dedicated clips on a camera card and dedicated audio clips on another card, right? And they said to us, whenever they prepare a project, it takes a day. They have to import the clips. They have to create those sub clips and those multicam clips, they have to link audio and video export those clips as combined clips and all those things. And they asked us if it's possible to do this in a complete automated way, David we've learned,
Speaker 1 00:28:22 And many things are possible given enough time, effort, and money.
Speaker 2 00:28:26 That's true. That's true. It wasn't really expensive. Okay. Good to know, to wrap our heads around it. So what we did was we wrote a small tool that monitors a specific watch folder, and that tool understood, alright, those are multi-cam clips or clips that need to be combined. Then we have kicked off a workflow in Hamad for that imports, all of those clips through the hidden panel extension into a premiere pro, it adds those clips to the timeline sorted by time code and by name. So camera a, B, C, and mix. Then it exports that sequence as an XML and reimport that back to Cosmo into the project those clips belong to. And whenever you open up the premier pro project through Amad four, it automatically imports the multicam clips. So what we did was we automated the way a user would import clips into premiere and create those multicam clips and export those multicam clips just by using the hidden panel extension and use extended script calls.
Speaker 2 00:29:32 So now they're doing this completely automated and they just save one day per episode, which is a lot, if you have a show on a daily basis. And that's just an example, one of the real life examples. And if we now combine all those things and paint the picture, well we have is the portion that enables us to keep track of the project themselves. The portion that enables us to keep track of the assets inside of the projects and the agent, the client on the local machine that enables us to execute everything that user would do manually, right. And that closes all the gaps we've seen. And if there is still a gap left, we just create an action node for the stream designer. Maybe we can show a picture of that one real quick. If you want me to show that we just create another portion that says, okay, if I receive information from any of the systems through ax board panel, extension, or watch folders or import through the web or whatever it is, I can take the payload.
Speaker 2 00:30:28 And now I want to do something with a third party tool. And then we create a portion. We call it note. And that note communicates with a third-party tool through the API and tells us third party tool. Hey, someone has export it, sequence X, Y Zed. And we have created a master file with Amy. And it's stored at that place. Please take it and do something. And if you're finished call back saying done, and then the stream does something else. And in fact, again, we just automate the things people will do manually. That's really the key to the success. And to answer your question, whenever we talk to customers about that, the problem is, it sounds like we're show offs, right? Because we always go for yet, we can do whatever you want. So if you go like this and try to sell a workflow engine and tell someone, you can do whatever you want. They're like, okay. And what, what is that I don't have?
Speaker 1 00:31:22 Yeah, exactly. The dance of what can I do? Well, what do you want to do? Well, I don't know. What can I do?
Speaker 2 00:31:27 So what we did is we edit a nice project management around it and we can tell a story. And the story is, Hey, you can log into that system. You can create your projects. You can find your projects, you can open the projects and work on the projects. You can export sequences, you can import files and you have your steps of the production. And on top you can add the things you would like to happen whenever someone is interacting with the tool to the buttons, and then it goes like, ah, gotcha. Okay. So I got one question and then that kicks off the discussion. And then there is a solution.
Speaker 1 00:32:05 It sounds like as simple and easy to understand interface, which can sort of start the discussion of, I see what all the sort of concepts are here, the buttons and the widgets. And then you can start to think about what functionality you want to have attached to those, those buttons and widgets.
Speaker 2 00:32:21 You're absolutely correct. What we do is we do not provide functions. We provide options for functions. It's a toolkit. In fact, and I really liked the fact that you said that word, because most of the tools you'll find in the market, they provide a set of, and you have to find your workflow within the functionalities, which leads to 70% of the functions are valid for your workflow and you're missing 30%. And then you take a look at another tool and at that provides another 10%. And then you'll find another tool that provides another 10%. Are you close to what you would like to do? What you're not able to combine them. What we do is we just provide 10% of the functions. As I described in the very beginning of that session, you can plan your production, create projects, import clips, added export, distribute, archive, delete. That's what we do, nothing more. And you can define the way the tool should behave. If you're ingest, should do a million things. Well, then do it. If you just want to hotlink your clips, because they're already stored at the right location, then do it. And that's the key to the success. And in fact, it enables us as a system integrator to do proper system integration. It's the fact. So we just created the tool to do a proper system integration, right?
Speaker 1 00:33:41 I think this has been a great discussion talking about some of the gaps that really, I think, need to be filled in that solution. Definitely addresses. Let's talk a little bit about Adobe. We just released a few episodes where we featured Dave homely from Adobe and Michael gambit who leads the Adobe certified partner program as Chesa has sieved that certification sooner, that, that move, it has had that certification for quite some time. In fact, weren't you guys, the first ones to receive it?
Speaker 2 00:34:06 Yes. We've been the first ones. How long has it been? It's more than a year now or maybe two. I'm not quite sure. I think it's more than a year. We know Michael comeback for a long time, gradings Michael, and he was working really hard to bring that program on the road and was super happy to be part of that program from the very first day. In fact, it has to do with our mindset and the things we did in the last 10 years, because the company was founded as I described before to bring video in it together in a better way. And we focused on final copy. Like first two years, we took a look at premiere all the time. And then one day we started to interact with the guys from Adobe and we use premier pro on, I think it was the, the world cup in Brazil, 2014.
Speaker 2 00:34:56 And we introduced to, at that time, I think with premier and a tool called AME renders server, which is now how would I, oh yeah. And then the whole discussion and the whole partnership with Adobe. And I started and yeah, since then we do a lot of integrations around premier because Adobe and premier enables us to do things we could not do with other Annelies at that time. And even can't do nowadays still can't do them. The certified partner program enables us to even shine more because people do recognize this, even our customers. So in the broadcast industry, this is like, and even in Germany, you know, if you have that certification, they're like, gotcha. You're the ones it's not that strict in the U S I guess. So
Speaker 1 00:35:42 If
Speaker 2 00:35:42 You can prove the things you do, and if you can just prove your ride, people trust you in Germany, it's more or less like, can I see your certificate or your graduate or whatever. So that really helped us to do more of those things. And it even helped us to get in touch with you guys and others in the world to do more business across the globe, which is a great thing.
Speaker 1 00:36:03 That's the other thing I wanted to just mention is that's one of the reasons we're talking right now is because being that we also have that certification. I know now for the last several months or month or so, it does give us sort of access, I would say to folks like yourself and the solutions that you guys are working on, which is great. I mean, like I said, this sort of a solution, I think is something that there really aren't too many other options out there for something like this that I'm aware of. Ben, I, I, you know, my wrong there. No, you're not wrong. Okay. So David, you did give us a little bit of history there. Tell us how you sort of got into the industry and how you got your start and what led you to where you're at right now?
Speaker 2 00:36:39 Oh, that's funny because that's a long story. Initially, I'm a musician, right? I started to make music when I was like three years old and I spent 20 years of my life just playing music, creating music. It was all about music. I don't know what
Speaker 1 00:36:57 Talking about Jason and I have not had a similar experience at all.
Speaker 2 00:37:02 And when my school career was, has ended and I needed to decide what to do, I was like, okay, do you want to be a musician? Or should it be something more let's say conservative? And I was like, Hmm, I'm not quite sure if I want to go that step and blah, blah, blah. And then I went for picture and sound engineering and I studied patients' sound engineering with focus on music and all the things like I played guitar for a few years, classical guitar. And I had all those things around music, but I got a technical degree. So I had physics and mathematics and it was all about video stuff and all the cool things. And then I got into that post production thing. And my very first job was head of post production at MMC studios in cologne, which is in fact, the biggest studio facility in Europe running around 52 studios from thousands of square meters down to just a few hundred.
Speaker 2 00:38:00 This is where they produce, like Germany's got talent, the tase and all those big shows. And I was running the post-production at that time. And I was really into workflows. I really had process processes. I just wanted to change things on a daily basis to make the production even better and faster and more reliable, more transparent for everyone and all that stuff. And at that time I had the chance to do some really fancy stuff for got to dance. I don't know if you know that show. We did a live logging on iPads with the camera teams. We had 10 teams out in the studio, they were running. And I used that technology from Sony to live logging on those XD cam cards. And when we ingest that, the clips into our Pam and the storage, everything was tagged, every metadata was set and they could easily find any clip and from shooting on Monday morning.
Speaker 2 00:38:53 So we started to shoot on Saturday and the first episode was completely edited on Friday evening, a two hours show. That was really when it clicked. Like there has to be more options that we need to find better ways and more options. But I was just part of a production team. I wasn't able to define it, create new workflows. And at that time I got in contact with the guys from move it because they were providing hardware and software. And they said like, yeah, we get someone who's doing some code stuff. And I had a lot of ideas and it's some little helper tools. And at one day they asked me to join the team and that kicked off all this stuff. So, and then it's all about the things I've learned in the productions. It's all about my experiences. And now I can do all the things I ever wanted to do as someone that was part of a production. Now I can create those tools and enabled people to do those things that I wanted to do when I was on that side, which is a great thing.
Speaker 1 00:39:48 Fantastic. As Ben said, you're telling a very similar story to actually a lot of us had Jessa, but Ben and myself in particular,
Speaker 2 00:39:57 We should, we should have a band.
Speaker 1 00:39:59 We should have a band. Exactly. Actually I just started playing the harp. And since you play the classical guitar, you'll know why I'm showing you my nails. I
Speaker 2 00:40:07 Don't know if you can see them.
Speaker 1 00:40:08 Yeah. Anyway, that's fantastic. And getting a little philosophical for a second. I think that this area of discipline that we were talking about, I find that the musical brain tends to tend to gravitate towards that sort of a thing. It's about the sort of underlying mathematics of sound and vibration and how that all relates to, you know, I don't know, it's, it's just a theory that I have, but
Speaker 2 00:40:34 I'm with you
Speaker 3 00:40:38 For me, video is just the higher end of the spectrum, right? It's all just vibration the sounds on the lower end, you get goes faster and faster. It becomes light. And then, you know, so it's all the same thing. It's just that I think you're right about music and how our minds are wired to understand relationships a little bit better because we understand the tonal center, you know, we can immediately say like, oh, this is the core, right? This is the most important thing. There's some ability that's in born there. And I think that as we become, you know, technically proficient with software and technology and kind of the networking aspect of that and how the human consciousness kind of grafts onto the network aspect of that, all of our ability as musicians to seal those patterns really kind of digs down and pushes out and makes it really a little bit easier for the R types of brain to understand signal flow and the process and how people work and creativity. So, yeah, good times.
Speaker 1 00:41:42 It's interesting. You mentioned it really speaks to a lesson that I just learned. I was trying to I'm in my band, I'm trying to incorporate this new instrument a little bit. And I realized, you know, the harp that it's as either it's a small autoharp. And the thing that I'm learning is, you know, this is the first string instrument I've ever played besides the piano, which is actually technically a percussion instrument, but, you know, strings that are being vibrated in a certain way. Yeah. It really does kind of look like a little piano, you know, but anyway, I learned that when you play with a felt pick, it really brings out the fundamental of the tone. When you play with a plastic or metal pick, it stresses the overtones, which really influences your perception on how loud the instrument is, which is amazing. You know, and this is the stuff that I love, you know, figuring out and finding out it's pretty cool.
Speaker 2 00:42:32 It's the same with workflows. As you said, at the very beginning, we're talking workflow orchestration. So it's, it's an orchestral, right? We have to orchestrate things together. We have a huge amount of people creating a piece, which is a video file. And they do this together. It's all of their faults. It's all of their creativity. And they use tools to enable themselves to do that. It's the same with music, right? People take instruments, they, they enter a room and together they build that piece of music at that time. Right. So workflow orchestration is close to this, right. And that's why I love that. And whenever customers do have questions and ideas, it's not like, oh, another idea it's more or less. Yes. Another idea, right. I love to figure out how to solve problems for customers, right. That brought us to that point and whisper room.
Speaker 2 00:43:28 We're working on a new version of that engine. And we want to decouple the, the whole stream logic and engine from what is nowadays create helmet five as a package, and enable other tools to directly interact with the stream engine. Right? Because what we now do is we do the integration to other tools, but we want other tools to use that engine as a workflow engine in the same way we use the workflow engine nowadays with Hamot four, which means whenever you do something in a third party tool, click on a button to log in, whatever the tool provides. It will be great. If that tool could behave in the same way as Hamill does nowadays, nowadays just send a call to a data engine and tell the client what to do at the time someone is interacting with an application like within an orchestra. So it's like a conductor helmet for is a conductor. Absolutely.
Speaker 1 00:44:18 That was a great button.
Speaker 3 00:44:20 We can talk that
Speaker 2 00:44:25 One of our developers, he is same story. He's a musician. I was like, you're a musician. Yeah. Yeah. And then we thought about code and structure and it's again, something similar, right? They solve problems and they build a picture. And at the very end, it is a piece they have created and they are super happy with that. And it shines and it works. And then it's like a recording and you can play it and play it and use it and use it. And they like it. It's all about the fact that you create things, right. And even we, as a manufacturer to create a tool like this is a creative process, it's less technical stuff. The technical stuff is just, it enables us to do what we want to do.
Speaker 1 00:45:04 It's very much like saying, you know, you've heard the expression whenever you have a new hammer, a shiny new hammer, everything's a nail, but really the hammer is the tool that you need to be able to distinguish what really is a nail. You know, what really is a nail and, uh, how long has the shiny new hammer gonna last and all this kinds of things. So you've really have to take that tool and apply it to what your, you know, what your, what your problem is or find the right tool for the problem. And that's, I think where the creativity comes in, you know, maybe what rules can we break with this tool versus that tool that might enable us to work better, faster, you know, less painfully that's one thing that we're really passionate about is let's work less painfully. That's what workflow therapy is all about. That's one of the reasons we use that term so much here is, you know, how can we get your work done without the pain?
Speaker 2 00:45:50 Yes.
Speaker 1 00:45:50 That's what Ben and I call ourselves and the rest of the folks at chest. So we call ourselves workflow therapists, consultation, and solving these problems is it's really like therapy, just like music is like therapy.
Speaker 2 00:46:03 Where's that problem coming from instead of long-lasting problem.
Speaker 1 00:46:06 Yeah, exactly. But that's how you tease out though. Having those discussions is really how you tease out. Like, you know, what are the roots of the, of the causes of the pain? Are they because we're working with a certain camera format, do we need to like maybe distance ourselves from that? Is it because we have a particular member of the team that's like stuck in their ways and they don't want to work differently, you know? So do we accommodate them or do we tell them, Hey, you know, this is the deal. That's the kind of things that we find ourselves talking about with our customers and that's therapy. It really is. And the other part of it is, is just letting people know, you know, you guys are not alone. You're not the first people to come to us with this. You know, you might have a slightly different take on it, but, uh, you know, everybody's got some childhood trauma, you know, it's just a question of what are some ways that we can move beyond that.
Speaker 2 00:46:53 Yeah. That's a great picture too, to explain that. Yeah.
Speaker 1 00:46:57 What are some of the most awesome workflows that you have been a part of creating for customers? You've talked about a few of them, uh, you know, I certainly thought they were really cool. Anything that stood out for you that you're able to talk about
Speaker 2 00:47:14 All of them, but let me pick one. So the one I've described with the multicam was really good. One, that was a really nice one. We did for a webinar three weeks ago with lucidly. It's quite an easy thing, but again, it is in combination, just an awesome piece. So what we set up was we set it up a lucid link on a Hetzler in Germany, which is just a cloud provider like AWS. We use men IO next to it as a storage, and we have put a helmet instance in place. And what we do is we access Homewood directly through the web without a VPN. And whenever you lock into Homewood for, it starts your local lucidly incline on your workstation. It logs into that lucid link client and mounds, the lucid link volume that is in fact, in the cloud on hat center and file space and it mounts your POS base.
Speaker 2 00:48:10 And then you can right away start to work. And what we did was we set up a workflow for the webinar where someone was located in London and Berlin and cologne, and we collaborated on a premiere pro project in real time. And that's it. And the funny, the funny thing about that was due to the fact that we could control the workstation with them with four installed the lucid mint client and do the lock-in. We could monitor that lucid link file space, and we could even track every father uploaded or not through the API and stuff like this. And we could show this in the dashboard so we could inform the user if, if everything is accessible and stuff like this, but from a user perspective, it was like this, it was, there was a real cloud workflow out of the box. And this is, this is the special sauce I think right here.
Speaker 2 00:48:56 Yeah. It was not a real, huge, shiny thing. And we did this and that and this, and then with a million combination there, the very end we had, no, it was, it was all about with a little tweak. We went for a real clap workflow out of the box and we were like, oh, that's awesome. This is just amazing. And that's what everyone was talking about since days now, how to do that and upload and download and mirror things. And it's a hybrid thing now. No, no, no. That is a real, and that can just enable you within five minutes to be part of that production by sending you the helmet client, the lucid link Klein, and then you just log in that's it. And that was really amazing that amaze myself, because I wasn't aware of the fact that we could do this easily as that.
Speaker 2 00:49:39 And yeah, really, we created split a split and stitch feature for premier now a month ago. And whenever we create a new workflow, we want to be sure that we do not have to put anything into our backend. We want just to create something within the stream engine, a new action note that does something. We took a look at what we have and what we want to do. And we thought of, okay, if we can kick off an export and distributed to a client and let the client split the sequence into pieces and create new jobs out of it and monitor those jobs, it will be awesome if we can stitch it together at the very end. And what we did was we have a thing called very able store. So we have one action that that creates splits. So you export a sequence from premiere pro run the stream, the stream picks up the sequence and creates multiple jobs in the dashboard by adding in and out points to the same job like in and out, in and out now.
Speaker 2 00:50:39 So we have 10 pieces. So Amy just run this the first part, the second part, the third part, and so on and on. And then we added something to the variables. So saying we did 10 splits and whenever one of the nodes or one of the clients was ready with one of the pieces, the client took a look at the store saying, is there a, still 10 things left or nine or eight? Or am I the last one? And if it says it's just one left, oh, that means I am the last one. Then we kick off the stitch process and stitch things together. So we just created two note to action, knows to do a really, really complex workflow. And now we can do split and stitch up on multiple render nodes. And as much as you like, so you can use a hundred or 10 or five it's limited to 25 frames per seconds at the moment.
Speaker 2 00:51:30 So it can't be used in the U S unless you use 25 frames per second, but this will be there in the future. That's, that's a real heavy task because of the periodic numbers stuff like the math because of the math. But that was a really impressive thing. In fact, what impressed me the most within the last two years is the fact that customers find ways to use our tools. We never thought of. So they call back say like, guys, we did this and that, and this and that. And when I take a look at what you do, wouldn't it be possible to do this and that? And then we're like, yes, it is. For example, we've got a customer in Germany, they're using open media. Whenever they use open media to create a placeholder in open media that creates a project enameled for, and now they want to use that project as a master ingest project.
Speaker 2 00:52:22 And whenever they import fall into that project, they want to distribute that file to all the projects that belong to that one. And so we created a workflow that they can create a new project of the existing ones. So as often as they want, and we keep track of the IDs. So whenever you create a new project, we take the idea of the new project and add it to a metadata field of the existing one, next one, existing one. So we update that list and whenever they import a file into that project, we add the file to all of the projects in our database. And it doesn't matter which one you're working on. They're always in sync. And at the very end, you just delete the masterpiece and everything's fine. And I wasn't aware of the fact that this is possible. They just asked me in a meeting where we had a training, if that is a possible workflow, was like guys, multiple projects. We did not supported that. And then we took a look at it and the very end we found a solution just out of the box again. And the funny part that there is always a new view on what we do from a customer perspective. And that's what I really like about the
Speaker 1 00:53:25 Tool. Fantastic. Yeah. Well, there's some great examples of, uh, some of the things that David Merzenich was able to do with a helmet for, and we were really glad to have you on the show today. Thank you for your time. We will talk to you soon. Thank you very much. It was a great show. Thanks for listening. The workflow show is a production of Chessa and more banana productions. Original music is created and produced by Ben Kilburg. Please subscribe to the workflow show and shout out to [email protected]
or at the workflow show on Twitter. Thanks for listening. I'm Jason Wetstone.