Jason Whetstone (00:00):
Welcome to The Workflow Show. I’m host, Jason Whetstone. The Cloud: it’s here and it’s going to stay. And many of us in the media and entertainment industry are endeavoring to understand the cloud. What is possible today? How should we use it? Maybe versus peers and other industries. And what makes sense for our own individual organizations? We may be receiving pressure from other parts of our organization to transition from traditional costly on-premise storage and workflows to a more cloud-first or cloud-centric infrastructure. Is this possible or will it be someday? Are the bandwidth speeds and pipelines to the Internet really there yet? And what about security? Is our data safe? Some of us are already leveraging cloud-based object storage for archive, delivery to other platforms and maybe even storage for co-located workgroups. But what about during production and post-production? And we’ve imagined a world where we can perform very complex edits, motion graphics, and VFX work, audio sweetening, and coloring without having a SAN or a NAS or an external drive attached to our workstations. So, when is this coming and is it any good? Will we even need workstations? So, let’s talk about these and cloud editing realities and myths here today on episode 43 of The Workflow Show. Joining us remotely over the Internet is creative technologist, Michael Kammes. Michael has been in the media industry for many years and you may have heard of his technology series, 5 THINGS at 5thingseries.com. That’s two S’s. And we’ll spell that for you later. Hello, Michael, and welcome to The Workflow Show.
Michael Kammes (01:37):
Hey, Jason, good to hear from you again.
Jason Whetstone (01:39):
Thank you. Thanks for joining us. Michael is also Director of Business Development for Bebop Technology, and that’s a group of creative individuals specializing in post-production and cloud-based media workflows. And also joining us today, as usual, is my co-host, CHESA’s Senior Solutions Architect, Ben Kilburg. Hey, Ben.
Ben Kilburg (01:57):
Jason Whetstone (01:59):
So, Michael, we’ll get to the topics at hand. But first, let’s talk about you. I want to hear your story. So how did you get into the industry? What got you here? You know, to tell- tell us where you came from.
Michael Kammes (02:10):
Well, in high school, I was the nerd that was playing with the camera. You know, the video camera on the shoulder and editing deck-to-deck. And I went to college to be a director and I realized I had no talent and that I thought I would be a video editor. You know, a film editor at that time. And I realized I didn’t have an original thought in my head. So, I started working with audio and coming from a mom and stepfather who were both, you know, loosely tied to the audio industry and my mom who had what we would call “golden years”. And my father who is a musician. I seem to gravitate towards audio. So, I ended up getting my degree in post-production audio for film and TV from Columbia in Chicago. So not the Ivy League one.
(02:57): And after tooling around the Chicagoland area for a couple of years, I realized I’d much more enjoyed talking to people and working with technology than having people look at the back of my head all day. You know, being in the chair, editing, and mixing. And at that point, what I was doing was more triage than creativity, right, with independent films. So, I worked for what we call “resellers” in the industry, folks who take the technology that manufacturers are making and install it and sell it and train on it.
(03:29): And I worked my way up from being a bench tech at the time to an installer to a trainer and demo artist. And eventually, I made my way to California and I’ve been a creative technologist ever since.
Jason Whetstone (03:43):
Gotcha. Awesome. I mean, there are several aspects of your story, actually, many aspects of your story that sound familiar.
Ben Kilburg (03:51):
Jason Whetstone (03:53):
Right. So, Ben and I also have audio backgrounds. We have very similar entries into the industry. Kind of, you know, I enjoyed the technology and putting the technology together and putting the pieces together a little bit more than, I don’t know, putting the pieces together on screen.
Michael Kammes (04:10):
Do you think it kind of holds true that you have one of two paths? If you go into audio and you’re not a musician, you become a roadie? And if you’re not a roadie, then you become a tech person?
Jason Whetstone (04:20):
Yeah, I guess so. I mean, I still do my own- actually Ben and I both still do our own independent music production. But it’s like not really connected to what we’re doing here at CHESA.
Ben Kilburg (04:32):
Except for the- what we’re doing right now.
Jason Whetstone (04:34):
Except for what we’re doing. Exactly. We both kind of-
Michael Kammes (04:36):
Well, congratulations for the rent of your soul and completely foregoing audio. Thank you.
Ben Kilburg (04:42):
Indeed. Still have a nice little studio. Still get down and funky with my guitar. Good times.
Jason Whetstone (04:48):
So, let’s talk about 5 THINGS. It’s really kind of a pretty amazing collection of, you know, just information, really. So, before we get to the cloud, I want to talk about that. So, when did you start working on that and what sort of inspired you to do it?
Michael Kammes (05:05):
I think it was around 2014, give or take. I’d have to look to be completely honest with you. I have kind of lost track. I worked for a reseller, Key Code Media, who was just fantastic. But when you start getting into marketing, especially here in the US, there- there aren’t as many laws about marketing. And you can now, let’s say, bend the truth, but let’s say “accentuate the positives”. And quite often, the accentuating the positives may cast a light or an opinion on technology that may not be right. Or maybe that feature isn’t shipping yet. Or maybe there’s an asterisk next to it. And as someone who was a former creative, I felt like if I was going to pontificate on technology and I was going to share technology, I needed to be as forthright as possible without, you know, you know, shooting a gun at a manufacturer or in public or even in private. So I went to the owner of Key Code Media and I said, “I would like to take a lot of the information that I’ve gleaned from here and I would like to put it into a semi-fun, at least, I think it’s fun, digestible format that folks who are not technical but creative can sort of grok that technical aspect of it.” And I got his blessing. And as long as I didn’t throw anyone under the bus, it was fine. And so, I started doing the series as a one-man band. And to this day, it, for the, most part, still is. I still have I occasionally call in resources when there’s something I don’t need, but it’s primarily a labor of love. And I try to do an episode every month. There’s been a couple of times where I’ve taken an extended hiatus for various reasons. I’ve just come back with a new episode, which I think you may have seen, and it’s a great way for me to reach out to those creatives and technologists who kind of walk the line and, kind of, the Venn diagram bleed over of both camps.
Ben Kilburg (07:05):
Great. I love that you said “grok” their water, brother.
(07:10): Yeah. Good times. If nobody knows that reference, it’s from, “A Stranger in a Strange Land” by Robert Heinlein, the dean of science fiction. If you haven’t read that book, it’s a really interesting, kind of, view into what the 60s were like from a science fiction lens. And apparently the myth is that Heinlein got together with L. Ron Hubbard at a science fiction convention one year, and they made each other bet about who could come up with a better religion. And Heinlein wrote “A Stranger in a Strange Land” and L. Ron Hubbard came up with “Dianetics”, which is what brought us Scientology.
Jason Whetstone (07:46):
That is an amazing segway. Yeah, absolutely. I did not know that. I know what it means. I didn’t say, you know, to understand deeply and you know, but that’s great.
Ben Kilburg (08:00):
Yeah, sorry, Sci-Fi geek. As you were. Please continue.
Jason Whetstone (08:02):
So, I just wanted to- yeah. 5 THINGS. I just wanted to talk about, you know, just a couple of these episodes. And I want to talk about one of my favorite ones, which, you know, I’m very passionate about. We have an episode about Blackmagic, a GPU and Mac Minis versus Final Cut Pro 10, Adobe Premiere, DaVinci Resolve, Avid Media Composer. It’s just comparing all of these, you know, these technologies, YouTube tips and tricks for your media, building a Hackintosh, building a Roku channel. The truth about video editing software and Hollywood. These are just the titles of some of the episodes. Just some really interesting stuff here. Some very technical, but yet really distilling it down so that the creative professional can get into it, really sink their teeth into it. Not too techie, very accessible to a wide audience, I find.
Michael Kammes (08:50):
I do what I can. I mean, I do want to get technical to some extent, mainly because, as you know, in the technical realm, everything is scrutinized. And whether it be from, you know, just trolls online to, believe it or not, I get a fair amount of manufacturers who send me emails and say, “Well, you know, you didn’t present this right.” And I politely respond to them, “Yes, I did. I just didn’t put your marketing spin on there.” But I like to be, I don’t want to say, “above reproach”, but I don’t want to get called out for being partisan or having an allegiance. So, it’s very important to me to have my facts straight and my tech straight. And that takes a fair amount of research and having almost a peer review sometimes of the content to make sure that it’s on the up and up.
Jason Whetstone (09:39):
Awesome. Yeah. And I think you do a very good job. I certainly appreciate what you’re saying about, you know, doing your due diligence to research and test and, you know, make it so that if someone comes to you and says, “Hey!” You’re like, “No, no, no. This is really how it works.” And people need to know that. So, I mentioned my favorite one so far, which is entitled, “Prepping for Post Audio”. It’s episode 212. And listen, if you remember nothing else today, listeners, all of my video editing colleagues, all of my video editing friends and even my not-video editing friends, please watch this episode because it basically covers everything you would ever want to know to not really, really frustrate your audio post engineer.
Michael Kammes (10:23):
You know, I think I have post audio experience. I’m sorry. Post audio is so compressed, no pun intended to begin with in terms of timelines. Again, no pun intended. If I have to- if I had to spend any more hours cleaning up an OMF and A, cleaning up, B, OMF, the fact that I have to do both of those things in 2019, just- I don’t get it. And I’m hopeful that some of the tips and tricks in there will get folks to not only limit the amount of cleanup work but also stop exporting damn OMFs.
Jason Whetstone (10:55):
Right. Right. And I mean, we’re talking about things like sort of thinking of your timeline, like it needs to be done in a DAW and a digital audio workstation. You covered things like if you have nested sequences, you’ve got to unnest those because they’re just gonna show up as mixes to the AE. You know, this is all great stuff that, I got to admit, I watched it and I was like, “I didn’t even think of that.” You know, this is something that I did for many years. And, yeah, it’s great information. So, thank you.
Michael Kammes (11:24):
No, thank you. I appreciate you sharing that. Thank you so much.
Jason Whetstone (11:28):
So, let’s segway a little bit into the cloud part of our discussion. So, the latest episode of- oh, by the way, 5thingsseries.com. So, there’s two S’s in there: 5thingsseries.com. Please check it out, everyone. The latest episode is called, “Intro to Using the Cloud for Post-Production”. And that’s kind of why we’re here today to talk to you. So, let’s talk through some of the things that I opened with. Like, what are people doing? What’s possible today? I mean, I think it’s worth mentioning you should watch this episode. It’s about 20 minutes long. But let’s just discuss it a little bit. So, investment strategies CapEx, OpEx. Let’s start with that. I mean, that’s typically why we have somebody in our organization coming to us saying like, “Hey, you’re not getting on a new SAN this year. You’re not getting a bigger SAN this year. We’ve got to start moving everything to the cloud.”
Michael Kammes (12:16):
I think we almost need to take a little bit of a step back and talk about what’s possible because I think there’s a lot of confusion there. And the word “cloud” does not help. Cloud is, again, no pun intended, very nebulous.
Jason Whetstone (12:33):
It’s kind of a buzz word.
Michael Kammes (12:34):
It is. And I remember six months ago sitting down with our V.P. in marketing and saying, “What other phrases can we use?” And we had a whiteboard just full of phrases and all of them just were like, well, “That makes sense to me but that doesn’t make sense to Jill Editor or John VFX Guy.” It made no sense. “Decentralized post-production.” It just- it doesn’t fly. So, we’re kind of stuck with “cloud”. So, what’s possible in the cloud? Can you edit video in the cloud? Yes. Are you going to use it for finishing? Probably not.
Jason Whetstone (13:07):
“Terms and conditions apply” is what we usually see.
Michael Kammes (13:08):
Yes, exactly. We find it’s best for creative editorial, string outs. We find a lot of three-letter stations, and I’ll leave it at that, who are launching streaming platforms. Well, how do we repurpose our back catalog? Right. We need to cut down everything. We never provision to have an additional 10 editors in this building. We want to hire editors where the talent may be elsewhere in the country. So, you can do editing in the cloud and there’s a market for it. Finishing color, not so much. You can use the cloud to transmit video from one place to another but you’re never going to get that, “Hey, I’m using a Sony BVM HDR monitor, $50,000 dollars”, and get that color fidelity playing from the cloud. That’s not feasible with our meager Internet connection here in the US. But video can certainly be done; audio not so much. As I point out in the video, we talk about latency and whatnot and it just doesn’t.
Jason Whetstone (14:11):
Yeah, that’s, I think, a huge misconception. It’s one that I like to clarify a little bit sometimes with the way like a DAW works versus the way an NLE works with the rendering and the way the frames are coming through. It’s not really the same with audio. Right. Everybody always says all the files are smaller. So, it should ‘t be a problem. Right.
Michael Kammes (14:30):
When we deal with video, we’re dealing traditionally with, and I’m going to round here so no one hammer me on frame rates, 24 and 30 and 60. And we’re dealing with those kinds of frame rates, 30 frames.
Jason Whetstone (14:43):
60 frames per second.
Michael Kammes (14:44):
Yeah. When we’re dealing with audio, we’re talking about tens of thousands of samples. And as an audio person, you need to get in at that sample level and notch out a pop or a click or remedy a transition or fade into this second take of dialog. You need to have that kind of finite, not only control but to hear it as well. And unfortunately, getting that the latency and sync at a sample level is not something that can be done currently. And that’s why the reasons why, you know, the big 800-pound audio gorilla, you know, Pro Tools, that doesn’t run in the cloud currently. And there’s a reason for that.
Jason Whetstone (15:24):
I always- I kind of like to explain it as- imagine if audio is 30 frames or 60 frames per second, you would just hear- it would sound like bugs or something.
Ben Kilburg (15:35):
Right. No good.
Jason Whetstone (15:36):
More like our ears hear closer to real-time than our eyes. See? I’m not sure how to how to explain it, but the whole difference between hearing and seeing, it’s very- it’s a very big difference. So, the way that the information is transmitted in a digital way is also very different.
Ben Kilburg (15:53):
Right. Right. The simple fact that we can trick our brains into thinking something’s moving by only showing at 24 frames a second tells you something, right? It’s like watching car tires spin backwards, that aliasing effect. They’re not spinning backwards; it’s just your brain being funny.
Jason Whetstone (16:14):
Or it’s your rearview camera or something like that.
(16:18): What are some good things we should be doing with the cloud? I mean, we talked a little bit in the intro. I mentioned archiving. I think that’s a pretty popular workflow for a cloud-based solution would be pushing my stuff that I don’t care about, for now, to the cloud archive. Right.
Michael Kammes (16:33):
Definitely. I mean, the one thing you get with cloud is that storage is virtually unlimited. So, it’s a perfect use case. I think we’re combating a few things. I think one of them is that there is an extreme aversion in the industry to the concept of subscription.
Jason Whetstone (16:52):
Michael Kammes (16:52):
And obviously, Adobe. Thank you, Adobe. Seriously, for being the pioneer to take all those arrows in the back, but they were one of the first ones to make that move. And I think, as you probably know, a lot of creatives, especially freelancers, hate it because they’re paying, whether- and it’s feast or famine in terms of getting a gig. But a lot of businesses like this because they’re guaranteed to get the updates. It’s a predictable expense. And when you get to cloud storage, you’re dealing with subscription. You’re paying a low monthly fee but it’s always there. And it’s a lot like a drug dealer, right? The first taste is free. You can upload; usually, that doesn’t cost much, but when you want to download, then you get egress fees. Right. Right. That’s when they ding you.
Jason Whetstone (17:38):
Another thing that’s worth mentioning is like if you have an archive that’s maybe, maybe it’s on-premise object storage or maybe it’s LTO tape based. I don’t know. I guess I feel like some people are under the impression or maybe, maybe, more accurately, that you have a little bit more control over that. I mean, it’s something that you physically can interact with. The cloud is this thing that, you know, you’re paying a subscription for. What if something happens to that subscription? What if we can’t afford it anymore? What happens to our data?
Michael Kammes (18:02):
No, you bring up a great point. And I think what’s important to remember is that cloud availability is five nines or better. And, as for some of your listeners who don’t know, five nines is kind of a baseline for uptime and that’s 99.999% uptime. And, if my memory serves, I think that’s less than one hour of downtime a year, I think if that’s the right number? And I don’t know how many folks on-prem can say that they have a SAN or a NAS. Oh, by the way, if I say “on-prem” to everyone, that means “on-premises”.
Jason Whetstone (18:44):
It’s in your building, in your data center, probably. It might even be in your edit suite or in your audio suite.
Ben Kilburg (18:49):
But we hope not.
Jason Whetstone (18:50):
Yes, exactly. We hope not.
Michael Kammes (18:52):
But to have an on-prem solution be completely available for 24 hours a day, seven days a week, 365 days a year. That’s a tall order and the cloud definitely gives you that.
Jason Whetstone (19:06):
Right. And I mean, some of these platforms, as you mentioned in the video, some of these platforms give you more nines, like 11 nines or something like that.
Michael Kammes (19:14):
Yeah, Backblaze, I think, is doing that.
Jason Whetstone (19:16):
Yeah. So, when you- when you say, “there’s more nines”, essentially, you’re taking that downtime down, even then some. So I think you had mentioned in the video, it was like five minutes a year of potential downtime, which is who- who can say that they’re spinning disk or whatever, their servers are going to be up, you know, and maybe down for up to five minutes a year? I don’t know. I mean, it seems, when you really crunch the numbers, it does seem like it’s quite a bit more stable.
Michael Kammes (19:47):
It is more stable. We’re also not just talking about availability; we’re also talking about resiliency. So, you may have a drive on-premises that always spins u, but the one time it doesn’t spin up, well, that’s a failure and there’s no recovery from that. In the cloud, you’re always going to have that recovery because there’s multiple levels of backup and archive to get that data back. So, it’s not just a matter of availability, it’s also resiliency.
Jason Whetstone (20:15):
Right. So, you’re paying for that subscription and it’s covering your redundancy as well. So, you’re not having to like, say, make a second copy of all your data manually or with a workflow or something like that.
Michael Kammes (20:25):
Correct. Yeah. And I’m not suggesting that you have your local storage that you’re using for- for local editing and just the cloud and that’s it. I think there’s still some good practices in play where you may have a second copy on-prem because if you have archived, let’s say, 5TB on the cloud or let’s say a backup 5TB in the cloud and you need to get that back. Well, A, it’s going to take a little while. B, you’re probably gonna have to pay to download that much content. So, it’s still good practice to have an additional backup or even archive locally. So, the cloud is not a replacement for a good backup strategy.
Jason Whetstone (21:03):
Yeah, absolutely. So, I see a lot of rather than saying we want to transition our archive into the cloud. It’s more of a we want to add another redundant location to the cloud. Maybe we’re only writing one LTO tape and keeping that on-site and keeping our backup of our archive up in the cloud.
Ben Kilburg (21:17):
Yeah, that’s where I’ve seen a lot of the adoption is in that hybrid approach because you can have your “oh crap” copy that’s living up in the cloud and then you get the benefits of the geographic distribution and your offsite copy. But like you were saying, Michael, it comes back to your recovery time objective and whether or not you want business continuity on-premises or how quickly can you get it spun back up to do your work. Right. That’s what’s important to people.
Michael Kammes (21:42):
You brought up a really good point and, I’m glad you chose this word, is that I find that a lot of folks are looking at on-prem verse the cloud or Mac verse PC and it’s not that; it’s- it’s you exploit the positives of each platform. So, if you can get the instant availability of stuff on-prem, fantastic. But you need the horsepower of the cloud. Great. Exploit both. You don’t have to have one or the other. So, the hybrid approach, as you mentioned, adding the cloud, I think that’s going to be the way to go. And no one should ever look at the cloud as a, “Well, I’m just seeing everything that I have on-prem and going purely in the cloud.” That’s not going to be a way to go either.
Jason Whetstone (22:27):
Great. Right. So, let’s talk security for a second here now. There’re two different types of security I want to talk about. One of the things that I’ve heard some people say is, “Yeah, we’re never gonna get the buy-in from IT on pushing things to the cloud because their security is too stringent and, you know, and they just won’t make it available to the world.” So that’s one thing. The other thing is just the physical security of the data. You know, in your building, you’ve got a data center. It may be protected by a locked door. It may not be protected at all. It may have card reader or some sort of a biometrics device. But at the end of the day, if you want your data to be safe, it’s got to be physically secure. The storage media, the servers, they have to be physically secure. So that’s another type of security. So, let’s talk about that a little bit like what do you get from people who are concerned about this concept that like, you know, “our IT isn’t going to buy into this”?
Michael Kammes (23:14):
Not that I want to disagree but what I find is that when we work with corporate IT, we look at larger IT companies, the cloud. Yes, no problem. We’re in it where that’s- that’s already been- that was vetted years ago. We’re good to go. It’s where- it’s where I find media-centric IT, where media takes precedence over Excel documents. And I find that because a majority of media facilities are doing post-production or even production, they don’t own their content. Right. They own someone else’s content. So, because of that, there is even more of a scare. Plus, the fact we’ve had these high profile and, I don’t want to call them hacks because hacks implies a lot more forethought, but there’s been a lot more leaks due to bad passwords or leaving your account unlocked on an open computer or-
Jason Whetstone (24:08):
Or maybe social engineering or something like that, yeah.
Michael Kammes (24:11):
Exactly. Exactly. If you were to take a look, you know, some of, you know, some of the movies we’ve seen where people are breaking into data centers and, you know, diving under laser beams and that kind of thing, that’s not far from the truth. If you were to look into some of these data centers, we’re talking about several layers of intrusion detection. We’re talking not only here, sign your name, here’s a badge but also retina scans. We’re looking at fingerprint scans, we’re looking at weighted plates on the floor when they get to the data center. Even getting into the data center, which is separated physically with no windows, even getting into the racks where the machines are is blocked. So, there is- there are so many physical layers, you know, that’s not really a concern that most of the clients I deal with have.
Jason Whetstone (24:57):
But that may be a concern in your own facility. Or maybe it should be. It’s just something to think about in your consideration for whether cloud workflows are going to- are going to work for your organization.
Michael Kammes (25:06):
I’m sure all of you have seen SANs or NASs is that or never provision for and now they’re in a closet with a fan or someone’s office. Maybe, maybe you- maybe you stuck the intern in there. I don’t know. But I’m sure you’ve seen this where it was never provisioned to have a secure area, let alone a room, for all that gear. And that means anyone with a thumb drive can plug in and get what they want. And unfortunately, you’re then screwed.
Jason Whetstone (25:33):
Absolutely. So, let’s move a little bit to what you call in the video flopping power, right? I love that it was said so every time. Yeah. It’s it comes up a couple of times and I just- I love it.
Michael Kammes (25:45):
Jason Whetstone (25:46):
So, we’re talking here about compute in the cloud, right. So, talk about compute in the cloud a little bit. Look, I mean, I’m sure a lot of our listeners are very familiar with running compute instances and, you know, and all that. But when we’re talking about a post-production workflow, we’re used to having these massive machines. You know, Apple just released the Mac Pro and it’s like amazing, but very expensive.
Ben Kilburg (26:06):
Well, apples to apples. Apples to HP. It’s pretty much on the nose.
Jason Whetstone (26:12):
That’s right. It is pretty much on the nose. You know, we all know what it takes to have like a machine that’s going to be able to do all this high-speed rendering and everything. So, we’ll talk about some of the solutions and a little bit. But this is another, I think, a selling point for a cloud-based workflow as well, right?
Michael Kammes (26:27):
Definitely. And I think we have to be very careful how the preaching or sharing this and here’s why is because there’s the freelancers who are going to buy a machine and then they’re going to run that thing into the ground. And then there are facilities who are going to buy it for a creative. Then creatives are going to use it. And then in three to five years, that machine isn’t going to be thrown away; it’s going to be moved to an assist station or it’s going to be moved to an office worker or it’s going to be used as part of a render farm and a news machine will be purchased. And I think also those paradigms change how you look at computing and what it costs. When we talk about VFX and rendering, how many facilities need to have these massive render farms? Right. Traditionally, it’s the effects houses, but by and large there aren’t a ton of vastly expensive render farms. It’s usually smaller facilities. You’re looking at using the machines off hours. So, you’re repurposing them. Right. And so, with all that being said, with all those caveats, in the cloud, these data centers are massively configurable. When you buy a Mac or even an HP, you’re saddled with that CPU, those cores that RAM, etc. and you have to outlay more cache to upgrade those. In the cloud, traditionally, they’re upgrading every year, year-and-a-half, which means without paying anything extra, granted, you’re paying every month, you’re getting more horsepower. And at the drop of a hat, virtually, you can say, “I need another GPU, I need you CPU. But you know what? I only need it for these hours.” So, it allows you to cut very close to the bone in terms of how much horsepower you need and for how long.
Jason Whetstone (28:10):
And potentially scale up very quickly in case you get a project that’s like really, really render intensive or something like that. Right.
Michael Kammes (28:16):
And that’s what, I don’t want to say worries me about these big juggernaut VFX facilities, but they obviously are paying for, you know, massive amounts of real estate, mass amounts of age of HVAC and electricity. And there’s now these pop up VFX houses that are saying, “Look, we’ve got artists all around the country or even all around the world. And we just send the frames to the cloud, let it render downloaded. And then when the gig is up, we don’t pay for it anymore.” And so, there are a lot more of these pop up VFX boutiques, virtual boutiques. You’re going to start to see these large footprint VFX facilities having to charge less, which, unfortunately, we’re already dealing with, and we’re going to start to see these pop up boutiques are doing everything in the cloud, sort of takes the other VFX facilities, lunch, so to speak.
Jason Whetstone (29:02):
Gotcha. So that’s something else to sort of keep in mind, is that as this some of this gets easier and cheaper, it gives that, some could say, opportunity but also, you know, it changes the landscape a bit. Right. I’m reminded of like ridesharing versus getting a taxi, you know, in a way.
Ben Kilburg (29:19):
So, when was the last time you took a taxi?
Jason Whetstone (29:20):
Ben Kilburg (29:22):
I think talking structurally through some of the ins and outs of the cloud, or we should say the ups and downs of the cloud. Right. Like, clearly, it’s your WAN connection, your wide area network connection, in and out of wherever you are, whatever facility you’re in. So that has to be really wide to handle the traffic you need to work in the cloud. And that’s for either if you’re using the cloud as an archive or if you’re using it for editorial. Right. You need to know how much data you’re going to be pushing through that link and whether or not you’re going to piss off everybody else in the entire building because you’re stealing everybody’s bandwidth. IT can’t do any of the things that they need to do to back up the entire organization’s footprint because of the media group within the organization might be monopolizing the pipes because their data tends to be that much bigger than everybody else’s. So, Michael, in your specific journey, how do you talk to people about, you know, WAN connections and what are some other unknowns that you see that people aren’t thinking about when they’re kind of thinking and strategizing about how to get into working in the cloud more?
Michael Kammes (30:35):
I think that a lot of creatives and a lot of IT folk have a snapshot in their mind of how working with a remote computer, how poor that experience can be. And, you know, I don’t want to shat on something like TeamViewer, which is, you know, ubiquitous through the industry and everyone uses it but tools like that are much more meant for IT. They’re not meant for media. And I think a lot of creatives have tried to use TeamViewer because who hasn’t been home on a weekend and has gotten a last-minute note and astro mode into their system and change a title or something? And they have to deal with that horrible audio and sync issues and, you know, extreme banding. And I think a lot of folks haven’t seen what can actually be done in a tuned environment. And when you start moving away from these kind of one size fits all remote desktop paradigms and start getting into things like PC over IP, which is made by Teradici, and you can see that you can get 30 frames a second and you can get, let’s see, sub-50 millisecond latency on top of what you already have. And it is possible to work remotely without the common issues you’ve seen. And I hate to say this because it sounds like I’m a pitchman. It totally sounds like it. But I swear to you, every person that I’ve put in front of a zero client, which is a desktop unit that links to a cloud editing system, has sat in front of it and said, I can’t tell the difference. If it’s done right. You cannot tell.
Jason Whetstone (32:15):
So, this is the system where you are not sitting at a Mac Pro or a really tricked out HP workstation with all the GPUs and everything. You’re sitting in front of something very simple and it’s like the ISDN line for the voiceover actor. You know, that you’re basically giving them a KVM switch into the very powerful machine.
Michael Kammes (32:38):
One thing I like to throw out there is and, I’m a throw some math teacher ya, right? So, everyone get your pens. There’s a website out there that, you can even do this yourself and prove it for yourself, it’s called “human reaction test.” I forget the exact URL, but it flashes something on the screen. You hit a spacebar. It tells you how long that latency is, how long it took for you to perceive a change on the screen, act on that change and hit a spacebar. The average is about 210 to 215 milliseconds. That’s a little bit less than a quarter of a second. OK, so if what we’re doing in the cloud, if you hitting space part of play and that signal getting to the cloud, the cloud computer, determining you hit spacebar, starting that play head and then getting that frame of video back to your desktop, if you can do that in under 200 milliseconds and often under 150 milliseconds, which is considered superhuman, the end-user won’t know or they might notice, but it’s like sitting down at someone else’s machine. Of course, something’s going to be a little bit different and they acclimate to it. And I know there’s folks who will say, “Well, I can certainly tell.” OK but does it impact your ability to edit? Is your ability to have more family time? Is your ability to sit at home and your T-shirt and jeans and cut at home, aside from driving an hour to work each way? Does that slight delay, if there is any, is that worth that? And we find that more often than not. Yeah. Yeah. It’s completely worth it.
Jason Whetstone (34:12):
Absolutely. So this is- it’s worth mentioning here that this is a departure from the traditional way that we are used to working with where maybe if someone is asking to sort of temporarily be able to work remotely with editing proxies or something, they might be able to mount a remote storage device like across a VPN or something like that. So, like I connect to my office, I mount the storage and then I can get to editing just like I was there in the building. We’re not talking about that. We’re just talking about transmitting what you’re seeing on the screen, what you’re pressing on the keyboard, what you’re mousing on the mouse. Right. And then that’s the connection. So, I just want to make sure everybody understands that, like this cloud editing workflow is- it’s almost like a misdirection of sorts. Right. It’s just different than the way we’re used to seeing a computer setup, essentially.
Michael Kammes (34:58):
And there are these sort of paradigms. Obviously, the approach I’m talking about is something that I just call, “all in”, where everything is in the data center. And, as you pointed out a few minutes ago, it’s a remote KVM. It’s a long HDMI cable. It’s a long mouse cable, a long keyboard cable, and you’re just controlling that computer in that data center within, you know, a couple hundred miles. There are other solutions. Obviously, Avid has been doing this for years with Media Composer Cloud Remote, and, don’t hold me to that name because it keeps changing but the ability to have an on-premises Avid Interplay, I mean, Avid Media Central UX system, that delivers proxies in real-time from your work facility, delivers these proxies to where you’re cutting remotely. That’s a completely different paradigm. Adobe Anywhere did it. And you know, now companies like IPV and Arvato Bertelsmann, are taking that same approach that Avid uses and Adobe Anywhere used to have and use that for Premiere. So instead of having everything in the cloud, you’re housing everything at your work, at the place where your bosses work, and then streaming those proxies to where you’re editing remotely. And so, you’re using your local machine.
Jason Whetstone (36:12):
Right. Right. So, this is a situation where we’re just using the cloud, basically, we’re just using the Internet part of the cloud; not necessarily the cloud compute and the storage and all that.
Michael Kammes (36:21):
It would be really just transmission medium.
Jason Whetstone (36:24):
Exactly. Exactly. Right.
Michael Kammes (36:26):
And I’m not saying one’s better than the other. I think if you’re looking to control your own costs, then doing it on-prem, I think is a wonderful solution. I think, obviously, we talked about security, and if you do it on-prem, we you have to bolster security. It also means you have to bolster your up and download speeds, your Internet connection. It means you’re going to have to hire an IT person that knows what they’re doing. This isn’t something an assistant editor can learn and be a sysadmin on something. You’re looking at hiring a full-time engineer to handle these sorts of things. So, there’s a lot of considerations for doing it on-prem verse having a managed service handle all that for you in the cloud and then all you do is sit down and work.
Jason Whetstone (37:09):
So, I have a question about that because I totally agree with you and I’m totally with you on hiring someone. What kind of qualifications would you expect that person to have? I mean, a lot of, you know, a lot of larger organizations that are looking for someone to manage networking or something like that would probably be looking for like maybe a Cisco certified engineer or something. Are we talking about somebody like that? Are we talking about somebody with a more broadcast background? I’m asking because we run into this in our industry a lot because as everything becomes more focused on IP and less on broadcast and, IP I mean Internet protocol, there’s like- there’s a gap there a lot of times in the sort of expertise.
Michael Kammes (37:50):
No, you’re completely right. In terms of Avid, Avid obviously has been doing this longer than anyone else. And they- I don’t want to say “stranglehold” because with the price point of what Interplay and now Media Central UX costs, it’s not like there’s 10,000 of these all around the world and you can find a job anywhere. But Avid, obviously, has their own certification program. They’ve had their 101 and 110 and 210 classes where you can get the, you know, purple stamp of approval and the purple tattoo on your back all day long. But when we start getting into things like Adobe and Arvato Bertelsmann and IPV, that takes a different kind of beast. A creative can certainly learn it, but you have to have computer chops. And what I find, typically, is folks who are administering these on-prem solutions also have to know the IT portion. So, it’s not just how does Media Central UX work and how do I assign a workspace and how do I assign permissions, but also how can I configure this VPN, you know, to handle these different clients? And what’s my aggregate throughput and how is it going to be set on this separate VLAN? There needs to be an IT component to that as well. So, I don’t think there’s one course you can go to college and take and come out ready to know how to do this. But it’s got to be someone who understands video workflows and certainly understands IT.
Jason Whetstone (39:12):
Absolutely. We’ve talked about this on the show before, actually somewhat recently, about that sort of gap in the industry of- we’ve kind of come to the conclusion that there needs to be like an educational program that people can learn some of this and because it seems very niche at this point in our history. But I just see it getting more and more needed as time goes on, as these things get more and more, sort of, prevalent in the industry, and they will.
Michael Kammes (39:35):
You had mentioned a few minutes ago. Ben, what do you come across that maybe you didn’t know or what end-users didn’t know? And we find the creative industry that a lot of people like using Macs. I can’t complain. I have- I’m talking to you on a Mac. I have a Hackintosh on my left. I love the macOS. I love working in its ecosystem. And despite the fact that, you know, the pricing of the new Mac Pro is close to an HP, when you start getting into cloud editing, there really isn’t a Mac solution. And I know someone will say, “Well, there’s Mac Stadium and there’s all these other Mac data centers around the country.” Yes. They, however, are not scalable in the way PCs are. Meaning I can’t say, “Give me other GPU.” No, you have to use a whole new machine at that point because you can’t configure Macs like that. Or the Apple OS or, I’m sorry, the macOS Apple EULA says it is for it’s illegal for you to run a macOS on non-Apple hardware. So, this means if you’re working in the cloud, you’re going to use Windows or Linux. That’s how it’s going to be and a lot of creatives balk at that, and I can’t disagree. If that’s what you like to work in, that’s what you’d like to work in. But you gotta admit, once you’re inside Media Composer, once you’re inside Adobe, once you’re inside and any of these other VFX applications, After Effects, Blender, et cetera, it’s the same interface on Mac or PC.
Ben Kilburg (41:05):
Michael Kammes (41:05):
And once you’re in there and, I hate to be the, “Well, if you don’t like it, find another job”, because there aren’t a lot of facilities that do that but there is definitely a, “It’s not just you, Mr. or Mrs. Creative. It’s the rest of the company and our viability as a long-term player, and staying relevant, we have to make these fundamental financial moves. And if it means that you don’t, you’re no longer getting your dock; you’re getting a start menu, that’s how it’s got to be.”
Jason Whetstone (41:36):
Ben Kilburg (41:36):
So that’s interesting. That’s a good segway into some of the finances of being able to do this versus, you know, if I’m going to drop anywhere from $10,000 to $50,000 on a really high-end workstation, what does it cost to run editing, editing and editorial in the cloud, you know, to do it on a yearly basis versus a machine?
Michael Kammes (41:57):
I can give you the numbers. Would you like that?
Ben Kilburg (42:00):
Oh! Do it!
Michael Kammes (42:01):
What we find and, I use the royal we, what I find is that when you’re spinning up things in the cloud, it’s best for gigs. It’s best for projects. If you’re looking, you say we’re going to move to the cloud and it’s going to be a permanent solution. That’s when the numbers really- you really have to scrutinize the numbers, because there comes a point in time where what you’re paying in the cloud is going to equate to what you could purchase on-prem and then at the end of that term, three years, five years, whatever, you still have the hardware. So, there’s definitely a tipping point.
(42:36): But if we’re looking into storage and, I’m saying this as of December 2019, if we look into storage at the big three, so we’re talking AWS, we’re talking GCP, we’re talking, well, Azure really doesn’t have a three-letter acronym, but if we talk about Azure, those three, we’re looking at about $150 a terabyte a month for edit worthy storage. So this means if you’re editing video in a cloud, in the cloud off a VM in a data center, storage that’s going to give you hundreds of megabytes a second read and right for renders, for editing video, is going to be about $150 per terabyte per month just for the JBOD. You then need to add a server. Anyone who’s worked with Xsan is probably familiar with the fact that you have the JBOD, the just a bunch of drives, the array and then you have a server that spits it out. The cloud dings you for both and that’s about $150 bucks as well. So, you’re looking at about $300 per terabyte per month for edit worthy storage in the cloud.
(43:38): When we then look at what is our compute time, OK, cloud providers traditionally charge you per month in terms of storage, but when it comes to compute, they charge you per hour. So, list price for a 16-core machine that is 2.30GHz with 122GBs of RAM and a 16GB GPU, that runs around and, I got to say “around” because it varies, about $2 dollars an hour. So, if you took that to $2 an hour, multiply that by 40 hours in a week, because I know creatives only work 40 hours a week, and then multiply that by how many weeks in a month. That’s going to give you a ballpark number as to what you can expect.
(44:23): What we found is that a lot of the Fortune 100, Fortune 500 companies, when we talk to them, when Bebop talks to them, they say, “Well, we’re not going to pay your rates. We have a sweet deal with Amazon and we’re paying a $1.50 an hour.” Cool. Great. You, corporate America, you pay that fee that you’ve already negotiated and then you pay for the additional Glue services that’s something like Bebop would handle. And we find that that brings the cost down drastically because they’re already getting those cost savings. And to that end, as I mentioned earlier, we there’s three big data centers, though, the three biggies. If we increase that to four, we get IBM and then we get what we call “data lakes” or “private data centers” where companies, well, I’m not going to name companies, but there are companies who are building one-off data centers that can drastically undercut all the biggies and you can save a ton of money if you’re located in that area.
Jason Whetstone (45:21):
I was really hoping we could get through the episode without mentioning data lakes.
Ben Kilburg (45:27):
Right. Just again, somebody else’s storage, kids.
Jason Whetstone (45:30):
Somebody else’s storage. Exactly.
Ben Kilburg (45:32):
Right. And one other thing I thought while listening to you talk, Michael, GCS, Google Cloud Service. Right. Just in case somebody doesn’t know the acronym.
Michael Kammes (45:43):
Jason Whetstone (45:43):
Ben Kilburg (45:44):
Jason Whetstone (45:46):
A cloud platform. Right? Thank you. VM, virtual machine.
Ben Kilburg (45:49):
Jason Whetstone (45:49):
So that’s that computer running somewhere else that’s not a real machine. I mean it is, but not really.
Ben Kilburg (45:56):
Right. So it’s a- it’s an abstraction layer where you’ve got multiple operating systems acting like individual computers running inside a central computer’s platform so that you can have, maybe, eight machines running on one machine and maybe that one machine has 28 cores and you subdivide those by a certain number of workstations.
Michael Kammes (46:21):
I would encourage anyone who is still on the fence, “Will the cloud be there for media?” There was a massive announcement, I think it was two months ago, between Microsoft and Avid and Disney about a “march to the cloud” where all three are going to be working together on Microsoft Azure to start moving and adding the cloud to collective media workflows. And what’s important about that is, once Disney starts to do that, that means the Disney tentacles and all the different companies that are supporting Disney are going to have to do that. And within about a month of that announcement, we had calls from everyone doing QA and QC to people who are doing foreign language dubs and all their questions were, “We have no idea how to approach this. Disney’s already calling and saying we’re going to start delivering everything in the cloud. You need to work there.”
Ben Kilburg (47:17):
So, talk us through some of that and how you coach others. Right. Because clearly there are cloud buckets, which is a strange terminology to use, but those are the object-based storage. Right. But what we’re talking about for editing, which is essentially somebody else’s file server up there in the cloud, that would be mounted on the virtual machine acting like a desktop, right?
Michael Kammes (47:45):
You’re completely right. Just like when you’re on-premises, you have various tiers of storage. And for those who aren’t really- we haven’t spent much time on this, don’t worry, I’ve done enough for all of us. You have your Tier 0 storage, which is your fast SSD or NVME sitting in your computer. And then we have your Tier 1, which is usually a SAN. Maybe it’s a high-performance NAS, but I’m not going to get into that argument today. Tier 1 is usually SAN. Then you have Tier 2 as a NAS and then Tier 3 is your LTO, etc. And they all vary on accessibility and availability. When we start getting to the cloud, the same paradigm applies. We have the object storage, as you pointed out, and that’s more of a Tier 2.
(48:31): Tier 1 is usually a file system. AWS calls it “EBS” and it’s a way of delivering content fast from spinning disk or SSDs without using the underpinnings of object storage. Object storage is fantastic for resiliency, but sucks for latency. And so, what we’re finding is that there are companies, I won’t name any of them, but there are companies who are saying, “Well, we can just put fast cache in front of the object storage, and you can edit off that.” And I tell people to run from that as fast as possible because you don’t get the performance you do from your Tier 0 or Tier 1 and you drop frames and you don’t know why.
Jason Whetstone (49:15):
Michael Kammes (49:15):
So, we have a lot of clients who are saying, “We’re going to move everything to object storage and then we’re gonna selectively move things or sync things from the object storage to the fast editing volumes in the cloud. And we’re going to do that in a methodical way so we’re only spending the minimum that we need for the fast storage but keeping everything else on slower and thus cheaper storage.”
Ben Kilburg (49:38):
Jason Whetstone (49:40):
Sounds like a great cost savings, but maybe at quite a cost of just efficiency and frustration of users potentially.
Michael Kammes (49:49):
Well, you know, just like we talked about having a full time IT person, when you’re talking media management in the cloud, that’s a little bit more tough because, I hate to sound like an old man here, but there’s nothing for you to grasp on to. There’s no LTO tape on a shelf or XDCAM disk on the shelf or no spinning disk you can point to; it’s all via a web page. So, I completely understand that that is weird and foreign to some people, but I think it’s a necessary evil and something that we’re gonna have to make that move to be able to handle.
Ben Kilburg (50:22):
Gotcha. So, are you guys building any automation platforms to help people do this? You know.
Jason Whetstone (50:27):
I’m just going to get into that. Thank you, Ben.
Ben Kilburg (50:29):
And so, I’m thinking like I’m a client. Right. And I know, you know, maybe I’m a big Fortune 100 company and I do some external comm stuff, and I only need to set up these edits with freelancers that I bring in. Maybe I’ve got one in-house video guy, but I bring in a team of freelancers and that every few months I might have a big project for some product announcement or something. If I have something like a BeBop account, how do I spin that back up again quickly so that I can get my folks editing quickly?
Michael Kammes (51:01):
When it comes to BeBop, BeBop certainly handles something like that where you use it as you need. You pay a subscription every month. You use it for as many workstations as you’re paying for. When you don’t need it, you spin it down. And then we use a couple of different tools: we have our own Rocket, what we call “rocket transfer”, which allows you to sync media from one location to another, whether it be on-prem to cloud, cloud-to-cloud or cloud to on-prem, and that certainly can be used when it’s needed. But in terms of storage, that’s obviously something you can’t not use. So that’s when we tend to either use Rocket to move media, move content to object storage, or we talk to the end-user about a MAM, a DAM, a PAM, and can that manage your media. And a lot of Fortune 100 and Fortune 500 clients do have their own DAMs, MAMs and PAMs and it’s just how much has the cloud been integrated into those solutions?
Ben Kilburg (51:55):
Gotcha. So, with our media asset management platform, we might be able to build some fancy automation is to say, let’s spin something up and BeBop and get back to work.
Michael Kammes (52:05):
Yes, and quite often, because cloud is expensive for storage but great for editorial, a lot of MAMs, DAMs and PAMs will generate proxies, automatically push those proxies up to the cloud, to your object storage, and when it’s time for the editors to edit, it’ll selectively move it to the EBS or fast storage, if you’re on Amazon, obviously, and then when you’re done, it moves everything back. So it keeps your fast storage at a minimum, keeps your costs down, and, a lot of times, the asset management systems are tracking all that media in real time, so you have an on-prem solution that’s tracking all that as opposed to something esoterically in the cloud.
Ben Kilburg (52:41):
Cool and EBS, elastic block storage, right, versus something like what is it, the elastic file system, EFS, if we’re talking about Uncle Jeff’s magic bucket.
Michael Kammes (52:52):
Whose Jeff and why does he have a bucket?
Ben Kilburg (52:58):
He’s got all the buckets. All the buckets.
Michael Kammes (53:02):
I think if you were to take a look at pricing, if you look at pricing now for storage, you’ll see that it’s, I don’t want to say, “stabilized over the past couple of years”, but the storage cost per gigabyte, cost per terabyte, has kind of stabilized a little bit, and I think we’re going to start to see that dropping probably mid to late next year as more, I’m going to say it, data lakes, pop up and these private data centers pop up because they’re able to undercut. If you looked at Backblaze, for example, they don’t do any compute; it’s all storage. And I think they’re, what, a quarter of the cost of, I think AWS or GCS or what not? The pricing is so inexpensive. So yeah, we’re going to start to see that happen a lot more.
Ben Kilburg (53:44):
They’re not charging you in the same way for egress and some of the others are. The same thing with Wasabi as well. Those are a couple brands that people really like because of their billing models are very different than some of the traditional cloud vendors. The Big Three, as we mentioned earlier.
Michael Kammes (54:02):
Another point to bring up and, this diverges, kind of, but in my mind, Ben, I’m keeping- I continually go over that something you brought up earlier, which is, “What are folks not realizing? What have they not thought of?” I mentioned earlier that the concept of latency and, you know, “We can do it faster and you won’t notice it.” And I realize you should probably put a little asterisk next to that.
Ben Kilburg (54:22):
Michael Kammes (54:22):
If you look at the speed of light or try to look at the speed of light as it shoots across the country, if you are in New York, let’s say, and your trying to edit with media that’s sitting in LA, that is an increased latency. That’s going to be 80 milliseconds, give or take. And when you add 80 milliseconds on top of what your computer is doing in terms of processing, the cloud is doing in terms of processing, that’s when you may start to hit issues. That’s why trying to farm out editing overseas is very difficult because your media is here and that means you have to push all your cloud media overseas and then sync it between the two places. So, when you start looking at decentralizing and having people all around the world edit, there comes a point in time where you say, “OK, I need to have a co-location of data centers. I need to have media in a bucket, maybe between the two. So, sitting in the middle of the U.S., so it’s equal latency for both people in LA and New York or we’re going to replicate things.” And I can give you a great example of how that worked.
(55:29): We were working with a client who had the, shall we say, lead edit in LA and the assistant was out in the Philippines. There’s no data center in the Philippines, so we had to spin up a data center in Singapore. So the editor would take all the raw content in the United States, do a selects pass, then media manage just that select pass to the data center in Singapore, which the editor in the Philippines would then edit and then return the edit list back and they would reconform. So, it can completely be done; you just have to worry about what the latency is in each location.
Ben Kilburg (56:04):
Right. And make sure you’ve got wicked good file acceleration going all over the world.
Michael Kammes (56:08):
Jason Whetstone (56:08):
Right. Right. And that was a great example of how workflow like that could happen.
Michael Kammes (56:13):
I can give you an even cooler workflow.
Jason Whetstone (56:15):
Ben Kilburg (56:15):
Michael Kammes (56:15):
And this is a client that I can actually name because they just got a spread in, I think, Nerd Weekly or something. I can’t remember what the publication was, but NHRA, the National Hot Rod Association. They have races all around the country and they have shooters that are going out and recording all these races. They’re then uploading that content using Sony C. Sony C is the asset management system that Sony has, which is all in the cloud. They upload that via Sony C. Sony C then geolocates where that media needs to be. Drops it in an archive bucket but also, they have a panel that the editors out here in LA can load up in Premiere, pull down the clips and keep working with them and then deliver everything in the cloud. So, the physical hard drives don’t have to be swapped. The graphics and video people are always accessing the same content from the same asset management system and it’s being tracked the entire time.
Ben Kilburg (57:08):
Jason Whetstone (57:08):
Yeah, that sounds really cool.
Michael Kammes (57:09):
Jason Whetstone (57:12):
Yeah. Yeah, I think we’ve taken enough of Michael’s time.
Ben Kilburg (57:14):
There you go.
Jason Whetstone (57:15):
I’d like to ask you a few questions that we have. So, I have five.
Michael Kammes (57:21):
Okay. I like that. Five different things. I like that.
Jason Whetstone (57:23):
It strikes me that a question a lot of people would have would be, “So can I edit 8K video on my laptop over a Wi-Fi connection?
Michael Kammes (57:31):
Yes, and let me explain-
Jason Whetstone (57:34):
And terms and conditions apply.
Michael Kammes (57:35):
Yeah, there terms and conditions because you’re editing 8K, great. If you’re editing, and I’m not saying VFX, if you’re editing, are you editing on an 8K display? No. So you’re already seeing it in a scaled version. So, if you’re doing this to the cloud, that’s going to apply as well. Most people on the cloud work with either HD or 4K GUIs. So already you’re only seeing a quarter or half. So, you can certainly do that if it’s running at 30 or 60 frames. If you’re looking at a Premiere interface, you can certainly see 30 or 60 frames a second at the resolution in that preview monitor. My concern is, “How long did it take you to upload that 8K file?” Right. And, “How much are you paying for it?” Because couldn’t you accomplish the same thing by creating an HD proxy, cut that proxy and then download the EDL or cut list and reconform to the 8K?
Ben Kilburg (58:27):
Yeah, and it’s like you mentioned earlier. Right? Maybe the first step in that might be uploading the high-res original and migrating that over to something like Glacier where it’s going to be safe as houses because you know, everybody has a vested interest in maintaining the best quality and making sure that those files are guarded and, hopefully, geographically distributed across multiple data centers. But you want to create that low res HD proxy that’s easy to stream and doesn’t take too much out of your wallet.
Michael Kammes (58:59):
And that brings up a good point. An even more robust workflow is uploading the 8K, having a transcoder in the cloud, then you create your proxy and then move the 8K elsewhere. What I should also mention is that any remote editing protocol worth its salt isn’t going to try and deliver an 8K payload while streaming. They’re only going to change the pixels that have changed from one frame to another. So, you’re never playing full 8K resolution down the pipe to your computer; it’s always playing what’s visible on screen and what’s changing from frame-to-frame.
Jason Whetstone (59:34):
All right. Awesome, awesome answer. Question number two. The name BeBop. Where did that come from?
Michael Kammes (59:40):
One of our co-founders actually is a gifted guitar player and he’s into jazz. And so, BeBop was kind of named for that. Also, if you look up the definition of bebop, it’s improvisational. And I think when you’re working in the cloud, you have to be nimble. You have to be able to improvise on the fly because technology is always changing.
Jason Whetstone (01:00:00):
I love that answer. I love that whole answer. That’s awesome. So, next question. Question number three, what goes into a 5 THINGS episode? And I already know the answer to this, I think.
Ben Kilburg (01:00:12):
There’s an episode of 5 THINGS about goes into a 5 THINGS episode that’s really awesome. But hey, we’ve got you here with us, so.
Michael Kammes (01:00:19):
Yeah, I was hard up for material that month.
Ben Kilburg (01:00:22):
Michael Kammes (01:00:25):
It’s approximately 40 hours per episode and it’s a lot of that’s me writing it. That’s then me, you know, sitting on it for a day and then going back and checking my zeros and ones. I then, if you’ve noticed, I do a lot of cutaways to movies or shows or things that I like. And I then I have to find out, is there a way I can weasel in the word “SAN” as a joke and then find a movie clip where someone says the word “sand”. So, there are some investigative journalism to find those clips. I then have to shoot that, which I do, and I’d explain what cameras I use and what lights I use and what sound I use. I then cut that together. I then do a color pass and then I have a myriad of exports because I do close captioning for everything. I like to make sure that I create streaming versions on different platforms. I have a Roku channel, so I have to create an output for Roku.
Jason Whetstone (01:01:19):
There’s 5 THINGS episode on that, too.
Michael Kammes (01:01:20):
Yes, there is.
Jason Whetstone (01:01:23):
That’s awesome. Great. Yeah. And you said investigative journalism, which is something that I consider all of us here, in a way, where we also kind of have to put the investigative journalist hat on when we’re trying to learn about how people work, because you’ve got to ask all kinds of interrogating questions, you know.
Michael Kammes (01:01:39):
It’s an audit. It’s a workflow audit.
Jason Whetstone (01:01:42):
Exactly. Exactly. So, it’s also a therapy session, too.
Michael Kammes (01:01:45):
Jason Whetstone (01:01:46):
You know. OK. Next one. Next one is from Ben.
Ben Kilburg (01:01:51):
Yeah. Yeah. So, on your website, you’ve got some of the interesting creative work that you’ve done. And one of the things I noticed on there were three Slayer videos. That’s pretty awesome. How did that happen?
Michael Kammes (01:02:06):
Yeah. That that was very interesting. As I mentioned, I kind of transitioned from being a fulltime creative to being more technical. And there is an editor, a very talented editor here in LA by the name of Ed Marx. And Ed Marx is also a Chicago guy. So, we kind of bonded over Italian beef and deep-dish pizza. And he knew that I used to do audio and he said, “Hey, I have this project coming up and it’s Slayer and it’s gonna be a trilogy. And working with B.J. McDonnell, who has done some fantastic stuff.” I did the cut of the first one of the music video, which is kind of the bookends, shall we say, because the central portion of the video was obviously the music. And then there’s a story that goes through all of them. And so, yeah, I did the first one and they didn’t hate it. And so, the second and third one I did as well. And recently that was all wrapped up with an ending piece. So now it’s like a final concert film. And that was released not too long ago. So, it was received real well. And it was- I like to cut an audio project a year just to see how bad my hearing has gotten and to make sure my chops are still there, and they didn’t hate it. So, I guess I’m OK.
Ben Kilburg (01:03:16):
Jason Whetstone (01:03:16):
Ben Kilburg (01:03:16):
Yeah, I certainly enjoyed it. It brought me back to the days of my heavy metal youth and Slayer was actually the pit, the first pit I was ever in, which is a good memory for me.
Michael Kammes (01:03:28):
I’m glad I could assist.
Ben Kilburg (01:03:29):
Jason Whetstone (01:03:31):
Awesome. So last question. As a fellow glasses’ wearer, I have noticed that you have some very unique glasses frames and I would like to know what brand they are.
Michael Kammes (01:03:40):
That’s a real good question. They are called mó, M with an O and like accent mark above it.
Ben Kilburg (01:03:50):
Michael Kammes (01:03:50):
Yeah. I, honestly, and you’ll probably get a kick out of this, as you look at my previous episodes and you see that I went for the traditional white guy black nerd frame and I realized, A. they didn’t stay up on my face for a well. And second of all, I look like everybody else. So, I went to like, no kidding, three different eyewear stores in LA and happened to find this at a place called Linden Optometry in LA and insurance covered them too. Even better!
Jason Whetstone (01:04:22):
Yeah, even better. That’s always a bonus.
Michael Kammes (01:04:23):
So, I think they were $150 bucks maybe. Yeah. They’re the 415M if you want to try and look it up online.
Jason Whetstone (01:04:32):
Okay. Yeah, I went into a designer framed store once and quickly realized that I couldn’t afford anything and none of it was covered by insurance. But they had some really cool stuff in there. I mean, you know, and I actually had run into a friend and I looked at her glasses and I was like, “I know where you got those.” And she was like, “Oh, yeah. Yeah, I got them there.”
Ben Kilburg (01:04:51):
Jason Whetstone (01:04:51):
Well, this has been great, Michael. Thank you so much for joining us today on The Workflow Show.
Michael Kammes (01:05:00):
Thank you so much for inviting me and letting me talk tech with you two. Thank you.
Ben Kilburg (01:05:04):
Absolutely. We really enjoyed having you. I would also like to thank my co-host, Ben Kilburg, Senior Solutions Architect here at CHESA, for his expertise.
Ben Kilburg (01:05:11):
Jason Whetstone (01:05:13):
Thank you. Thank you so much for listening. We hope you like listening to the podcast as much as we enjoy creating it. If you have any questions or concerns or stories about media asset management, cloud storage, cloud workflows, media infrastructure, integrations or any of that technology, we would love to hear from you. So, e-mail us at email@example.com. And as always, you can visit our website at any time: chesa.com. The Workflow Show is produced at the CHESA Church, our home office in Baltimore, Maryland. The show is co-produced by my co-host, Ben Kilburg, and CHESA’s Sales Operations Manager, Jessica Mantheiy. Thanks for listening. I’m Jason Whetstone.