End of Week Update - 004

By A.A. Lopez

    This week has a bit of ups and downs, but feel like it was mostly positive, and it's ending on the start of my vacation for the week. Also the fact I got a lot more out than I originally thought I might, and looking at using new tools.

    Let's take a look:

• Monday
• I cancelled the meeting on Monday, I didn't have any updates on it
• Did do some work on the OCR tool, but it's only catching items with @ symbols
• And did some minor work on Project://JUNO, actually got a few UI mockups built up for the Seeker and Matchmaker home screens
• Tuesday
• I got a lot more documentation of the Logger extension done, noted by that full article I put out
• Still lacking on R.v.1.0
• Also worked on the production planner as noted in the Journal entry 
• Wednesday
• Didn't show off R.v1.0 either
• Left for home right after being in the office, so had to drive from 5pm to 10:30pm
• Thursday
• Just did my day job and relaxed afterwards
• Friday
• Got to the hotel, but worked until 1 for my half-day
• And now getting prepped for my vacation

    So now I'm just getting ready to rest up, enjoy my time at the beach and then sleep. I'll probably skip this week on my SoW and EoW updates. See you next time.

 

LESSON OF THE WEEK: Senoritis also affects right before vacation 

Building Up Anytime and Always Studios - Studio Journal - Issue #1

Building Up Anytime and Always Studios

Studio Journal - Issue #1

Creative Brief - Voices in Transit - Portfolio Build

By A.A. Lopez

Ey, I Don't Have a Boss. Do I?

    I'm not the biggest supporter of generative AI in the creative field. I don't see it as much more than something that takes away the hard work of other people and mashes it together to a marketable paste. Now do I think there is no market, or value to AI? No, I think it definitely has a space in certain areas, and for people who need certain help in avenues that they didn't realize they could use. After all I've had ChatGPT help me in learning how to code because I don't learn in the usual fashion.

    When I was growing up I learned the Pythagorean theorem before I knew 1+1. Why? Because I had my brother and mom teach me them to me. I have always learned backwards. So using ChatGPT to tell me how to build a code let's me reverse engineer it and understand what the code means more than slowly building up from the annoying "hello, world" script that each language starts with. I'm not saying the traditional path isn't good, I'm saying it doesn't fit me.

    Similarly, I tend to be more logical but I'm still a creative at the end of the day. I love to create, but I'm better at building up ideas or working off of them. Or I can give the idea, but I need help fleshing the ideas out more. For instance a few years ago I was working on putting together a horror anthology podcast, I had the format and idea, but I got some friends to be writers. I did my own script, but I was unable to maneuver it the same way as I was able to direct them.

    So the hardest part is building a portfolio when you don't know what it is you want to do or how you want to go about it. This is where ChatGPT comes back into the conversation. After thinking on it for some time I decided to ask our little AI friend for the following prompt:

    "Act as my creative director at a simulated job in a creative field (e.g., media production, advertising, content design). Assign me projects with realistic constraints: timelines ranging from a week to several months, varying stakeholder demands, and specific creative briefs. For each assignment, provide:
The product or subject
The project goal
The target audience
The tone and style
The final deliverable format (e.g., video, blog, storyboard)
What assets are already available in-house (e.g., art, music, scripts)
The budget constraints (hypothetical, for planning)
Throughout the process, simulate real-world changes: stakeholder feedback, shifting priorities, scope creep, deadline moves, etc. At the end of each project, require me to package it as if submitting a real-world deliverable."

    And it gave me an ad campaign for an anthology podcast focused on creepiness (so just up my alley). It did it best to give me a breakdown of what it was looking for and a schedule. 

    The goal was: a 60 second audio spot to tease the series and drive first week listens, with the ad also running on other fiction podcasts. Giving me the target audience, tone & style, as well as the expected deliverables. To be honest when I read through it it was just word soup, I didn't know how to really parse it. So I did what felt the most natural, utilize a new tool, Lucid.

My Own Crazy

    Utilizing the textboxes and group mechanic of Lucid I put together this Creative Brief, a way for me to understand. Showing off the Product, the Target Audience, the Tone, the Budget, In House Resources; all in a way that was easy for my mind to comprehend. Then I was able to start asking questions.

    You can see that an Episode Guide was added because I noticed when I put this together that it was part of the In House Resources but it wasn't given right away. Since I found out I asked for the episodes. Which in turn gave me some ideas.

    Pictured at the bottom you can see them:

Option 1: A person goes to sit down on a bus and the person they sit next to what they're listening to and tells them about "Voices in Transit" and plays snippets a la The No Sleep S6 intro

Option 2: Two coworkers get ready to leave and one asks the other what they're listening to and talk about "Voices in Transit" but then ask "wait aren't we on route 88" with the bus driver coming up and cackling like the cryptkeeper

Option 3: Standard voice read of what the podcast is about

    Each felt like a part of a story that could be told. Well 3 was just literally the boring safe version any stakeholder who just wants it done to have. Nonetheless, it did help me figure out which ideas to present back to my "boss" AI.

Arguing the Case

    The stars show how they were received. Option 3 was in the top 2 originally, with option 2 being the best and considered to be better between them all. However, I did like option 1 the most and decided to fight for it. Citing reasons to go for it, such as, only requires one person so don't have to worry about scheduling, gives off an air of creepiness and directness, gives just enough of a free taste of what the show is about, and a bit of nostalgia for people who enjoyed the early days of podcasts.

    It actually took 3 tries to get the AI to agree to it, but in the end I agreed to do two temp scripts for option 2 and 1 and let the one that felt more true to the series be the one with no ego to it.

    From there once I finished both scripts and presented, 1 won out easily. Just had to refine it and now it is ready for the actual recording and will be put together...

What's Next

    Well it would, but I'm actually going to be on vacation come this Friday. I won't be near my equipment that I would use to do any of this. I have a few ways around this. Use my brother's equipment when I'm home (depends on how he feels about it). Record on my phone and use very limited resources, like my low power laptop (which overheats playing youtube videos). Or old school route: tape cassette. Or the option I don't like, ask for an extension from robo-boss.

    We'll just have to see what wins out, but I'm proud of how far I've gotten here. Honestly, I figured I'd just diddle around and pretend to work. Yet here's actual proof of what I'm capable of when I put even a pinky's worth of effort into something. Once I have this one done, I'll think I'll ask for another one to try out and build the portfolio more. After all who wouldn't want someone who can do this on only half a day's work?

Logger Extension - Devlog

By A.A. Lopez

    For the past couple of weeks I've been working on a new tool for my team based on an idea another member brought up. The idea comes from the fact that during our quality control checks in our media asset manager(MAM) caption issues are under reported. Various factors are the reason behind this; from the notes section being outdated, too many caption issue types to remember, to even just missing the timecode of a certain instance. So she put forth the idea of updating our note system in the MAM, however, due to future planning there's little to no actual movement on that front, which leads us into this tool.

    Instead of relying on implementing a new tool inside the MAM, why not create an extension that does the exact same thing? With the help of ChatGPT, a lot of googling, and various nights with a monster or redbull the tool, right now called the "Logger Extension" (name will change once I have a better one), is on a decent path.

Pseudo-Code 

    Starting off we have the Pseudo-code that I've put together. Here in the following image you'll see the pseudo-code laid out by both a simple path and through sticky notes thanks to Lucid.

 Here's the breakdown of the path:

>> Open MAM Video/Audio Tab
>> Extension opens automatically, Dropdown window tab appears on right hand corner of the screen
    >> Extension finds Title in <div id="cc-nav-button-label" class="label"> QC Tools - Title - S# - E# </div>
        >> Places title at the top of Extension window ignoring "QC Tools"
    >> Extension finds FPS in <div class="time-area" slot="time-input-area">FPS [NDF/DF]</div>
        >> Places FPS at top of Extension window by title
>> QCer selects timecode by selecting "New Note" or Hotkey (z)
    >> New note section generated containing a spot for the Timecode, Dropdown Menu of Issue types, Note section
        >> Extension grabs timecode found in <span class="duration">HH:MM:SS:FF / HH:MM:SS;FF</span>
            >> Places timecode into new note section
        >> Dropdown menu of issue types
            >> Nigh Impossible (8 Issue Types)
            >> Difficult (2 Issue Types)
            >> Moderate (4 Issue Types)
            >> Easy Fix (6 Issue Types)
            >> KidVid Content (1 Issue Type)
>> QCer selects issue type. Optionally may add note in note section. Optionally marks an "Unsure" check box for instances the QCer is unsure of the issue or if one does exist
>> Repeat process until complete
>> QCer then selects save/export
    >> Option to export as CSV file or Txt file. CSV is default
    >> Option to save as input for an Airtable Form as well
>> Close Browser Window
>> End

     The breakdown is pretty straightforward with what the intent is. And as you can see there's some future planning for beyond the captions.

Flowchart 

    Next is the flowchart showing off how the breakdown will be interpreted by the code. Starting as the QCer begins and then encountering a CC issue. From there they will choose to make a note and pulling the timecode and FPS from the MAM, then they will choose the issue type and include a note if they need. They will do this until all issues they can find are flagged, afterwards they will choose to save, copy, or send the notes to an Internal Storage in the extension, to a Text or CSV file, or to upload directly to the Airtable that contains the CC issues for the team that reviews them to check and fix them.

UI Mockup

     Finally, the UI design for this tool. The original design, the Silveria Design, focused more on the consistent issue types and a checklist, which seemed to put many other captions to the wayside. The updated design focuses more on a consistent level and allows each type to be considered. 

    The design is focused to be small and convenient without getting in the way of the QCer's work. With each new note it will grow in size, but max out at 3 notes and then become scrollable.

    The color scheme is focused to align with the company's colors of black, yellow, white, and a muted grey. Where the color scheme of the issue types aligns more with the teams set colors found in Airtable. However, the color for KidVid has been altered from a grey to a purple due to some minor instances "KidVid should be more colorful."

    The buttons below the Issue Type and Notes are:

• Save Note || To save the note to the internal extension's storage
• Copy to Clipboard || Copy over the note for a quick copy/paste from the note to the Airtable
• Delete || In case a note was created by accident

    The yellow buttons below are:

• Screenshot || For instances where a QCer finds something that warrants an image
• New Note (z) || To create a new note, with the hotkey present
• Export to CSV || Export to the CSV file format
• Export to TXT || Export to the TXT file format

  Conclusion  

    This Logger Extension is meant to help out the team. Be more cognizant of where and what the issues are and give a standardized note taking experience. The intent is not to limit the team, but to allow them to be proactive without sacrificing time or energy. And as we build this tool I am confident that we will see a marked improvement. Thank you for looking this over and hope to see you on the next log.