Projector has its own vocabulary, and the words map onto the settings you will see in the app. Learn the words and the interface is easier to read. Skip them and you can end up configuring the wrong level, such as a payload setting when you meant one video.
This page is a working glossary for a producer, developer or agency lead who is about to run a real pipeline through Projector. We split it into the terms that belong to Dataclay Station, the video concepts behind the settings you will see, and a worked example that uses both. If you want the overview of what the product is first, read what Projector is and come back.
The terms that belong to Projector
Start with the unit of work, because everything else hangs off it. A mission is one job. It can include any or all of five tasks: export data-merged Autografs to digital files, transcode the versioned output, deploy the output to a storage and streaming platform, log metadata from those platforms back to your data source, and clean up the scratch files the other tasks created. One video, start to finish, is one mission. You do not have to use every task, and a mission that only transcodes is still a mission.
What a mission does
One mission can run five tasks
-
ExportExport data-merged Autografs to digital files.
-
TranscodeTranscode the versioned output.
-
DeployDeploy the output to a storage and streaming platform.
-
Log metadataLog metadata from those platforms back to your data source.
-
Clean upClean up the scratch files the other tasks created.
A payload is a collection of one or more missions. This is the word to hold onto, because many settings are configured at the payload level. The data source, the transcode settings, the streaming service, the archive location and the global streaming settings all belong to the payload, and they apply to every mission inside it. When you ask “where do I change this?”, the answer is “on the payload” for the settings listed here. The consequence is simple. If two groups of videos need different settings, you can split them into two payloads.
An Autograf is the digitally authored file at the start of the chain. It holds compositions that you set up with Templater to use data elements merged from a data source. You set Autografs up with Templater in After Effects before Projector is involved. If you are new to that half of the pipeline, how to render versioned video in a batch process with Templater covers it.
A replicate is related but different. It is a Templater feature that generates a new After Effects project (AEP) file based on Templater versioning, one file per version. These files are the replicates. Projector renders them into full videos using parallel processing. So an Autograf is the authored template, and a replicate is one versioned copy of the project made from it. Our support docs describe a replicate as an authored document, not a rendered video. That distinction matters later, when we get to export.
Template vs. copy
Autograf and replicate are not the same file
- What it is
- AutografThe authored template
- ReplicateOne versioned AEP copy
- Made with
- AutografTemplater in After Effects
- ReplicateTemplater versioning
- Rendered video?
- AutografNot yet rendered
- ReplicateAuthored document, not a rendered video
An Autograf is the authored template, and a replicate is one versioned copy of the project made from it.
The data source tells Projector which videos to process and stream. Two are supported: Dataclay Que and Google Sheets. A payload takes one data source, not several. If you use Que, you can tell Projector to process all ready jobs or only the ready jobs in selected campaigns, and you can filter a campaign by job ID or by date created. If you use Google Sheets, the sheet must always include the columns render-status and stream-status. Those two columns are required. Our post on Projector stream status walks through what happens to a row as it moves, and Que has its own page if you are choosing between the two sources.
The export feature is where Projector earns its keep on speed. By default After Effects processes one video at a time. Export lets Projector render full videos from Templater-versioned replicates so that several missions run at once. The unit that makes this possible is the worker. Projector can use multiple cores on your computer to process missions in parallel, and each core used for this is called a worker. A slider at the bottom of Mission Control sets how many of your cores Projector may use. By default it uses half of them. If you notice slow rendering or heavy RAM use, you drag the Number of Workers slider to the left and save.
The docs give a quad-core machine as an example, where you can process four videos in roughly the time After Effects would take to process one. Read that as a rule of thumb and not a benchmark. Your own footage, effects and codecs decide the real number, so time a real batch before you promise a deadline.
Before your first payload
What to confirm before you configure a payload
- Count how many payloads you need, split videos with different settings into separate payloads.
- Choose your data source: Dataclay Que or Google Sheets (one per payload).
- If using Google Sheets, confirm the sheet includes render-status and stream-status columns.
- Time a real batch before you promise a deadline.
- Know your video lengths before you type a Poster Time number.
- If you create GIF clips, enable archiving so those settings are not ignored.
Mission Control is the part of the interface for watching distribution activity and individual mission progress. It shows each mission’s status graphically. If a payload has a few short missions, it may complete quickly. Mission Control becomes more valuable when the payload holds many missions, when the videos are long, when you use several of Projector’s capabilities at once, when you need to show someone else the progress (a screenshot of Mission Control works), or when you are juggling other tasks while the batch runs.
Projection is the automation tool. Once Projection is enabled, every step you configured for the payload is automated. You enter a number of seconds in the Check every field, click Enable, and Projector polls your data source on that schedule, looking for new jobs. Once you enable it, it keeps running until you disable it. After each check, Projector displays a message saying what it found.
The word distribution is the core activity Projector is built for. Everything above is in service of it.
Two more terms cover where the finished video goes. A streaming service is the platform where users find and stream your videos. Projector supports Vimeo, JW Player and Frame.io. You can authorize several of them and choose which one each payload uses, but one payload cannot send videos to more than one streaming service. A storage service is the cloud location where archived videos and their assets are stored. If you choose not to archive, you do not need one. When you do archive, Projector writes to an S3 bucket that you provide and manage, in an account of your own.
Where output goes
Streaming and storage services Projector supports
| Service | Type | Limit per payload |
|---|---|---|
| Vimeo | Streaming service | One streaming service only |
| JW Player | Streaming service | One streaming service only |
| Frame.io | Streaming service | One streaming service only |
| S3 bucket | Storage service | Bucket you provide and manage |
Finally, streaming globals are the default settings you configure for videos deployed to a streaming service. They cover privacy, posters and grouping, such as a JW Player playlist. You choose per payload whether to apply them.
The video concepts behind the settings
The second table in our docs covers ordinary video production ideas. They are there for anyone who comes from the data or development side and has not had to think about delivery before. Every one of them shows up as a field in Projector, so it helps to know what the field is asking.
Serial processing and parallel processing are the pair that explains why Projector exists. After Effects supports only serial processing: one video, then the next. It is also called sequential processing. Parallel processing lets each core in your computer work on a different video at the same time. The more cores you have, the more videos you can process in parallel. One video per core is the picture to keep in mind. That is why a batch of hundreds of versions stops being a single long queue.
Transcoding converts an audio or video file from one codec to another. The reason to do it is delivery. Raw, uncompressed video is very large, which makes it hard to stream and store cost-effectively, so files are compressed before storage or transmission and decompressed on playback. If you are sending video straight to viewers on limited bandwidth, you pick a codec that produces a smaller file. Not every video needs transcoding, but when yours does, Projector can automate the step.
A codec is the compression and decompression method. Video and audio each have one. For video files, the audio codec has to be compatible with the video codec. The Projector transcoder asks you for both. For video, the docs name H.264 as the choice for MP4, optimized for streaming and good quality even on older or less powerful devices, and they say you may want something different for very high-definition work. For audio, if you deploy to a streaming service, the docs say to always select AAC, because the streaming services require audio in that format. They recommend an audio bitrate of 128.
A file format is not a codec. MP4, AVI and WMV are wrappers around the package of video, audio and other streams that make up a rendered video. A codec is the algorithm inside the wrapper. Different formats support different codecs. In Projector the codec you select determines the file type of the finished video. One caution from the docs: if you do not enable export, Projector uses that file type for every video in the payload, even if your Google Sheets data source has an output-ext column that says otherwise.
Bitrate and FPS trade quality against size. FPS is frames per second. Higher FPS means better quality, but also bigger files and longer render times. The docs give 30 FPS as the figure for video watched on a computer monitor, and for that case they say video bitrate should generally fall between 3,500 and 6,000 Kbps. Higher bitrates give better picture and sound at a larger file size. Lower bitrates shrink the file and cost you quality. The docs say that if you are new to streaming you may need to experiment within the recommended range, and that is honest advice. Run a short test payload, watch it on the device your audience uses, and adjust.
The transcoder also asks for a pixel format, and the docs recommend yuv420p unless the videos will be viewed on something other than a standard computer screen or mobile device. And you pick an output width or an output height. Choose one and Projector adjusts the other to keep the aspect ratio. If you do not know how your audience watches, you can leave sizing unset. Larger dimensions mean larger files, and that raises upload, delivery and bandwidth costs.
A poster is the still image used to promote your video on a streaming service. It is also called a thumbnail. Projector lets you pick the exact frame. This is worth the effort, because the streaming service otherwise builds the poster from the time you enter, and that frame may or may not be a personalized or attention-grabbing image. Dataclay strongly recommends that you create the poster inside Projector. They also warn about a trap with the Poster Time field. If a video is three seconds long and you enter 3.5 seconds, the service falls back to the first frame. When your customized videos have different lengths, a single timestamp lands on the intended frame for some videos and not for others. Know your video lengths before you type a number.
Posters have a cost too: Projector takes time to extract a still and convert its format for each mission. Across hundreds of videos lower image quality saves time, but the docs suggest the best quality you can configure while still controlling production time.
A GIF clip is a short excerpt of your video saved as a GIF, which is friendly to bandwidth. You set when the clip starts and how long it lasts. Streaming services do not use GIF clips. You use them to promote the video on social media, in email or in text campaigns. If you only plan to stream the video on a streaming service, you probably do not need one, and the more videos you produce, the more time it takes to make a GIF for each. If you do create them, the docs say to always archive them as well.
Archiving saves completed assets to a cloud location for historical reference or possible future use. Along with the rendered and transcoded video, you can keep the poster, the GIF clip or both. The location has to be an S3 bucket that you provide and manage. One warning is worth repeating here, because it bites quietly. If you set a payload to archive posters or GIF clips in the transcode settings but never enable and set up archiving, Projector skips archiving completely and those settings are ignored. Projector skips archiving completely.
Streaming itself is the process by which a video on a remote server or cloud platform plays on a viewer’s own device. That is why bitrate, codec and size matter.
One pipeline, using the words correctly
Here is the glossary put to work. Maya is a video producer. She renders 500 short promotional videos from one composition. She personalizes 499 of them, and she will send those to individual customers in an email campaign. She also plans to stream a non-personalized version on her Vimeo channel and promote that link in a social media campaign.
Her first move is to count payloads. The two groups need different settings, so she builds two. The personalized payload is aimed at mobile viewing, since she expects her customers to watch on phones, and that shapes its transcode settings. It gets a GIF clip for the email campaign, and a personalized frame as the poster to raise engagement. The non-personalized payload is transcoded for streaming on Vimeo. It needs no GIF clip, and its thumbnail comes from a different frame. From her data source, Maya picks which videos belong in each payload.
Notice how the vocabulary did the planning. Each video is a mission. The settings differ, so the payloads differ. The privacy default, the poster and the grouping are streaming globals, set per payload. The GIF clip is a transcode choice, and because she creates GIF clips, she archives them. What she needed was to know what level each word lives on.
Now the same example from the engineering side. Her Autograf is built in After Effects with Templater. Templater generates a replicate for each version, one AEP file each. With export enabled, Projector renders those replicates into finished videos, with workers handling several at a time. With Projection turned on, Projector checks the data source on its timer, picks up new jobs and moves them through the steps configured for the payload. When a mission finishes deploying, Projector writes metadata back to the data source from the streaming service, the storage service or both. Our docs say that happens but do not list the exact fields, so check your own sheet or campaign after a test run to see what lands there.
Pairs that look alike and are not
A few words invite mix-ups, so it is worth pinning down the pairs.
Exporting and rendering are not the same thing in Projector. In After Effects, rendering a composition into a finished video is sometimes called exporting. With Autografs you have more options. You can render separate Autografs, create Autograf replicates within your AEP file, or create separate AEP replicates from each Autograf. An AEP replicate is an authored document and not a rendered video, and the docs say export and render are two separate steps in Projector. Do not use the words interchangeably in a ticket or a handoff note, or the next person will configure the wrong thing.
A mission is not a payload. The mission is one job and the payload is the group. Settings live on the group.
A codec is not a file format, as covered above. A poster is not a GIF clip. The poster is a still that the streaming service uses. The GIF clip is for places the streaming service does not reach.
Streaming globals are not the last word. They are defaults. If only one or two videos in a payload need something different, change those on the streaming service itself. If a larger set needs different treatment, split it into a second payload, as Maya did.
The data source is not the streaming service. The data source says what to process. The streaming service is where the result goes. The limit sits on the streaming side: one streaming service per payload.
The bottom line
Start with five words: mission, payload, replicate, worker and Projection. They cover what a job is, where its settings live, what gets rendered, how much of your machine does the work, and what makes it run without you. The video concepts matter in proportion to how much you care about delivery, and for a streaming pipeline that means codec, bitrate and posters. Treat the ranges in our docs as starting points and test a short payload on the devices your audience uses. And when two videos need different treatment, make a second payload instead of fighting one set of settings. If you are still deciding which Dataclay product covers which part of the work, Que versus Projector is the next read, and the full set of guides is on our Learn page.
That is the vocabulary. Here are the questions we expect a producer to bring to it before they set up a first payload.
Frequently Asked Questions
What is the difference between a mission and a payload in Projector?
A mission is one job: it can export Autografs to files, transcode them, deploy them to a streaming or storage platform, log metadata back to your data source, and clean up scratch assets. A payload is a collection of one or more missions. Many settings are configured at the payload level, including the data source, transcoding, streaming service and archiving. So every mission in a payload shares those settings. If some videos need different settings, put them in a different payload.
Is exporting the same as rendering in Projector?
No. In After Effects, rendering a composition into a finished video is sometimes called exporting. In Projector, they are two separate steps. Replicates are authored AEP documents and not videos. The docs say export and render are two separate steps in Projector, so the terms are not interchangeable, and Projector can also transcode an already finished video into new files. Keep the two words apart when you write handoff notes so nobody enables the wrong option.
How many cores does Projector use, and can I change it?
Each core Projector uses is called a worker, and by default Projector uses half of the available cores. You can change this with the Number of Workers slider at the bottom of Mission Control. If you see slow rendering or excessive RAM use, drag the slider to the left and click Save. The docs give a quad-core machine as an example, where you can process four videos in roughly the time After Effects needs for one. That is a rule of thumb, so time your own batch.
Which data sources and streaming services does Projector support?
Projector takes its jobs from Dataclay Que or Google Sheets, one data source per payload. For streaming it supports Vimeo, JW Player and Frame.io. You can authorize several services and pick one per payload, but a single payload cannot send videos to more than one streaming service. For archiving, it supports S3 buckets in an account you provide. The docs say more options are planned.
Which columns does a Google Sheets data source need?
Two columns are always required: render-status and stream-status. The docs also say that when a mission finishes deploying, Projector writes metadata back to the data source. If you set the payload to transcode without enabling export, remember that the codec you choose sets the file type, even if the sheet has an output-ext column.
Do I need an S3 bucket to use Projector?
Only if you want to archive. If you choose not to archive, you do not need a storage service. If you do, Projector archives to an S3 bucket that you provide and manage, and it needs credentials for a user account that can work with that bucket. If you did not set up your own cloud infrastructure, ask your administrator to do it. Also remember that if you set a payload to archive posters or GIF clips, you must enable archiving, or those settings are ignored.
What privacy setting should I use for personalized videos?
Our docs recommend the setting where no one except the owner can access the video as the best default. It lets you play each video through, check that everything is right, and stage videos before a marketing campaign. When you are ready, you change the privacy setting directly on the streaming service. The settings that let anyone, or any member of the platform, view the video are described as poor fits for personalized videos, because not everyone should see every video.
Should I create a GIF clip for every video?
Only if you will use it. Streaming services do not use GIF clips, so if you plan to stream and nothing else, you probably do not need one. A GIF clip is useful for promoting a video on social media, embedding it on a website, or sending it in an email or text campaign. The more videos you produce, the more time it takes to create a GIF for each one, and there can be storage implications. If you do create them, archive them too.
