Preview a handful of rows, then render spot checks, and only then start the big batch. Templater gives you two preview modes and a spot check feature, and each one catches a different kind of mistake before it is multiplied across thousands of videos.
The cost of a bad batch grows with its size. A wrong font size in one video is a five-minute fix. The same mistake across a long data range is a long render you throw away. Our documentation says as much in plain terms: before initiating an intensive batch process, you may want to cycle through composition previews to verify how Templater will process your data source. This article explains how each preview mode behaves, where spot checks take over from previews, what each one is bad at, and an order of operations that keeps surprises small.
Two preview modes, two different jobs
The Preview feature shows what Templater’s layout engine will do to your dynamic layers. You get to it after you have mapped data to the layers of your composition and configured their layout rules. From there you can preview a new version of the composition one at a time, or you can have Templater cycle through several versions automatically. Those two choices are the manual and cycle modes. One note on versions: we changed the preview method recently, and the instructions for modes from before Templater 2.9.0 are kept available in our help pages for anyone still on an older release.
Manual mode
In manual mode, Templater retrieves data from your data source and updates all of your dynamic layers with it. You choose the range of data Templater uses. You then step between jobs with the step icons in the transport controls, and the layers update from consecutive rows in your spreadsheet or consecutive objects in a JSON array. A Refresh icon forces Templater to pull row data again, from the start row of your range, from the end row, or from a row in between.
Use manual mode when you are looking for something specific. Say you want to see how a long name fits in a text layer. You jump to the row with the longest name and look. Say a customer reported that one row looks wrong. You refresh to that row and study it. Because you control the pace, you can stop on a row, adjust a layout rule, and refresh to see the effect on the same data.
One fact about manual mode matters a lot when you are editing your data and previewing in alternation. Templater re-reads your data source each time you preview and does not use caching. That is good news. If you fix a typo in the sheet and refresh, the fix shows up. You never have to wonder whether you are looking at yesterday’s version of the row.
Our documentation also offers a tip for speed: specify a small range of rows or objects. A preview that has to prepare a short range runs faster than one prepared for a very large sheet. When all you need is to see six rows, there is no reason to load six thousand.
Cycle mode
In cycle mode, Templater retrieves a range of data and cycles through each preview on its own. Layers display with updated data and layout, one row after another, with no clicking. You begin by clicking the clock icon. While it cycles, Templater uses the range you set in the Start and Finish fields in the transport section.
Cycle mode is for scanning. You are not studying one row, you are watching the whole range pass by, and your eye catches the oddities: the row where the text spills out of its box, the row where an image is the wrong shape, the row where the layout jumps. It is the closest thing to a flip-book of your batch. If you ever wondered whether a specific row would break the design, cycle mode is the quick way to find out.
It has a limit that you must keep in mind, and our documentation states it plainly. Templater reads the data source only one time before it begins cycling previews. Updated or newly added data will not show up in a preview while Templater is in cycle mode. If you notice a problem while cycling, fix your data, then stop and start the cycle again so the source is read fresh. Forgetting this is the classic way to chase a ghost. You correct a cell, the cycle keeps showing the old value, and it looks as though the fix did not work. It did work. The cycle simply has not looked again.
Which mode, when
Use manual mode to investigate and cycle mode to survey. A reasonable workflow starts with cycle mode over a modest range to find the rows that look suspicious. Then you switch to manual mode, jump to those rows, and work on them one at a time with the freedom to refresh after each change. After a round of fixes, run the cycle again to confirm nothing else moved.
What previews cannot tell you, and where spot checks begin
Previews live inside After Effects. They show you the composition on screen, row by row, while you watch. That is useful and it has two limits. You have to be there to see them, and you see them one at a time. A preview does not leave behind a pile of images you can hand to a colleague, scan quickly, or feed to another process.
Spot checks fill that gap. Our documentation describes them as a way to quality assure batch render or replication processes, so you can quickly see how data merges across many jobs. When you generate spot checks, you create still images or movie clip exports of individual frames or groups of frames from data-merged compositions. The exported assets are the spot checks themselves.
Which frames get exported? You decide with named markers on the composition timeline. Place markers where you want Templater to look, and give them names, because Templater only recognizes markers that are named. If you set no markers, Templater samples an evenly distributed number of frames across the target composition’s duration. That default is a sensible fallback, but markers beat it whenever you know where the risk lives. If the lower third appears at four seconds and the call to action lands near the end, put a named marker on each. Then every spot check shows you the frames that matter.
You also choose the file format for the output, and you can pass the absolute path of any spot check to a script that you have registered to a supported Templater event. That second option turns spot checks from something you look at into something a pipeline can use, for example by posting the images where a reviewer can see them. Our spot check write-up also has a section on exporting spot checks through the command line.
Two requirements apply. Spot checking needs Templater 3.0 or newer, and it is supported in the Pro and Bot editions only. If you work in another edition, the preview modes described above are still your main tool, and the sensible way to use them is to be stricter about range. Preview more rows by hand, cycle the whole range at least once, and keep a written list of the unusual rows you checked. The spot check feature itself is not available to you.
What spot checks are good at
Spot checks make it quick to spot issues with text, footage and colors, which makes problematic data easier to deal with. Think of a batch where a few rows have a missing image, an overflowing headline, or a color value that clashes with the background. A set of spot check stills lets you look for all of those across many jobs at once, which is faster than watching the render of every video.
Two newer features deserve a special mention, because spot checks are how you verify them. If you use Dynamic Essential Properties, available in Templater 3.5 or later, spot checks help you confirm that data is controlling the values of the properties promoted to the Essential Properties panel. Templater updates those properties on pre-comp layers, not on the primary composition itself, which is worth knowing when you go looking for the effect of a data value. Our post on dynamic Essential Properties in Templater explains the feature, and the one on what gets exposed and what the primary comp’s value does covers the promotion side.
If you use Time Sculpting, spot checks let you verify the effect of time-based rules that crop, trim, shift and stretch layers. To see those effects more clearly, a setting in Templater Preferences can be used while previewing target compositions. It is disabled by default and should be used only with Time Sculpting. For a longer look at the feature, read time sculpting with Dataclay Templater.
What spot checks are not
A spot check shows the frames you chose. It does not show the frames you did not choose. If a problem lives at a moment between your markers, or in an animation that plays between two sampled frames, the spot check will not show it. They are a sample, not a full review. That is exactly what makes them fast, and it is also why we describe them as quality assurance and not as proof. For anything where a missed frame is costly, choose markers deliberately and add more of them.
An order of operations for a big batch
Here is how we would put these tools together on a real job. The scenario is hypothetical, but every step uses a feature described above.
Imagine a team preparing a large batch of videos from a spreadsheet. Before anything renders, someone confirms that the project has a target composition and a valid data range, because a successful batch needs both. Then the team sets a small range and uses manual mode to look at the first few rows, then at the rows they already suspect: the longest names, the shortest, any with unusual footage references.
Next they widen the range and switch to cycle mode, and they watch. When something looks wrong they stop the cycle, fix the data, and start again, remembering that cycle mode reads the data source once at the start. Only after a clean pass do they move on.
Then they place named markers on the frames where the template is most at risk and render spot checks over the full range. A reviewer looks through the stills for text, footage and color problems. Anything odd gets traced back to its row. The team fixes the rows and generates the spot checks again for the affected range.
Only now do they start the batch itself. If you want to understand what happens when a run goes sideways anyway, what happens when a Templater batch fails mid-queue is the right companion, and batch rendering software and how Templater handles it sets out the bigger picture. Our full write-up on generating spot checks with Templater goes deeper on the feature, and the Templater help pages on spot checking, batch rendering and replicating and on preview modes cover the exact controls.
The bottom line
Use manual mode when you are investigating a row and cycle mode when you want to scan a range. Remember that manual mode re-reads your data on every preview while cycle mode reads it once. Use spot checks, in Pro or Bot, to review many jobs at once and to hand something visible to other people, and place named markers on the frames that matter. None of it replaces a human looking, but all of it moves the looking to before the render and not after.
The thing to remember is that a few minutes of previews and spot checks is cheap, and a long batch rendered from a bad data range is not.
Here are the questions we hear most about previews and spot checks.
Frequently Asked Questions
What is the difference between manual and cycle preview?
In manual mode you step between jobs yourself with the step icons, and you can refresh to pull data from the start row, end row, or a row in between. In cycle mode Templater moves through the range automatically after you click the clock icon, using the Start and Finish fields in the transport section. Manual mode suits investigating one row. Cycle mode suits scanning a whole range.
Why does my fix not show up while a preview is cycling?
In cycle mode, Templater reads the data source only one time before it begins cycling. Updated or newly added data will not appear in a preview until you stop and start the cycle again. Manual mode behaves differently, because Templater re-reads your data source each time you preview and does not use caching. If you edit data often, manual mode shows the change right away.
How can I make previews faster?
Specify a small range of rows or objects. Our documentation gives this as the tip for faster previewing in manual mode. A short range means Templater has far less to prepare, so you see each preview sooner. Widen the range later, once the layout looks right, when you want to scan more of the data in cycle mode.
What is a spot check in Templater?
A spot check is a still image or movie clip export of individual frames or groups of frames from a data-merged composition. You generate them to see how data merges across many jobs and to quality assure batch rendering or replication. You can choose the output file format, and you can pass the absolute path of a spot check to a script registered to a supported Templater event.
How do I choose which frames a spot check exports?
Add named markers to the composition timeline. Templater exports the frames at those markers, and the markers must be named for it to recognize them. If you set no markers, Templater samples an evenly distributed number of frames across the duration of the target composition. Place markers on the moments where your template carries the most risk, such as where a lower third or a call to action appears.
Which editions and versions support spot checking?
Spot checking requires Templater 3.0 or newer, and it is supported in the Pro and Bot editions only. Verifying Dynamic Essential Properties with spot checks relies on that feature, which is available in Templater 3.5 or later. The preview modes described here are part of Templater’s preview feature, and the method changed in Templater 2.9.0, with the older instructions still available in our help pages.
Can spot checks help verify Time Sculpting rules?
Yes. Spot checks let you verify the effects of time-based rules that crop, trim, shift and stretch layers. A setting in Templater Preferences can make those effects easier to see when you preview target compositions. It is disabled by default and should only be used with Time Sculpting. Place named markers where the timing changes so the spot checks capture the frames you want to compare.
