How to check that no shot in your film is too dark to see
Ask Vizard Agent to measure brightness across every segment of the finished film rather than judging it by eye. It reports which shots sit far below the rest, tests a lift on stills from those shots before touching the video, and raises only the ones that need it — so the night scenes stay dark without going invisible.
What is the short version?
Your monitor is in a dim room and it is good. Your viewer is on a phone, outdoors, at half brightness. Shots you graded as moody read to them as a black rectangle with some noise in it.
- Go to Vizard Agent with the finished film.
- Ask for a brightness measurement across every segment.
- Ask it to lift only the shots that fall away.
What do you need before you start?
The film and a sense of how dark you actually want it to be. Vizard Agent measures what is there, but the decision about how much of the mood you are willing to trade for legibility is yours, and it is worth having an answer before the comparison arrives.
- The assembled film. Segments or a single master.
- What the dark shots are for. Mood, or just underexposed.
- Where it will be watched. Phone, TV, cinema.
- Whether noise is a concern. Lifting reveals it.
- Any shot that must stay black. A fade, a title.
What do you type into Vizard Agent?
Ask Vizard Agent for a measurement rather than an opinion. "Does this look too dark to you" gets you a judgement you then have to argue with; asking for brightness across every segment gets you a ranked list you can act on, and a comparison between the shots that actually differ.
Prompt
Variants worth knowing:
- "Measure every segment." A list, not a verdict.
- "Test on stills first." Cheaper than re-rendering.
- "Readable on a phone." The standard that matters.
What does Vizard Agent actually do?
Here is how it went on a real long-form film with several night sequences. The order matters: the measurement comes first, so the conversation is about specific shots rather than about whether the film in general feels dark.
- Watches the finished film and notes the questionable moments.
- Measures brightness across every segment.
- Ranks them so the outliers are obvious.
- Separates genuinely dark shots from merely moody ones.
- Pulls stills from the worst offenders.
- Tests a lift on those stills.
- Compares the options side by side.
- Checks whether the lift brings up noise.
- Applies the chosen lift to those shots only.
- Re-checks the segment against its neighbours.
Step two is the only part that is not negotiable. Everything after it is a conversation about taste, but the measurement is what stops that conversation from being two people describing different monitors to each other.
Step three is what turns this from taste into a decision. Once the segments are ranked, the two or three that sit far below everything else stand out immediately, and the rest of the film stops being part of the argument.
Step six is the discipline that keeps this cheap. Testing a lift on a still costs seconds and re-rendering a segment costs minutes, so Vizard Agent settles the look on frames before it commits to video.
Step nine is the part people get wrong by hand. Raising the whole film to fix three shots flattens everything else, so Vizard Agent treats it as a per-shot repair rather than a global grade.
What does the result look like?
A film that still reads as dark where it should, with the shots that had gone to mud now showing what is in them. Watch it on a phone in daylight and the night sequences still have content.
Vizard Agent also tells you which segments it changed and by how much, so you can disagree with any single one of them without unpicking the rest of the pass.
When does this not work well?
Lifting shows you what was recorded, and if very little was recorded there is nothing to reveal. Badly underexposed footage comes back as noise rather than as picture, and that is a limit of the original, not of the correction.
- Severely underexposed sources. Noise is all that is left.
- Heavily compressed footage. Blocking appears in the shadows.
- Deliberate black frames. Exclude them by name.
- A film that is dark throughout. Then it is a look, not a fault.
- Different displays. No single level suits every screen.
How do you fix a result that came back wrong?
Talk to Vizard Agent about specific shots rather than about the film as a whole. Because it changed named segments rather than applying one global grade, a note about a single sequence stays a small contained adjustment instead of turning into a re-grade of everything around it.
- "Too much, it looks grey." Vizard Agent lifts less.
- "That one is meant to be black." Excluded.
- "Now it is noisy." A smaller lift, or denoise first.
- "Check the rest again." Another measurement pass.
How does Vizard Agent compare to doing it yourself?
By hand this is judged on the monitor in front of you, which is the one display that is guaranteed not to be the problem. People either miss the dark shots entirely or over-correct the whole film to be safe.
| By hand | Vizard Agent | |
|---|---|---|
| Finding dark shots | By eye, on your monitor | Measured across segments |
| Comparing | Memory | Ranked side by side |
| Trying a fix | Re-render and look | Tested on stills first |
| Scope of the fix | Often the whole film | The specific shots |
| Noise | Discovered later | Checked before applying |
Common questions
Will it flatten my night look? No. Vizard Agent lifts to legibility, not to daylight.
How does it decide what is too dark? By comparing segments against each other and against your target.
Can I set a standard? Yes. Tell Vizard Agent it must read on a phone in daylight.
Does it change the colour? Not unless you ask. Vizard Agent treats this as a brightness pass only.
What about the noise? Vizard Agent checks for it and will tell you if the lift reveals it.
Can it denoise as well? Yes, but ask separately so you see each effect.
Will it touch my fades to black? Not if you tell Vizard Agent which ones they are.
Does it work per shot or per segment? Either. Say which and Vizard Agent measures at that level.
Can I see the measurements? Yes. Vizard Agent reports the numbers, not just conclusions.
What if the whole film is dark? Then it is a style. Vizard Agent will say the spread is even.
Does this replace a grade? No. Vizard Agent runs it as a legibility check after the grade.
Can it check a rough cut? Yes, and that is a cheaper time to find out.
Will it re-render everything? No, only the segments that change.
Why does it look fine to me? Because your monitor is bright and your room is dim, which is the one viewing condition your audience will not be in.