How to turn a market analysis recording into a news broadcast
Upload the recording and a photo of the presenter and tell Vizard Agent the channel branding. Vizard Agent transcribes the commentary, cleans the photo, drives a presenter from it in segments short enough to hold, sources market footage to cut against, and builds the broadcast graphics around it.
What is the short version?
A recorded market commentary is audio with a person's name attached. Turning it into something watchable means putting a presenter on screen, cutting to relevant footage while they talk, and wrapping it in the visual furniture that makes a broadcast read as a broadcast.
- Go to Vizard Agent and upload the audio and a photo of the presenter.
- Say the channel name and how the branding should look.
- Say whether you want subtitles and in which language.
What do you need before you start?
Audio and a photo. Vizard Agent transcribes the commentary, sources the supporting footage and builds the graphics, so those two files are the whole input, and a clean front-facing photo matters more than anything else you could supply.
- The recording. Your commentary, any length.
- The photo. Front-facing, clean background, decent resolution.
- The channel branding. The name and how it should appear.
- The language. For the subtitles and the fonts.
- The date. Broadcast furniture usually carries one.
What do you type into Vizard Agent?
Name the branding and the presenter. Vizard Agent builds the lower thirds, the title cards and the station identity from what you tell it, and being specific about the channel name saves a round of corrections on graphics that appear in almost every frame.
Prompt
Variants worth knowing:
- Broadcast graphics. Lower thirds, a ticker, a station bug.
- A specific date on screen. Vizard Agent will check it against the recording.
- A specific sector. Say if a particular index or industry should be shown.
What does Vizard Agent actually do?
Here is the order Vizard Agent worked in on a real market broadcast. Two problems shaped the whole job: a stray mouse cursor sitting on the presenter's photo, and a narration far too long to drive a presenter in one pass.
- Opens the photo and checks the audio file, then transcribes the commentary.
- Searches for market and economic footage and reviews the candidates as an overview sheet.
- Checks the available fonts, downloads one for the language and verifies it installs.
- Checks the photo's resolution and cleans it before use.
- Locates a mouse cursor left in the corner of the photo and tests removing it against the white background.
- Starts the presenter render and finds it fails on a six-minute pass, then reads the logs to see why.
- Tests what segment length actually holds, reads the word-level timeline and plans the split points.
- Generates the presenter in batches, renders the background footage and joins it.
- Builds the station graphics, sizes the subtitles by trial on frame, measures speech against music, then swaps weak footage, corrects the card text and re-renders the final programme.
Step 7 is the one worth knowing about. A presenter generator has a practical ceiling on how much audio it will hold in one pass, and the way through is to find that ceiling by testing and then split on natural pauses rather than at arbitrary intervals.
What does the result look like?
From the run this page is written from, probed on the delivered file: 1920x1080, H.264, 30fps, 392.0 seconds, AAC audio. Widescreen, six and a half minutes, a presenter driven from a single photograph across the whole programme, cut against market footage, with station branding and subtitles throughout.
Six and a half minutes from one still photograph is the part that surprises people. It only holds because the narration was split into segments that each render cleanly and then joined on pauses where the seams do not show.
When does this not work well?
Market commentary is regulated speech in most jurisdictions, and wrapping it in a broadcast look makes it more persuasive without making it any more accurate. Vizard Agent builds the programme to a professional standard, and every question about what may actually be said stays with you.
- Financial commentary is regulated. What you may claim, and what disclaimers are required, varies by jurisdiction.
- A broadcast look implies authority. That is exactly why the content needs to be sound.
- The presenter is generated from a photo. Say so; do not present it as filmed footage.
- Only use a photo you have the right to use. With that person's agreement to appear this way.
- Figures on screen need checking. Vizard Agent puts up what the commentary says.
How do you fix a result that came back wrong?
Name the segment or the graphic. Vizard Agent keeps the transcript, the segment split points, every generated presenter clip, the sourced footage and the graphics, so a correction is a re-render of one part rather than the whole programme.
- "That footage does not fit the section." Swapped, as this run did.
- "The date on the card is wrong." Corrected and re-rendered.
- "The subtitles are too small." Resized by trial on frame.
How does Vizard Agent compare to doing it yourself?
By hand this is booking a studio or settling for audio only, licensing market footage clip by clip, building every broadcast graphic in a design tool, and then discovering that your presenter tool refuses a six-minute file with no useful error message.
| By hand | Vizard Agent | |
|---|---|---|
| Getting a presenter on screen | Film one, or go without | Driven from one still photograph |
| A long narration | Fails, unclear why | Ceiling found by test, split on pauses |
| A blemish in the source photo | Ships with it | Found by coordinate and cleaned first |
| Broadcast graphics | Build each by hand | Generated and checked on frame |
Common questions
How long a recording can it handle? This run was six and a half minutes, split into segments by Vizard Agent and joined back together on the pauses.
Do I need video of the presenter? No. One clear front-facing photograph is enough for Vizard Agent to drive a presenter across the whole programme.
Will it look like a real broadcast? It carries the graphics and pacing of one, which is why the content underneath has to be sound.
Can it work in my language? Yes. Vizard Agent downloads and verifies a font that renders the script before building subtitles.
Where does the market footage come from? A stock library. Vizard Agent reviews the candidates as an overview sheet and swaps out the weak clips after seeing them in the cut.
Why split the narration? Because the presenter generator has a practical limit per pass. Vizard Agent finds that limit by testing and splits at natural pauses so the joins are inaudible.
Can it add a date or a ticker? Yes, as part of the station furniture, checked against the recording.
Does Vizard Agent measure the mix? Yes. Vizard Agent measures the speech sections against the music-only sections and re-measures after each change to the programme.
Does it check the finished programme? Yes. Vizard Agent reviews frames from both halves and measures the mix balance before delivery.