September 8, 2025 • 8 mins
Today on Blue Lightning AI Daily, we break down ShengShu Technology’s latest Vidu Q1 update: multi-reference video generation. Imagine keeping your characters, props, and backgrounds consistent across a whole sequence—not just a shot—using up to seven reference images. Vidu Q1’s Reference-to-Video workflow aims to crush the classic problems like disappearing hats and morphing hero products, letting creators lock in continuity and reduce time lost to patchwork fixes. We compare it with rivals like Runway, Pika, and Luma and explain how more references mean fewer weird wardrobe jumps or identity drift. Find out why this helps everyone from YouTube regulars to ad agencies, plus solo filmmakers tired of reshoots and cleanup. We cover new features like 'First-to-Last Frame' seamless transitions and API-powered automation for brands. Hear practical workflows for character scenes and branded product shots, plus vital tips for prepping references and matching lighting. We also flag limitations like over-rigid faces and ambiguous input images, explain pricing and availability, and dig into how the update fits the evolving battle for AI video control. Will this replace your whole toolkit? Not yet. But for continuity, Vidu Q1’s seven-ref system might be your secret weapon. Tune in for the meme send-off and practical advice you can use today.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Happy Monday, and welcome to Blue Lightning AI Daily from Monday,
September eighth, twenty twenty five. I'm zaying your resident AI
video nerd, and yes, today's episode was made with Microsoft
Vibe Voice seven B. Pretty cool.
Speaker 2 (00:12):
What up, fam PIPA here, new week, new upgrades. We've
got a spicy one. Vdoo q one just dropped a
multi reference update. Up to seven image inputs to keep
your characters, props and backgrounds consistent across shots. I love
this for.
Speaker 1 (00:27):
Us, same quick context. This is Shngshu Technologies vdoo platform.
The headline is reference to video with up to seven
stills guiding the look of a sequence, not just a
single shot that's aimed straight at the continuity problem we
all fight in AI video.
Speaker 2 (00:42):
Yep, no more hat disappearing between cuts or your hero
product morphing into a canned ham. The announcements on PR Newswire,
if you want the receipts, they position it as more
control and less drift across edits.
Speaker 1 (00:54):
So what's actually new? Multi reference isn't unheard of, but
seven references per sequence is a pretty assertive cap. It
lets you anchor multiple characters, props, and your set like
front side back angles, plus a prop sheet and a
background plate.
Speaker 2 (01:07):
And they're saying it'll still get your prompt if you
add new actions or objects that aren't in those images.
So if your references show a red jacket and your
prompt says he sits on a motorbike, the motorbike shows up,
but the jacket stays the same. That's the idea anyway.
Semantic understanding plus visual.
Speaker 1 (01:21):
Lock Compared to the last wave of updates in the
Space runways, character reference features, Pikas style locks, Luma's identity prompts.
This leans hard into multi shot continuity. It's more than
a vibelock. It's a sequence lock.
Speaker 2 (01:33):
Big mood for who. This helps YouTubers doing recurring characters,
agencies doing branded variants, indie filmmakers doing previz, and social
marketers who ab test thumbnails and cuts. If you're doing
episodic or a mini series, this saves time you'd normally
waste patching.
Speaker 1 (01:49):
Drift in post exactly workflow impact. This won't make your
cuts for you, but it should reduce reshoots and clean up.
You can storyboard with frames. These seven images define the
world than prompt seen. It's not a total replacement for
manual comp tools, but it could knock out thirty to
fifty percent of continuity fixes in a solo pipeline.
Speaker 2 (02:08):
I'm thinking, do I ditch my save as Zoo three
final final file habit? Maybe not yet, but this gets
me closer. Also, they're folding this into the vd Q
one model, which has been pitching cinematic visuals and smoother transitions.
The PR newswire piece also nods at audio fidelity, boosts,
SFX and music generation at forty eight kilohertz.
Speaker 1 (02:27):
That's nice for drafts availability. They say it's live now
in the VDO platform's reference to video workflow. No special
invite mentioned in the press release. If you're already in
VDO Q one, you can try it.
Speaker 2 (02:39):
Price question, did you catch the bit about ten ADP
up to five seconds. PR mentions ten ADP clips and
low per clip costs. Exact pricing always moves around by plan,
but budget wise, this feels approachable for creators, not just enterprise.
Speaker 1 (02:52):
Also, Shangshu's been pushing an API track for enterprises. They've
got a VDO API and even launched a creative Leap
program for automated dot e com video production. Also on
PR newswire. That means this multi reference flow could hit
ad tech stacks and brand asset managers, not just the
WebUI translation.
Speaker 2 (03:09):
Your brand's PIM system could feed product shots straight into
a reference set, then crank variants for TikTok reels, YouTube, whatever.
If you're a growth marketer, that's chef's.
Speaker 1 (03:18):
Kiss, competitive field check. We've got Runwaygen three, Pica, Luma,
hyper kling. Most have some character style reference where vdoo's
pushing is the count up to seven, and explicitly framing
it as continuity across cuts, not just within one shot.
If it's stable, that's a meaningful advantage if we had
to meme it.
Speaker 2 (03:36):
POV, you finally stop fighting your AI about what color
the jacket is or continuity supervisor unemployed kidding, don't alt me.
Speaker 1 (03:45):
Let's talk practical setups for a two character scene ref
One two character A front and three quarter Ref Three,
four character B front and side. Ref Five wardrobe prop
detail sheet, close ups of the jacket, badge watch Ref six,
background plate of the cafe interior Ref Seven. Exterior establishing
shot for transitions. Now prompt they argue at the counter.
Barista drops a cup thunder outside the system infers the
(04:07):
cup and thunder preserves.
Speaker 2 (04:09):
The looks and for brand teams front dot back dot
side of your hero product logo, close up the brand
color card, the set background, and a lifestyle context shot.
Then prompt your seasonal line fall mood, warm, lighting, city, rain.
The model keeps the product crisp while you swap settings.
Speaker 1 (04:26):
Limitations to watch over anchoring. Too many references might make
the model rigid. Faces can look stuck. You'll need to
curate the set. Varied angles, consistent lighting, Ambiguous refs can
confuse it. Mismatched outfits or lighting directions can create contradictions.
We don't see details on watermarking or export caps beyond
that ten ADP short duration guidance in PR. If you
(04:46):
need four K long form, you're still stitching.
Speaker 2 (04:48):
Also guardrails. They don't spell out heavy content restrictions in
the announcement, but every platform has lines. If your creative
needs edgy or niche stuff, test early and shoot your
references well. Clean backgrounds, clear angles don't feed it chaos.
Speaker 1 (05:03):
Speed verse quality verse control. This leans into control. You'll
likely get steadier identity retention and fewer weird wardrobe jumps.
Speed gains are indirect you save time and revisions. Output
quality looks cinematic per vdoo's positioning, but we haven't seen
independent benchmarks in the wild yet. Facts.
Speaker 2 (05:19):
I want to see user tests, creators, side by sides,
single reference verse seven if any of you post results
on X or insta tag blue lightning so we can peep.
Speaker 1 (05:28):
Who benefits most. Indie filmmakers doing previz agencies, cranking multivariant ads,
and solo creators building repeatable characters. Hobbyists can use it too.
It's beginner friendly in concept, but your reference curation skill
matters a lot.
Speaker 2 (05:41):
Daily YouTubers, this is a w if your channel has
recurring characters or segments. For podcasters doing video promos, you
can lock your visual identity while you iterate scripts. Designers
could use it for animated lookbooks. Writers previz your scenes
before you commission art.
Speaker 1 (05:57):
One more detail from pr Newswire. They've touted first to
last frame transitions, seamless links between unrelated images or clips.
If that pairs with multi reference, you could bridge scenes
while holding identity. Think montage with consistent characters.
Speaker 2 (06:10):
That's cool for music videos and hype reels. Also, Shengshu
got named a twenty twenty five Technology Pioneer by the
World Economic Forum, also on PR Newswire. It doesn't mean
the model is best, but it signals they're serious about
creator facing controls, not just slick demos.
Speaker 1 (06:26):
Does this replace other tools? Not yet. I'd still keep
a compositor and a color tool in the stack, but
it could kick something like ad hoc frame painting to
the curb. For a lot of cases.
Speaker 2 (06:35):
Time savings guess for a thirty sixty second sequence built
from five second clips, maybe you cut your cleanup by
a third or more, depending on how chaotic your references
were before. If you're scrambling on deadlines, that's clutch risks.
Speaker 1 (06:48):
If you rely on seven references for every scene asset
preps sucks time pro tip build a reusable reference pack
per project model sheet, wardrobe sheet, hero prop set environment plates,
then drag and drop per sequence.
Speaker 2 (07:01):
And keep lighting continuity in mind. If your references are
Golden Hour and your prompt says noon office fluorescence, the
models like Girl pick a Vibe, match your reference lighting
to your intended scene lighting when you can.
Speaker 1 (07:13):
Where this fits the trend line, we're moving from single
shot wow to multi shot stories. Control and continuity are
the battleground. Vdoo's seven ref play is one clear move
in that direction.
Speaker 2 (07:23):
I'm into it. If you're comparing tools, try the same
scene across vd Q one, Runway and Peka with identical refs.
Measure identity drift and time to fix. Post your findings,
science quick source.
Speaker 1 (07:35):
Roll call before we wrap. The multi reference update and
availability are from PR Newswires vd Q one announcement. The
API and Creative Leap program details are in a separate
PR Newswire release, and the WEF twenty twenty five Technology
Pioneer nod is also PR Newswire.
Speaker 2 (07:50):
All recent meme send off seven refs, enter one consistent
timeline leaves mad max vibes. Okay, I'm done.
Speaker 1 (07:57):
That's the show. Thanks for hanging with us on blue
Light AI Daily. If you're testing vdoo Q one's multi
reference flow, let us know what holds up and what breaks.
Speaker 2 (08:05):
And hit bluelightningtv dot com for news updates and new
video tutorials on your favorite AI tools. We've got breakdowns
of the latest video model upgrades and how to put
them to work.
Speaker 1 (08:15):
Appreciate you all, Catch you next time.
Speaker 2 (08:17):
Bye,
Happy Monday, and welcome to Blue Lightning AI Daily from Monday,
September eighth, twenty twenty five. I'm zaying your resident AI
video nerd, and yes, today's episode was made with Microsoft
Vibe Voice seven B. Pretty cool.
Speaker 2 (00:12):
What up, fam PIPA here, new week, new upgrades. We've
got a spicy one. Vdoo q one just dropped a
multi reference update. Up to seven image inputs to keep
your characters, props and backgrounds consistent across shots. I love
this for.
Speaker 1 (00:27):
Us, same quick context. This is Shngshu Technologies vdoo platform.
The headline is reference to video with up to seven
stills guiding the look of a sequence, not just a
single shot that's aimed straight at the continuity problem we
all fight in AI video.
Speaker 2 (00:42):
Yep, no more hat disappearing between cuts or your hero
product morphing into a canned ham. The announcements on PR Newswire,
if you want the receipts, they position it as more
control and less drift across edits.
Speaker 1 (00:54):
So what's actually new? Multi reference isn't unheard of, but
seven references per sequence is a pretty assertive cap. It
lets you anchor multiple characters, props, and your set like
front side back angles, plus a prop sheet and a
background plate.
Speaker 2 (01:07):
And they're saying it'll still get your prompt if you
add new actions or objects that aren't in those images.
So if your references show a red jacket and your
prompt says he sits on a motorbike, the motorbike shows up,
but the jacket stays the same. That's the idea anyway.
Semantic understanding plus visual.
Speaker 1 (01:21):
Lock Compared to the last wave of updates in the
Space runways, character reference features, Pikas style locks, Luma's identity prompts.
This leans hard into multi shot continuity. It's more than
a vibelock. It's a sequence lock.
Speaker 2 (01:33):
Big mood for who. This helps YouTubers doing recurring characters,
agencies doing branded variants, indie filmmakers doing previz, and social
marketers who ab test thumbnails and cuts. If you're doing
episodic or a mini series, this saves time you'd normally
waste patching.
Speaker 1 (01:49):
Drift in post exactly workflow impact. This won't make your
cuts for you, but it should reduce reshoots and clean up.
You can storyboard with frames. These seven images define the
world than prompt seen. It's not a total replacement for
manual comp tools, but it could knock out thirty to
fifty percent of continuity fixes in a solo pipeline.
Speaker 2 (02:08):
I'm thinking, do I ditch my save as Zoo three
final final file habit? Maybe not yet, but this gets
me closer. Also, they're folding this into the vd Q
one model, which has been pitching cinematic visuals and smoother transitions.
The PR newswire piece also nods at audio fidelity, boosts,
SFX and music generation at forty eight kilohertz.
Speaker 1 (02:27):
That's nice for drafts availability. They say it's live now
in the VDO platform's reference to video workflow. No special
invite mentioned in the press release. If you're already in
VDO Q one, you can try it.
Speaker 2 (02:39):
Price question, did you catch the bit about ten ADP
up to five seconds. PR mentions ten ADP clips and
low per clip costs. Exact pricing always moves around by plan,
but budget wise, this feels approachable for creators, not just enterprise.
Speaker 1 (02:52):
Also, Shangshu's been pushing an API track for enterprises. They've
got a VDO API and even launched a creative Leap
program for automated dot e com video production. Also on
PR newswire. That means this multi reference flow could hit
ad tech stacks and brand asset managers, not just the
WebUI translation.
Speaker 2 (03:09):
Your brand's PIM system could feed product shots straight into
a reference set, then crank variants for TikTok reels, YouTube, whatever.
If you're a growth marketer, that's chef's.
Speaker 1 (03:18):
Kiss, competitive field check. We've got Runwaygen three, Pica, Luma,
hyper kling. Most have some character style reference where vdoo's
pushing is the count up to seven, and explicitly framing
it as continuity across cuts, not just within one shot.
If it's stable, that's a meaningful advantage if we had
to meme it.
Speaker 2 (03:36):
POV, you finally stop fighting your AI about what color
the jacket is or continuity supervisor unemployed kidding, don't alt me.
Speaker 1 (03:45):
Let's talk practical setups for a two character scene ref
One two character A front and three quarter Ref Three,
four character B front and side. Ref Five wardrobe prop
detail sheet, close ups of the jacket, badge watch Ref six,
background plate of the cafe interior Ref Seven. Exterior establishing
shot for transitions. Now prompt they argue at the counter.
Barista drops a cup thunder outside the system infers the
(04:07):
cup and thunder preserves.
Speaker 2 (04:09):
The looks and for brand teams front dot back dot
side of your hero product logo, close up the brand
color card, the set background, and a lifestyle context shot.
Then prompt your seasonal line fall mood, warm, lighting, city, rain.
The model keeps the product crisp while you swap settings.
Speaker 1 (04:26):
Limitations to watch over anchoring. Too many references might make
the model rigid. Faces can look stuck. You'll need to
curate the set. Varied angles, consistent lighting, Ambiguous refs can
confuse it. Mismatched outfits or lighting directions can create contradictions.
We don't see details on watermarking or export caps beyond
that ten ADP short duration guidance in PR. If you
(04:46):
need four K long form, you're still stitching.
Speaker 2 (04:48):
Also guardrails. They don't spell out heavy content restrictions in
the announcement, but every platform has lines. If your creative
needs edgy or niche stuff, test early and shoot your
references well. Clean backgrounds, clear angles don't feed it chaos.
Speaker 1 (05:03):
Speed verse quality verse control. This leans into control. You'll
likely get steadier identity retention and fewer weird wardrobe jumps.
Speed gains are indirect you save time and revisions. Output
quality looks cinematic per vdoo's positioning, but we haven't seen
independent benchmarks in the wild yet. Facts.
Speaker 2 (05:19):
I want to see user tests, creators, side by sides,
single reference verse seven if any of you post results
on X or insta tag blue lightning so we can peep.
Speaker 1 (05:28):
Who benefits most. Indie filmmakers doing previz agencies, cranking multivariant ads,
and solo creators building repeatable characters. Hobbyists can use it too.
It's beginner friendly in concept, but your reference curation skill
matters a lot.
Speaker 2 (05:41):
Daily YouTubers, this is a w if your channel has
recurring characters or segments. For podcasters doing video promos, you
can lock your visual identity while you iterate scripts. Designers
could use it for animated lookbooks. Writers previz your scenes
before you commission art.
Speaker 1 (05:57):
One more detail from pr Newswire. They've touted first to
last frame transitions, seamless links between unrelated images or clips.
If that pairs with multi reference, you could bridge scenes
while holding identity. Think montage with consistent characters.
Speaker 2 (06:10):
That's cool for music videos and hype reels. Also, Shengshu
got named a twenty twenty five Technology Pioneer by the
World Economic Forum, also on PR Newswire. It doesn't mean
the model is best, but it signals they're serious about
creator facing controls, not just slick demos.
Speaker 1 (06:26):
Does this replace other tools? Not yet. I'd still keep
a compositor and a color tool in the stack, but
it could kick something like ad hoc frame painting to
the curb. For a lot of cases.
Speaker 2 (06:35):
Time savings guess for a thirty sixty second sequence built
from five second clips, maybe you cut your cleanup by
a third or more, depending on how chaotic your references
were before. If you're scrambling on deadlines, that's clutch risks.
Speaker 1 (06:48):
If you rely on seven references for every scene asset
preps sucks time pro tip build a reusable reference pack
per project model sheet, wardrobe sheet, hero prop set environment plates,
then drag and drop per sequence.
Speaker 2 (07:01):
And keep lighting continuity in mind. If your references are
Golden Hour and your prompt says noon office fluorescence, the
models like Girl pick a Vibe, match your reference lighting
to your intended scene lighting when you can.
Speaker 1 (07:13):
Where this fits the trend line, we're moving from single
shot wow to multi shot stories. Control and continuity are
the battleground. Vdoo's seven ref play is one clear move
in that direction.
Speaker 2 (07:23):
I'm into it. If you're comparing tools, try the same
scene across vd Q one, Runway and Peka with identical refs.
Measure identity drift and time to fix. Post your findings,
science quick source.
Speaker 1 (07:35):
Roll call before we wrap. The multi reference update and
availability are from PR Newswires vd Q one announcement. The
API and Creative Leap program details are in a separate
PR Newswire release, and the WEF twenty twenty five Technology
Pioneer nod is also PR Newswire.
Speaker 2 (07:50):
All recent meme send off seven refs, enter one consistent
timeline leaves mad max vibes. Okay, I'm done.
Speaker 1 (07:57):
That's the show. Thanks for hanging with us on blue
Light AI Daily. If you're testing vdoo Q one's multi
reference flow, let us know what holds up and what breaks.
Speaker 2 (08:05):
And hit bluelightningtv dot com for news updates and new
video tutorials on your favorite AI tools. We've got breakdowns
of the latest video model upgrades and how to put
them to work.
Speaker 1 (08:15):
Appreciate you all, Catch you next time.
Speaker 2 (08:17):
Bye,
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.