precise synchronization via USB port, for stereoscopic 3D

[kaliberswiss]

Dear GoPro Team,

First of all, thank you for creating the best and most versatile action camera on the market!

As an immersive video producer, I have the opportunity to collaborate with leading companies like Meta and Red Bull. Currently, my goal is to develop a reliable camera setup that allows for precise stereoscopic 3D recordings. To achieve this, I am using two GoPro cameras set to 5.3K at 30fps to capture the highest possible quality.

However, I’ve encountered a challenge when trying to accurately synchronize the two GoPros for fast motion capture. Even when using the "Precise Time QR Code" feature, there is a noticeable shift during fast movements. My assumption is that while the timecode is being synced, the GoPro sensor clocks themselves are not being reset.

From my understanding, achieving precise sensor clock synchronization via the QR code isn’t currently possible. However, I came across the discontinued “Dual Hero System,” which allowed two GoPro Hero 3+ cameras to be synchronized using a cable connection. Unfortunately, this feature is no longer available with the newer versions.

This brings me to my question: Would it be possible to develop a GoPro Labs firmware that enables precise sensor clock synchronization between two GoPros via the USB port? Even if an intermediary device is required, this would be an invaluable feature.

If this is feasible, how could I contribute to the development of such a product?

This feature would give GoPro a significant advantage in the immersive media industry, as "Genlock synchronization" is currently a feature reserved for professional cinema cameras.

Thank you for your time and consideration. I look forward to your response.

Best regards,
Jussuf Lieberherr

[dnewman-gpsw]

Yes it is interesting. Thinking on it.

[3diver]

David, what is the best way for those interested in this feature request to know whether it is advancing?
Thank you very much for thinking on it! It would be so great for multi cam and 3-D stereo users. There are more 3-D stereo enthusiasts today than there were for the original hero through hero 3+ black generations.

[dnewman-gpsw]

No news to report. I will not post on any interim progress, as that would be no guarantee of any release. This is not a promised feature, but an area of interest.

[vedlai]

Hopefully there will be progress with this, GoPro has worked with Apple on Wide FOV APMP and Wide FOV Stereo from 2 GoPro 13s is incredible.

Just not perfect since even with 120 FPS it is still a tiny bit off no matter what I do.

[vedlai]

@dnewman-gpsw If you have access to an Apple Vision Pro on visionOS 26 go to this website on Safari and expand to full screen. It will just look like a regular 2D video in any other browser/version.

It is an example of APMP Wide FOV Stereo.

Just a glimpse into the potential of this format, I am not a videographer just an enthusiast.

[dnewman-gpsw]

I am do have an AVP. I will try it out, once I get it updated.

[RikshaDriver]

This sounds like a good use case for implementing PTP

[kaliberswiss]

Hi Davit, I have a quick question , why has the 3d hero system been discontinued and does gopro have any interests in VR?

I really do think that, with new VR headsets like the Vision Pro, there's a lot of potential for immersive stereoscopic 3D content.

Would love to know your thougts on it.

[dnewman-gpsw]

Discontinued due to insufficient demand, 3D was hot for a while, then went dormant. This happened years before VR's temporary resurgence (still well below market expectations.) However camera sync has many use cases well beyond stereoscopic imaging, the whole market for sync is more interesting. Camera sync of HERO HD/2/3/3+/4 was dropped due to waterproofing, not use to lack of interest, we just could put connectors on the outside of the camera. With the new Contacto door, wired camera connections and waterproofing finally coexist. Here is some HERO HD Sync from 2010-11 https://www.youtube.com/watch?v=d0x52u2yzgI the BTS https://www.youtube.com/watch?v=jmGNwqKH2Yk (I was tangentially involved.) So yes, I personally want sync again.

[kaliberswiss]

Ok i understand now, thanks for your fast response.

So do you think it could be possible to have multiple Gopros synchronized with the contacto doors? I could also imagine connecting them once , when the gopro is powered to sync them.

Just need a way to sync them, that would be the best thing ever!

[dnewman-gpsw]

It is a potential mechanism, and a clean setup, but a lot more to learn.

[kaliberswiss]

I see you 're working on it.

I would love to give you a Meta Quest VR Headset, so you would also see how revolutionary immersive 3D with the Gopro looks!

If it's possible, tell me where I can send it to. [email protected]

[3diver]

dnewman, has the development team decided whether to pursue a genlock feature or equivalent?

As a 3D photographer, I also support the recommendation for a triggering method, both for 3D video and for 3D stills photography. My experience with many SLR cameras has been that Bluetooth wireless synchronization is never accurate enough for shutter speeds enabling 3-D stills photography with flash.

[uhdtv3d]

It would be great if a firmware update included a wireless sync method, eliminating the need for a USB cable. This would be especially useful when the two GoPros are spaced more than a meter apart to shoot hyper-stereo.

I am using two GoPro Hero 10s with identical settings for stereoscopic 3D stills and video recording. To start and stop recording, I use GoPro's 'The Remote'. However, I’ve noticed they don’t sync perfectly; there can be a 1- to 4-frame difference between them when recording at 25 fps, for example. This discrepancy makes it difficult to capture still photos of moving subjects effectively. I have to check whether burst mode would provide a few synced frames!

To synchronize the left and right videos, I import them into VideoRedo TV Suite, review the video thoroughly to identify scenes with fast motion to note the frame number of the two where they are exactly matching, and adjust the clips by trimming frames from the start and/or end as needed to match the total frame count. I then export the synchronized videos without re-encoding and merge them in Stereo Video Maker. This manual method takes a lot of time and sometimes it is difficult to understand the exact frame where they are matching just by eye inspection!

[rkanaga]

Yes, I would love this too.

I bought one of the older GoPro 3D Hero cases with the cable on eBay hoping to use with some Hero 4 Blacks, but then discovered that I needed to track down a couple of ancient GoPro Hero 2 cameras to make it work! I even tracked down a copy of the discontinued GoPro Cineform software and got it to work on an ancient copy of Mac OSX Mavericks from 2013 (pre 64 bit) running in a partition on my trashcan Mac Pro.

There is huge potential in stereo video with the advent of Meta 3, Pico 4 and of course Apple Vision Pro, so some way of synchronising modern GoPro cameras via USB or even bluetooth would be awesome, as well as maybe a mount?

360 video on a headset is great, but stereo brings true depth which is even more immersive sometimes.

[ByteAccess]

I would absolutely like to see this feature implemented!

I have a project that requires stereo VR180 with a smaller IPD (2x Hero11 + Max Lens 2.0 running in 8:7 @50/60fps). It will be imperative to providing learners/viewers with a closer look at objects demonstrated in the video. These shots will be intercut with Canon VR180 footage. It's not a perfect solution as the synchronisation disparity for each eye may be as high as 8-10ms, but it's fine for shots with slow or limited motion and will maintain the stereo VR180 format.

Another project that I'm currently exploring is sparse array light field capture (7x Hero11 cameras running at 100/120 fps). The cameras are controlled with a PC over USB. Synchronised cameras will improve the results and enable a plethora of new and exciting applications.

The synchronisation doesn't need to be perfect, if each GoPro sensor can be synced to packets periodically sent over USB by a device such as a laptop or Raspberry Pi and it achieves an error within 1ms over an array of cameras, this would still be a fantastic result!
(I used the Hero3 3D system back in the day)

[3diver]

I concur, except that I believe the synchronization needs to be at least good enough for video which, absent a global sensor, may mean line by line synchronous. I don’t know if you want GoPro dual system was achieving ing that level of sync, but it did appear adequate for a video.

[msu-gopro]

I would highly appreciate any type of proper genlock sync like the old hero 3D system.
Everyone who viewed my 3d scuba diving videos or variable interocular distance landscape videos on a 100” projection screen wanted to buy the system - is was already out of production at that time unfortunately. There is now replacement that comes close, especially because of the usability under water.

To all GoPro Devs who read this: Please please please with cherry on the top: return genlock sync via cable.

[BobVe]

Earlier this year I entered 2 short 3D movies (shot with my GoPro 11s) to the stereo theatre at the 3D Con in Twin Cities. Sync was dependent on the fastest frame rate available (240 fps). My 2 short videos tied for first place out of all the presentations in the theater! They were shown again at the International Congress in the Netherlands, and now they’ve generated a bit of interest.

One respondent from Germany inquired about my methods, and had some thoughts. I had told him that I start my cameras using the remote, and align frames in post. He responded:

“Wireless (Bluetooth) may not be the best choice for our purpose. There are two reasons: Bluetooth has only a one-to-one protocol. If two cameras react, you have luck, but that cannot be guaranteed. The 2nd point is more important: For Bluetooth remote, the cameras must have been already switched on, but that is, what we need for both cameras at the same time. Let me go more into detail: When the cameras are switched on, the internal oscillators and PLLs start and are continuously running. When the command comes "start the video", the camera will wait, until the pixel counter is at pixel zero, then the video starts. It doesn't matter at all, at which time the command "start the video" will come, the camera always waits until the pixel counter is at pixel zero, and that is only dependent on the switch on time of the camera. So you need an electric contact to switch the camera on. All Sony Multiport cameras has this contact, and that is, what I have asked for the GoPro.”

I know that another videographer from Japan had once achieved excellent synch with a certain Panasonic model camera, and vital to his setup was a common external battery with a switch. Both cameras were connected switched on, but only when the battery switched on did they receive power.

My question is, would something like this work with the GoPros? And if so, is there a way to implement it?

[3diver]

Bob, congratulations on your win at 3-D con! I wish I’d been there to see your entries.
As for the sync technique, I utilize something similar
on one of my Sony camera twin rigs, albeit not for the purpose of video.
I insert dummy batteries into the cameras and wire them to a single power source controlled by a switch. I put the cameras in “on” mode and then turn the power on and off with the switch.
I see no reason why this couldn’t be done with GoPro cameras provided that one can find dummy batteries or make them. I haven’t searched to see if that’s an existing product.
The other important thing to confirm would be that the two GoPro cameras retain their settings when turned on and off by this method. You could try this, simply by popping a battery in and out of your camera.
I am really hoping that Labs will offer a synchronization option in the future.

Congratulations again,
Mark

[BobVe]

Thank you, Mark! It was unexpected, to say the least.
regarding the common battery, I think the first question is whether the GoPros can be left “on” when they have no power. The on switch on the GoPro is what I think of as a soft toggle—not a hard on/off switch. If you remove a battery from the GoPro while it’s on, then reinsert the battery, it’s off until you turn it on again.
I’m confident the settings would remain the same.
I did another test recently of my cameras by running them each for 5 minutes at 240 fps and flashing a strobe at the start of the recording and again at the end. So near 72,000 frames. The difference between the two was by about one frame.
Bob

[vedlai]

Earlier this year I entered 2 short 3D movies (shot with my GoPro 11s) to the stereo theatre at the 3D Con in Twin Cities. Sync was dependent on the fastest frame rate available (240 fps). My 2 short videos tied for first place out of all the presentations in the theater! They were shown again at the International Congress in the Netherlands, and now they’ve generated a bit of interest.

One respondent from Germany inquired about my methods, and had some thoughts. I had told him that I start my cameras using the remote, and align frames in post. He responded:

“Wireless (Bluetooth) may not be the best choice for our purpose. There are two reasons: Bluetooth has only a one-to-one protocol. If two cameras react, you have luck, but that cannot be guaranteed. The 2nd point is more important: For Bluetooth remote, the cameras must have been already switched on, but that is, what we need for both cameras at the same time. Let me go more into detail: When the cameras are switched on, the internal oscillators and PLLs start and are continuously running. When the command comes "start the video", the camera will wait, until the pixel counter is at pixel zero, then the video starts. It doesn't matter at all, at which time the command "start the video" will come, the camera always waits until the pixel counter is at pixel zero, and that is only dependent on the switch on time of the camera. So you need an electric contact to switch the camera on. All Sony Multiport cameras has this contact, and that is, what I have asked for the GoPro.”

I know that another videographer from Japan had once achieved excellent synch with a certain Panasonic model camera, and vital to his setup was a common external battery with a switch. Both cameras were connected switched on, but only when the battery switched on did they receive power.

My question is, would something like this work with the GoPros? And if so, is there a way to implement it?

The new Wide FOV Stereo APMP settings would let you view this in 3D straight from Safari on Apple Vision Pro.

I don't think anyone has published a Wide FOV Stereo video. I have a few amateur ones but nothing of the same caliber as yours.

[BobVe]

I need to make a correction to what I posted earlier regarding my synch. The clips were limited to 5 minutes and strobes were fired about 4 minutes 27 seconds apart. And the difference between the left and right clips was about 3 frames.

[3diver]

I strongly support this feature for 3D and multicam use! Is there a way to track whether and when a request gets put into a work pipeline?

[3diver]

David,Thanks very much for your response. I’ll keep an eye on the Labs forum to see if this feature may eventually be implemented and released.Kind regards,Markwww.undersea3d.comSent from my iPhone. Please excuse any typos.On Feb 24, 2025, at 5:35 PM, David Newman ***@***.***> wrote: No news to report. I will not post on any interim progress, as that would be no guarantee of any release. This is not a promised feature, but an area of interest. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[BobVe]

That’s good to know! One thing I wanted to test for is drift over time, so I was going to run the cams for multiple minutes at 240 fps. At 2.7k the files would be quite large. But you think that would give me better test results than shooting 1080?To clarify, I eventually want to shoot short clips at the highest possible res for my movies but I thought long tests for drift would make more sense at a lower resolution. Bob VeneziaSeattle, WashingtonOn May 24, 2025, at 10:29 AM, David Newman ***@***.***> wrote: They are not hurting, correct. Setting via QR Code are same a settings via the touchscreen, so setting are preserved the same way. For 3D, you will want stabilization off, as there are no guarantees each camera will select the same stabilization paths. You will want the highest frame rate that supports your needed resolution, e.g. 5K60, 4K120, or 2.7K240. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: ***@***.***>

[dnewman-gpsw]

Drift would be the same at all resolutions, as the drift would due to the small difference in the hardware crystal's timing in each camera.

[chungnm1]

I would love to see gen lock available via GoPro Labs on supported cameras.
I enjoy shooting 3D with a GoPro 7 Hero Black and while I can get acceptable results in video with locked shutter speeds, it would be awesome if a pair of newer GoPros (9 and above) can all be synced and also start and stop simultaneously via the labs firmware (Better than using the GoPro remotes).
I have asked GoPro to make a stereoscopic 3D GoPro... but gen lock would be a good feature if they are not interested in building such a device. (Just refashion the GoPro Max by flipping and folding one of the lenses around into VR180 configuration!)

[3diver]

Yes to genlock!

[furtovin]

For now, the only option left is to use an external timecode generator, which is constantly connected via an audio adapter.

[chungnm1]

Atmos Ultrasync Blue transmits a timecode via Bluetooth. Adding GoPro support to it would be a good start.
(On a side note.. I’m looking at my old GoPro Fusion which records separate left and right tracks to a pair of SD cards. Hacking the lens modules would make a interesting 3D camera)

[furtovin]

Yes, if you disassemble the camera (GoPro Fusion 360) and turn the lenses into one plane, you will get a normal VR 180 3D. And there are no problems with synchronization. I have already done this with Samsung Gear 360 (in the form of a ball. model 2016. they are cheap now ~20$), it shoots video poorly, but VR photos are quite good (3800x3800 for each eye).
The only problem: it's not easy to deploy cameras on a GoPro Fusion - the cables are very short. And you'll have to come up with a new housing and print it on a 3D printer.

[3diver]

Thank you for the very exciting news David! You are very kind to notify us of potential progress and it’s much appreciated. 😁Any additional detail would be welcome of course. For example, is the work with Apple on APMP using GoPro 13’s achieving its synchronization via GoPro Labs or by some other method at this time?It sounds like the sync isn’t tri-level genlock at this stage (which would be perfection!), but how close is it?What are your thoughts on the level of sync that might realistically be achievable if progress continues?Many thanks again for keeping us in the loop.MarkSent from my iPhone. Please excuse any typos.On Aug 13, 2025, at 11:17 AM, Vedlai ***@***.***> wrote: Hopefully there will be progress with this, GoPro has worked with Apple on Wide FOV APMP and Wide FOV Stereo from 2 GoPro 13s is incredible. Just not perfect since even with 120 FPS it is still a tiny bit off no matter what I do. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[vedlai]

APMP/Wide FOV is a presentation standard not synchronization. Basically GoPro videos provide the Apple Vision Pro lens information for presentation. Synchronization is not done by Apple or GoPro I use standard methods and inject APMP information (my own app). It is a unique setup that wasn't even working until a bug report I submitted was resolved on Tuesday.

The video is decent but apparent in shimmering water or rapid movements such as birds flying.

[3diver]

Thank you for clarifying David. If I understand you correctly, now that your app allows proper presentation in APMP/wide FOV, you are hopeful that fact might provide more incentive for GoPro Labs to work on synchronization. Is that right?MarkSent from my iPhone. Please excuse any typos.On Aug 13, 2025, at 3:03 PM, Vedlai ***@***.***> wrote: APMP/Wide FOV is a presentation standard not synchronization. Basically GoPro videos provide the Apple Vision Pro lens information for presentation. Synchronization is not done by Apple or GoPro I use standard methods and inject APMP information (my own app). It is a unique setup that wasn't even working until a bug report I submitted was resolved on Tuesday. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[BobVe]

3diver, are you confusing Vedlai with David Newman? I’m not aware of any new developments on sync via GoPro labs. Or maybe you know Vedlai separately, and know his name?

[3diver]

Hello David,I was responding to your emails by replying to all. My apologies if I have created some confusion. I don’t know Vedlai.MarkSent from my iPhone. Please excuse any typos.On Aug 14, 2025, at 11:24 AM, BobVe ***@***.***> wrote: 3diver, are you confusing Vedlai with David Newman? I’m not aware of any new developments on sync via GoPro labs. Or maybe you know Vedlai separately, and know his name? —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[3diver]

Sorry folks, I have been replying to emails in Outlook instead of going to GitHub. Ain Outlook, all the emails appear to come from David but I can see now that they are not. Again, my apologies. MarkSent from my iPhone. Please excuse any typos.On Aug 14, 2025, at 11:24 AM, BobVe ***@***.***> wrote: 3diver, are you confusing Vedlai with David Newman? I’m not aware of any new developments on sync via GoPro labs. Or maybe you know Vedlai separately, and know his name? —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[3diver]

Hi Bob,Of course you’re absolutely right. I wasn’t thinking about the soft “on” toggle switch  on the GoPro cameras.I wonder if it is possible to hack into that switch and wire it to an outside the body hard on/off switch. I know you can hack the shutter release switches on Sony A7 series cameras this way, but they might have entirely different architecture.MarkSent from my iPhone. Please excuse any typos.On Sep 7, 2025, at 9:18 AM, BobVe ***@***.***> wrote: Thank you, Mark! It was unexpected, to say the least. regarding the common battery, I think the first question is whether the GoPros can be left “on” when they have no power. The on switch on the GoPro is what I think of as a soft toggle—not a hard on/off switch. If you remove a battery from the GoPro while it’s on, then reinsert the battery, it’s off until you turn it on again. I’m confident the settings would remain the same. I did another test recently of my cameras by running them each for 5 minutes at 240 fps and flashing a strobe at the start of the recording and again at the end. So near 72,000 frames. The difference between the two was by about one frame. Bob —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[BobVe]

Mark, I have an electronics wizard friend who would probably be able to do that if it’s possible/worth the fuss. But I think there’s maybe a larger question of whether the cameras have a built in speed that’s slightly different and that they will drift over time. Ideally I would be able to test multiple cameras and see which 2 show the least amount of difference in drift. I may have gotten unlucky with my current pair. The longer the clip, the more drift between these two.

[3diver]

Bob,I realize you’re talking about your test video here and not your award-winning videos, but in practice, how do you feel that a frame drift of three frames affects the stereoscopic viewing of the clips? Did they have moving objects in them and could you notice temporal displacement?I’ve shot the hero black 3+ dual stereo rig and it is frame accurate. Unfortunately, the image quality is low, particularly in very low light where I tend to shoot. Do you have a workflow you like for improving IQ?Best,MarkSent from my iPhone. Please excuse any typos.On Sep 7, 2025, at 3:39 PM, BobVe ***@***.***> wrote: I need to make a correction to what I posted earlier regarding my synch. The clips were limited to 5 minutes and strobes were fired about 4 minutes 27 seconds apart. And the difference between the left and right clips was about 3 frames. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[BobVe]

Mark, I’m talking about both the test videos and at least one of the videos that won an award.
The video of the hummingbirds (Tiny Dancer) is barely over a minute, including title and credits. The first image is at frame 138. Synch looks great. Synch stays mostly okay through around frame 1200. After that, for around the next 400 frames, there are very few pairs I would enter into still exhibitions. There’s rivalry. Anna’s hummingbirds wings beat at around 80 beats per second , so it’s a tough test. And the frame pairs are all on the same wing beats. But for the last 400 frames or so, the wings might have a slightly different shape, or one wing might be invisible on one side, one clearly defined and the other blurred.
Obviously it’s bugging me more than it bugged the judges!

This is a frame pair from the beginning, with good synch:

And these are random frame pairs from later in the movie:

Bob

[3diver]

I understand Bob and was simply responding to your question about the triggering. We need at least frame accurate synchronization. Apple has now beat GoPro to the punch on that one, at least since the demise of the hero 3+ black dual system.Sent from my iPhone. Please excuse any typos.On Sep 11, 2025, at 10:19 AM, BobVe ***@***.***> wrote: Mark, I have an electronics wizard friend who would probably be able to do that if it’s possible/worth the fuss. But I think there’s maybe a larger question of whether the cameras have a built in speed that’s slightly different and that they will drift over time. Ideally I would be able to test multiple cameras and see which 2 show the least amount of difference in drift. I may have gotten unlucky with my current pair. The longer the clip, the more drift between these two. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>