To the gods of sound knowledge and timecode drifting...

[paulcurtis]

So Epic-W and Zoom F4 (upgraded from a H4n recently so not used yet). When we've hired in sound for bigger jobs i place some blinkers on my head and don't pay too much attention. But i usually end up with a tentacle sync on my epic shoving TC in and the post side is smooth.

For my own stuff, hence the F4, i'm working out whether i should be doing similar.

But what i struggle to understand is why the inter webs are full of cameras drifting and boxes loosing frames - because from a techy view i just don't understand why anything *digital* should drift? So maybe the tech inside these things is something i don't understand. wondering if anyone can shed light on why all cameras aren't rock solid at timecode?

So in the case of F4 + EpicW, in theory i could start a session jamming it to the TC of the F4 via a BNC to Lemo cable. Then both devices are synced at that point. As i understand though this will drift in time? Or won't it? Is it an issue of recorders and cameras being turned on and off and loosing the sync. AFAIK the Epic W will maintain the TC when off? But perhaps the F4 doesn't.

Is that why people use tentacles and ultrasyncs?

However as i understand the tentacle, one sync'd, is basically a bunch of independent units whereas the ultrasync is broadcasting on an RF network to ensure all devices are the same? In this case why isn't the innards of a tentacle inside every camera?

I'm not against getting a couple of tentacles or ultrasyncs but i'd love to understand why in a single cam/single recorder environment.

thanks!
Paul

 

[Brian Trim]

Being a former sound mixer, I can tell you that any camera system will drift pretty quickly after a cable sync. This is mainly due, I believe to the way camera manufacturers deal with TC in their cameras. I don’t know the tech behind how they handle TC in camera other than that it’s a battery backed up TOD clock as you all already know. Even if you rejam the signal on a regular basis throughout the day, all the camera systems will drift to varying degrees from the master clock device used to jam said camera. The only way to avoid this is to jam a box like the tentacles etc and keep them permanently on the camera. That being said, it’s still a good idea to rejam both systems after lunch and any shutdowns of the camera

 

[paulcurtis]

Being a former sound mixer, I can tell you that any camera system will drift pretty quickly after a cable sync. This is mainly due, I believe to the way camera manufacturers deal with TC in their cameras. I don’t know the tech behind how they handle TC in camera other than that it’s a battery backed up TOD clock as you all already know. Even if you rejam the signal on a regular basis throughout the day, all the camera systems will drift to varying degrees from the master clock device used to jam said camera. The only way to avoid this is to jam a box like the tentacles etc and keep them permanently on the camera. That being said, it’s still a good idea to rejam both systems after lunch and any shutdowns of the camera

Thanks for replying Brian.

This is exactly what i read everywhere. But taking a step back - why? If tentacle can keep time in a $200 box why can't my $20k camera? That's the crux of my pondering question. Internally a camera is scanning a sensor at an incredibly precise rate otherwise we'd be getting dropped frames and odd cadences, surely? So as well as working out a practical solution for me, i'm just interested in understanding the why. Sometimes the general knowledge is out of date, perhaps modern cameras are fine. Don't you find it curious that cameras are drifting in this day and age (assuming they are!)

cheers
Paul

 

[Brian Trim]

Hey Paul,
I’ve always been curious as to why the manufacturers don’t develop more robust TC. I think it comes down to priorities. Let’s take RED for instance. I would guess that the bulk of their engineers time is spent developing sensor and color technology and things like audio and TC are left as the least important part of their universe. As important TC is to professional workflows, I’ve never felt the manufactures felt it had to be dealt with on a higher level then they have. It’s all of them too. I’ve been on shoots with massive Alexa setups and have spent all day dealing with TC error messages even though it had a box on it. And that’s a well documented issue on that camera. Perhaps that has been addressed by Arri but fact is, a lot of pro mixers I know lament the fact that TC can be an issue on a lot of cameras but again I’ve always felt as I just stated it comes down to priorities. Not enough post people are yelling at the manufacturers and saying “fix this”. They just blame the sound department when clips don’t sync.

 

[paulcurtis]

Hey Paul,
I’ve always been curious as to why the manufacturers don’t develop more robust TC. I think it comes down to priorities. Let’s take RED for instance. I would guess that the bulk of their engineers time is spent developing sensor and color technology and things like audio and TC are left as the least important part of their universe. As important TC is to professional workflows, I’ve never felt the manufactures felt it had to be dealt with on a higher level then they have. It’s all of them too. I’ve been on shoots with massive Alexa setups and have spent all day dealing with TC error messages even though it had a box on it. And that’s a well documented issue on that camera. Perhaps that has been addressed by Arri but fact is, a lot of pro mixers I know lament the fact that TC can be an issue on a lot of cameras but again I’ve always felt as I just stated it comes down to priorities. Not enough post people are yelling at the manufacturers and saying “fix this”. They just blame the sound department when clips don’t sync.

So after some digging about how they work and what a quartz crystal is then i guess that you're probably right - skimping on parts. However that still doesn't really make sense - because the cost difference doesn't appear to be that great. The TC side is probably separate than the sensor electronics, more an after thought.

Just seems such an industry anomaly!

So i guess i will need tentacle/utrasyncs running as well as the supposed built in TC generators. What a waste! Understand if it's a huge multi camera / device set up but other than that...

cheers
Paul

 

[Marc Wielage]

But what i struggle to understand is why the inter webs are full of cameras drifting and boxes loosing frames - because from a techy view i just don't understand why anything *digital* should drift? So maybe the tech inside these things is something i don't understand. wondering if anyone can shed light on why all cameras aren't rock solid at timecode?

Short answer: because the internal parts even expensive cameras use for audio and timecode are often shitty.

Longer answer: multiple cameras need to be synced in terms of both sync and timecode, so for that reason you need to have a Lock Box (Ambient, Denecke, Tentacle, etc.) that supplies both genlock reference sync and jammed code to everything. Ideally, you rejam to the master audio recorder every 4-5 hours, and the very stable, high-quality Temperature-Controlled Crystal Oscillator (TCXO) in the audio recorder will stay dead in sync all day long. I find that it helps to have a jammed TCXO timecode slate on set as well, just so that you can see immediately if the cameras are out or if the slate is somehow out from the audio deck.

The Lock Boxes nowadays are so small, and their batteries last so long, it's not that big a deal to velcro one to the side of each camera and go all day. When they're factory-aligned, I've had them last 9-10 hours and be exactly in sync, right on the subframe, with zero drift.

 

[paulcurtis]

Short answer: because the internal parts even expensive cameras use for audio and timecode are often shitty.

Longer answer: multiple cameras need to be synced in terms of both sync and timecode, so for that reason you need to have a Lock Box (Ambient, Denecke, Tentacle, etc.) that supplies both genlock reference sync and jammed code to everything. Ideally, you rejam to the master audio recorder every 4-5 hours, and the very stable, high-quality Temperature-Controlled Crystal Oscillator (TCXO) in the audio recorder will stay dead in sync all day long. I find that it helps to have a jammed TCXO timecode slate on set as well, just so that you can see immediately if the cameras are out or if the slate is somehow out from the audio deck.

The Lock Boxes nowadays are so small, and their batteries last so long, it's not that big a deal to velcro one to the side of each camera and go all day. When they're factory-aligned, I've had them last 9-10 hours and be exactly in sync, right on the subframe, with zero drift.

Thanks Marc,

How does a Red interpret TC in vs Sync in? There's one lemo on the vlock for 'Sync'. A tentacle cable is basically 3.5mm jack which is sending just TC AFAIK? An ultrasync can send both (it has two outputs). When not running two cameras in a 3D rig do we really need Sync? I believe this is about frame accuracy within the TC base of 1/25th as TC only goes down to the base frame rate but two cameras can see waver (be at different positions) within that 1/25th of a second.

Aside from the issue that these boxes are quite cheap so there's no real reason why the same TCXO can't be included in the internal design (lets not go there) i guess that idea of rock solid sync over longer than a day is a physical challenge. This is a tech issue that i wasn't really aware of so at least this has informed me a bit more.

So seems that i ought to head down external boxes.

So tentacle vs ultrasync. The ultrasyncs broadcast over a RF network but the tentacles are jammed then left for the day. I don't know whether one is more reliable than the other..?

cheers
Paul

 

[John Brawley]

Aaton's and Alexa's and some BMD cameras have been the only cameras that I've worked with that provide accurate TC over 8 hours.

Both purposely had temp controlled TC clocks fitted, I believe the Alexa's are made by Ambient at some expense.

When I asked my friends at BMD about clocks they tell me that there's a couple of things. One is that the system clock that runs the whole computer side inside the camera needs to be seperate from the TC recording clock for various engineering reasons. So you need two clocks (at least in their cameras)

And that accurate for more than 8 hour clocks are pretty hard to do consistently and the only way to get them is to pay. They're rated over a certain lifespan, so you can get an accurate clock, but it's only "good" for a couple of years, vs a much more expensive version that's got a 5 year lifespan.

I know for BMD ( who are making low 4 figure cameras) it's a hard choice about how much to invest in a clock internally when the standard workflow is an external clock anyway.

By the way, the BMD UMP and the new 4K are both "good" internal clocks, but I still use an external clock because that's the workflow for all the other cameras on set, including the Alexas. It's an extra backup, and oyu don't have to worry as much about keeping the camera powered constantly, nor remembering to re-jam the camera when you switch between offspeed and sync speed rates.

Having an external clock fitted that has days worth of batteries that will always run is a good backup to the power cycling and frame rate cycling of a camera.

JB

 

[paulcurtis]

Aaton's and Alexa's and some BMD cameras have been the only cameras that I've worked with that provide accurate TC over 8 hours

I've not used either of those for that length of time so this isn't a personal observation but since i've been googling my way to enlightenment i've read numerous places where the Alexas at least are off. Could be a per camera issue - or as you say length of service. That's an interesting nugget of info - that stability of these crystals can be an issue over time.

I'm in the acceptance phase now, so just need to decide between tentacle and ultrasync...

cheers
Paul

 

[Björn Benckert]

Synchronizing computers with each other is seriously difficult. Think thats one part of it. Cameras are basically computers these days.

 

[John Brawley]

I've not used either of those for that length of time so this isn't a personal observation but since i've been googling my way to enlightenment i've read numerous places where the Alexas at least are off. Could be a per camera issue - or as you say length of service. That's an interesting nugget of info - that stability of these crystals can be an issue over time.

I'm in the acceptance phase now, so just need to decide between tentacle and ultrasync...

cheers
Paul

Paul it depends which Alexa. The Mini doesn’t NOT have the same clock as the full size cameras and are notionally “not good” aka a system clock, not the seperate clock the full size bodies have.

JB

 

[paulcurtis]

Thanks everyone. I'm in a slightly more knowledge place now when it comes to sync/TC!

I guess it's turned into a Tentacle vs Ultrasync situation - does anyone have any opinion. I've used tentacles but one of the benefits (built in mic) doesn't really apply in red land.

cheers
Paul

 

[Bob Gundu]

Ive used the tentacles for many years now without issues. They're so small!

 

[David Peterson]

So Epic-W and Zoom F4 (upgraded from a H4n recently so not used yet). When we've hired in sound for bigger jobs i place some blinkers on my head and don't pay too much attention. But i usually end up with a tentacle sync on my epic shoving TC in and the post side is smooth.

For my own stuff, hence the F4, i'm working out whether i should be doing similar.

But what i struggle to understand is why the inter webs are full of cameras drifting and boxes loosing frames - because from a techy view i just don't understand why anything *digital* should drift? So maybe the tech inside these things is something i don't understand. wondering if anyone can shed light on why all cameras aren't rock solid at timecode?

Why?

There are three main reasons why:

1) "image 1st, sound last"

Unfortunately it seems most camera manufacturers just don't give a sh*t about sound department's needs, they just pour all their resources into making cameras which produce the prettiest pictures instead.

From random Chinese brands up to even ARRI themselves, they've all done things which make me as a soundie facepalm hard.

2) "human error"

Unfortunately half the human population has below average intelligence, and won't do their basic research beforehand or follow simple logical instructions step by step.

Doubt me? Just hang out on any filmmaking forum for a while....

3) "gremlins"

 

[David Peterson]

Thanks everyone. I'm in a slightly more knowledge place now when it comes to sync/TC!

I guess it's turned into a Tentacle vs Ultrasync situation - does anyone have any opinion. I've used tentacles but one of the benefits (built in mic) doesn't really apply in red land.

cheers
Paul

The original Tentacle Sync was a game changer for its small size and very low price, however now in 2019 the Tentacles have stiff competition when it comes to smallness (in size or price). In my personal opinion, the Ultrasync ONE is the best buy at the moment if you're sure you'll never need the scratch mic. (for a few reasons, such as: genlock, LED display, :blink, etc)

I did a short video here on my YT vlog channel, which compared three different timecode boxes:

And later on, I got the newer Tentacle Sync E as well:

 

[David Peterson]

This is exactly what i read everywhere. But taking a step back - why? If tentacle can keep time in a $200 box why can't my $20k camera? That's the crux of my pondering question. Internally a camera is scanning a sensor at an incredibly precise rate otherwise we'd be getting dropped frames and odd cadences, surely? So as well as working out a practical solution for me, i'm just interested in understanding the why. Sometimes the general knowledge is out of date, perhaps modern cameras are fine. Don't you find it curious that cameras are drifting in this day and age (assuming they are!)

There are varying levels of "precise", sure, it is doing it accurate enough not to drop frames. But from a timecode perspective, that is a very rough level of "precise", this is why we have genlock.

So after some digging about how they work and what a quartz crystal is then i guess that you're probably right - skimping on parts. However that still doesn't really make sense - because the cost difference doesn't appear to be that great. The TC side is probably separate than the sensor electronics, more an after thought.

Yes, because timecode and genlock are two different (but related) things.

Just seems such an industry anomaly!

Nah, not an anomaly, just the usual business of noone giving a sh*t about the sound department....

So i guess i will need tentacle/utrasyncs running as well as the supposed built in TC generators. What a waste! Understand if it's a huge multi camera / device set up but other than that...

What "supposed built in TC generators"? They don't exist! Nobody ever claims to have them in their cameras.

Hey Paul,
I’ve always been curious as to why the manufacturers don’t develop more robust TC. I think it comes down to priorities. Let’s take RED for instance. I would guess that the bulk of their engineers time is spent developing sensor and color technology and things like audio and TC are left as the least important part of their universe. As important TC is to professional workflows, I’ve never felt the manufactures felt it had to be dealt with on a higher level then they have. It’s all of them too. I’ve been on shoots with massive Alexa setups and have spent all day dealing with TC error messages even though it had a box on it. And that’s a well documented issue on that camera. Perhaps that has been addressed by Arri but fact is, a lot of pro mixers I know lament the fact that TC can be an issue on a lot of cameras but again I’ve always felt as I just stated it comes down to priorities. Not enough post people are yelling at the manufacturers and saying “fix this”. They just blame the sound department when clips don’t sync.

Yup, ARRI is "best" when it comes to dealing with timecode gracefully, but yet.... even ARRI isn't immune to their problems.

 

[paulcurtis]

David,

Thanks, i appreciate the details!

I've moved over to your camp about sound related stuff. I think the ultra syncs are perhaps best in my case for the reasons you say (plus they're a bit cheaper in fact)

I'm just putting together a small sound bag for my own stuff around the F4.

thanks again
Paul