ATEM Mini series latency spec

Questions about ATEM Switchers, Camera Converter and everything live!
  • Author
  • Message
Offline

jbreher

  • Posts: 3
  • Joined: Wed Mar 17, 2021 8:52 pm
  • Real Name: Joe Breher

ATEM Mini series latency spec

PostWed Mar 17, 2021 9:00 pm

I'm a live streaming musician. I was recently slapped with a dose of reality in 'upgrading' my video source from webcam to a mid-grade digital mirrorless camera through a (crappy?) HDMI/USB converter device.

While subjective picture quality went way up, the video ingest latency has skyrocketed from ~50 ms to ~1s. I'm a one-person crew -- production and talent -- and this delay in my 'confidence monitoring' has knocked my performance off my game.

I rather suspect that the HDMI/USB convertor -- not to be named here, but under $100 -- is the likely source of the huge jump in latency. I am accordingly looking for a better link in this chain.

Am I just blind, or is there no published spec for the latency of video conversion from HDMI in to USB out for the ATEM Mini family?
Offline

jbreher

  • Posts: 3
  • Joined: Wed Mar 17, 2021 8:52 pm
  • Real Name: Joe Breher

Re: ATEM Mini series latency spec

PostWed Apr 14, 2021 6:33 am

Nobody?
Offline

Tim_Partridge

  • Posts: 88
  • Joined: Sun Sep 06, 2020 7:40 pm
  • Real Name: Tim Partridge

Re: ATEM Mini series latency spec

PostSat Apr 17, 2021 11:35 pm

If I feed the oiutput of my laptop into the ATEM and view the program output in a window I can see the infinite regress building. So it is not instant, but from memory it is under a second per stage. I'm away from my hardware right now, so can't check.

If you connect a monitor to the HDMI out of the ATEM that might be an alternative. Be aware that most inexpensive monitors only work at 60 Hz, so the ATEM would have to be set to that which may impact streaming bandwidth needs. There is the option to directly send input 1 to the output with minimal latency instrasd of program, but this has to be enabled using the software control. (May also need Pro model or above, I'm not sure.) Pro and above lets you have multiview which is for monitoring inputs program and preview. (You can't do anything more exiting than one picture in picture for program output unless you go for an extreme model - those have super source.)

Tim
Offline

Jim Simon

  • Posts: 17109
  • Joined: Fri Dec 23, 2016 1:47 am

Re: ATEM Mini series latency spec

PostSun Apr 18, 2021 3:44 pm

What's the convertor for, Joe? Normally you'd just go HDMI out of the camera into the Mini Pro, no conversion necessary.
Offline

jbreher

  • Posts: 3
  • Joined: Wed Mar 17, 2021 8:52 pm
  • Real Name: Joe Breher

Re: ATEM Mini series latency spec

PostSun Apr 18, 2021 4:20 pm

Jim Simon wrote:What's the convertor for, Joe? Normally you'd just go HDMI out of the camera into the Mini Pro, no conversion necessary.


To be clear, I do not currently own or employ an ATEM. I am trying to determine if replacing the converter with a newly-purchased ATEM might solve my latency issue.

The converter ingests the HDMI which the camera spits out and converts it to UVC over USB so the computer can use it. Which, if I am not mistaken, is one of the many functions that the ATEM also performs.

Before switching to a 'better' system of a mirrorless camera with the converter, latency was virtually undetectable - about 50ms seemed to line up audio and video. Now it is close to 1 second. The faster solution -- albeit with lesser picture quality -- was a Logitech Brio webcam.
Offline

hatschiii

  • Posts: 3
  • Joined: Wed Apr 21, 2021 4:11 pm
  • Real Name: Sven Weber

Re: ATEM Mini series latency spec

PostWed Apr 21, 2021 4:22 pm

I use the ATEM Mini for Zoom and Teams. Now it comes to the setup what is better regarding latency

1. Microphone into DSLR Cam and then into the ATEM Mini
2. Microphone directly into the ATEM Mini and the DSLR Cam seperate into the ATEM Mini

Most save is for sure the first option, but this means I need the DSLR allways on even if I want to use the microphone for other things in my MacBook Pro.

I saw a video on Youtube where someone said NEVER put the microphone direclty into the ATEM Mini (can't remember where this was)

Best would be option 2 for me if...
a) I can handle the latency
b) I can use the microphone with the audio settings of the ATEM Mini (equalizer etc.) in the MacBook Pro without the DSLR

What would you recomment?
Offline

hatschiii

  • Posts: 3
  • Joined: Wed Apr 21, 2021 4:11 pm
  • Real Name: Sven Weber

Re: ATEM Mini series latency spec

PostFri Apr 30, 2021 9:46 am

I'm pretty new to all this video / audi stuff... and there are many things confusing me. hope someone can help me. Maybe I don't see the forrest because of all this trees at the moment.

When I connect my USB audio interface to the computer, the microfon signal gets into the computer and all the other audio from the computer gets back to the audio interface and from there to the loud speaker.

When I now connect the ATEM Mini to the computer the attached cams are send to the computer. Easy.

...but how about if I attach the microfon signal additional to the DSLR directly? In this case I will get the microfon signal through the computer back in the audio interface and from ther tho the loudspeakers. Any chance to "mix minus" the audio signal which goes Cam > Atem Mini > Zoom or Teams, so that it will not be used be the rest of the computer (have both, Mac and Win)?

I need to connect the microphone to the DSLR to have less latency. If I would use the microfon through the computer for Zoom / Teams I wouldn't have an issue.
Offline

Frank Engel

  • Posts: 671
  • Joined: Thu May 17, 2018 11:09 pm
  • Real Name: Frank Engel

Re: ATEM Mini series latency spec

PostSun May 02, 2021 11:41 am

jbreher wrote:
Jim Simon wrote:Before switching to a 'better' system of a mirrorless camera with the converter, latency was virtually undetectable - about 50ms seemed to line up audio and video. Now it is close to 1 second. The faster solution -- albeit with lesser picture quality -- was a Logitech Brio webcam.


I don't have a number for you, but just to point it out, a direct answer to your question is not sufficient to address your concern.

A webcam has latency between the camera and the computer, via the conversion into the USB format.

Most mirrorless cameras have latency between the camera and the HDMI port, and the converters will add more latency on top of that between the HDMI input and the USB output.

Without knowing how much of your latency issue is in the converter and how much is in the camera, knowing the latency of the ATEM itself would not be sufficient to determine whether or not it would be an improvement.

You could try to estimate this by connecting the HDMI output of the camera up to a TV to see what the latency is like?

Note that you could practically zero the latency of the ATEM itself by monitoring on a separate display via the HDMI output of the ATEM, which would normally have less than one frame of latency (maybe a frame or two if format conversion is going on for some reason), but this would not eliminate the latency of the camera itself, so if the camera latency is too high going out to a TV, the only thing you can do to deal with that is to replace the camera.
Offline

rstreber

  • Posts: 68
  • Joined: Tue Dec 08, 2020 4:55 pm
  • Real Name: Ryan Streber

Re: ATEM Mini series latency spec

PostSun May 02, 2021 6:02 pm

I've been relying on small Marshall SDI cameras in my studio for a number of years now to do very-low-latency closed-circuit HD video connections between multiple rooms for musicians tracking together. I can tell you that the ONLY way that I've found to keep latency sufficiently slow for comfortable live music-making is to go point to point between SDI cameras like these and monitors with SDI inputs and minimal hardware in between. I can get away with basic distribution boxes, but any multi viewers or other items in the chain that do any processing increase the latency to an unacceptable degree. If I take a basic DSLR or decent-looking camera with an HDMI output, the latency is inherently too high straight off the camera regardless of anything else that's going on.

So, in short, I'm guessing that the ATEM is not going to solve your problem.

All that being said, I can tell you that when I'm recording audio on a Pro Tools HDX rig at 48K and sync'ng it in real time to my ATEM connected to BMPCC4Ks, at I have to delay the send from Pro Tools to the ATEM about 6300 samples to get it roughly sync'd to picture. That would be on top of whatever the total system delay of my pro tools rig and dante network are, although I'm pretty sure that's all at most around 1ms.
Ryan Streber
Oktaven Audio
www.oktavenaudio.com
Offline

davincithe

  • Posts: 5
  • Joined: Wed Apr 22, 2020 9:43 am
  • Real Name: Roger Krebs

Re: ATEM Mini series latency spec

PostTue May 11, 2021 8:07 am

@rstreber Hi Ryan,
thank you for your great explanation.
My goal is to get 3 BMD Pocket Cameras in sync with the ATEM extreme iso and sound coming from a RME UFX+.
Will this work to have either the streaming in sync and also 4K recording on each cameras SD or USB SSD and also the iso files on the ATEM for post?
Are the BMD Pockets optimized to less delay in combination with the ATEM or is this only a wish or illusion?
I am in the same situation as the threat opener an looking for a good solution.
Offline

rstreber

  • Posts: 68
  • Joined: Tue Dec 08, 2020 4:55 pm
  • Real Name: Ryan Streber

Re: ATEM Mini series latency spec

PostTue May 11, 2021 1:02 pm

Hi Roger,
I'm not totally sure I understand what you're asking, but if you're just worried about sync, then that shouldn't be a problem at all between the BM cameras and ATEM. The system takes care of sync and you shouldn't have to do anything special either for recording or in post (especially if you're editing in resolve) to sync them up. Your audio from the UFX will likely need to be delayed by some amount to properly sync it with both the live audio / video feed off the ATEM as well as in post. The amount of delay will totally depend on your setup, so you'll just have to figure out what works best. But (if I'm understanding you correctly) that should be the only part of the chain that needs to be delayed for good sync.

Any delay that you apply to the RME output should also only apply to live viewing and streaming. I record a scratch stereo feed from my console (with appropriate delay) to the input on the ATEM, and then just sync up the "real" audio (recorded to my main Pro Tools rig) in post using the scratch audio as reference.

I don't believe that there's any special optimization for the Pocket cameras when working specifically with the ATEM. The only relevant detail there is that inputs 1 and 2 on the ATEM are capable of lower latency monitoring when viewed directly via HDMI out 1 / 2, but that would be a separate scenario from the program out which will have a uniform latency regardless of (but in addition to the self-latency of) whatever cameras are being input via HDMI.

Anyone else here, please correct me if I'm wrong!
Ryan Streber
Oktaven Audio
www.oktavenaudio.com
Offline

Frank Engel

  • Posts: 671
  • Joined: Thu May 17, 2018 11:09 pm
  • Real Name: Frank Engel

Re: ATEM Mini series latency spec

PostTue May 11, 2021 7:01 pm

rstreber wrote:Anyone else here, please correct me if I'm wrong!


Well... sort of wrong.

Someone who is live streaming doesn't really have a chance to edit in Resolve before it is too late (already streamed) so the sync between the cameras being translated into Resolve is certainly nice, but irrelevant.

The notion of both inputs 1 and 2 having the low-latency mode, the presence of HDMI out 2, and the sync between cameras in the Resolve project seems to assume the Extreme ISO model which is by no means a given from the conversation at large.

The concern seems to be monitoring latency - for example, if the musician is playing guitar live while streaming, he is hitting the strings then after some delay hears the sound or sees his hand move on the monitor.

That would be VERY difficult to play to.
Offline

rstreber

  • Posts: 68
  • Joined: Tue Dec 08, 2020 4:55 pm
  • Real Name: Ryan Streber

Re: ATEM Mini series latency spec

PostTue May 11, 2021 7:22 pm

Right. I was responding specifically to the previous post from davincithe. He seemed to be talking about something very different from the OP who I already responded to in an earlier post, saying basically that there's not really a good way to monitor one's self via program out of an ATEM with sufficiently-low latency for real-time musical performance. In my recording studio context, I've only been able to get truly low-latency video interconnections using Marshall cameras with SDI outputs feeding directly into monitors with SDI inputs.
Ryan Streber
Oktaven Audio
www.oktavenaudio.com
Offline

davincithe

  • Posts: 5
  • Joined: Wed Apr 22, 2020 9:43 am
  • Real Name: Roger Krebs

Re: ATEM Mini series latency spec

PostFri May 14, 2021 8:29 am

rstreber wrote:In my recording studio context, I've only been able to get truly low-latency video interconnections using Marshall cameras with SDI outputs feeding directly into monitors with SDI inputs.


Thank you so much Ryan and Frank for your answers.

Which Marshall Cameras did you used?
I ordered a BMD Pocket 6K Pro to try out the on or other thing.
I am a little bit surprised that BMD in conjunction with the ATEM
mini’s is not able to coordinate the Cameras to the ATEM without a delay.
They know this Product best and they should bring them together without
delay etc... or I am wrong?
ATEM stands for live streaming and it should work with live music to?
Offline

Frank Engel

  • Posts: 671
  • Joined: Thu May 17, 2018 11:09 pm
  • Real Name: Frank Engel

Re: ATEM Mini series latency spec

PostFri May 14, 2021 8:59 am

rstreber wrote:Right. I was responding specifically to the previous post from davincithe. He seemed to be talking about something very different from the OP


Thanks, somehow I missed that post in between.

davincithe wrote:I am a little bit surprised that BMD in conjunction with the ATEM
mini’s is not able to coordinate the Cameras to the ATEM without a delay.


The first issue is that most cameras have a delay internally when sending output to HDMI (cameras sending SDI out generally do better... most of them anyway). If your audio is coming through the camera then it will be delayed to match the video and there is not a problem with sync at that stage. The first problem being discussed here is that you are sending audio independently of the camera. There needs to be a delay to get the audio from the mixer in sync with the output of the camera; sync with the HDMI output of the ATEM should only come into play if the video is being processed within the ATEM (format conversion or DVE can add 1-2 frames of delay).

The other side of it is the USB output from the ATEM. That output does impose further delay due to the time taken to encode the video into a compressed format. The output is intended for live streaming, so this delay is not normally an issue for its intended purpose - there is FAR more latency imposed by the time taken to send all of the video across the internet to the streaming host and from there to the viewers, so the latency imposed by the USB connection is a drop in the bucket for its intended purpose.

Again, if your audio were fed into the ATEM it would automatically be delayed to match the video from the ATEM. As you are feeding the audio independently of the ATEM, that delay is not present and they can get out of sync unless you delay it separately yourself.

Return to Live Production

Who is online

Users browsing this forum: No registered users and 15 guests