Rank: Advanced Member
Groups: Registered
Joined: 12/21/2016(UTC)
Posts: 39
Location: Germany
Thanks: 23 times
Was thanked: 10 time(s) in 7 post(s)
|
Hi everyone! It might be a very easy question, but I couldn´t find something via search, so:
Which is the best output option with the smallest amount of latency?
I have a bmd quad2 and i´d say that, when using this as an output I get about (estimated, did´t measure it) had a second of latency on my output. I have to do a live show in 2 weeks where I am responsible for the video on the video wall and this latency is to high for my taste. I read somewhere that I should use the second output of my main graphics card (nvidia gtx 960 4GB) to have the least amount of latency, but can not find the thread again. Is vMix intended to use this output as "MAIN" Output? Because I always thought that using the BMD Outputs via SDI would be the more "Pro" like output. Thanks for your help and greetings,
Thomas
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 3/20/2014(UTC) Posts: 2,721   Location: Bordeaux, France Thanks: 243 times
Was thanked: 794 time(s) in 589 post(s)
|
Hi Thomas I have no definitive answer for you but here's an idea for you to test and find the best solution. An easy solution to measure latency is to output a webpage containing a chronometer with at least hundreds of seconds Like this one for example : http://www.proglogic.com...ipt/time/chronometer.php(it's easy to modify the code to increase font size) You can connect an external monitor to each output option (next to your main vMix monitor) and then take photos at 1/1000 with both screens in frame. The resultings photos will allow you to measure the latency. BTW no solution will give you 0 latency. The best I ever had (depending on vMix load) is 6/7 images (that is to say 0.24 seconds) at 25fps. Guillaume
|
 1 user thanked DWAM for this useful post.
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 3/7/2012(UTC) Posts: 2,636 Location: Canada Thanks: 33 times
Was thanked: 506 time(s) in 475 post(s)
|
@ Harty
There are a lot of arguments that can be made about what is "Professional" and what is not, but in the end, what really matters is what is going to work best for you in your scenario...
Multiple factors contribute to latency in our digital world but essentially, from a simplistic viewpoint, each link in the chain adds a degree of latency to the signal. Typically, grabbing the signal straight from the GPU is the shortest chain in the vMix workflow and will give you the least amount of latency (how much is more dependent on actual hardware).
With a "live" on stage screen, anything more than a couple hundred milliseconds is going to be noticeable and should be avoided where possible and why many vMixers (including myself) choose to use the "Fullscreen" Output in those situations.
Hope that helps some.
Ice
|
2 users thanked IceStream for this useful post.
|
Harty on 2/6/2017(UTC), PFBM on 2/7/2017(UTC)
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 12/21/2016(UTC)
Posts: 39
Location: Germany
Thanks: 23 times
Was thanked: 10 time(s) in 7 post(s)
|
Thanks guys! I'll measure my different output optios and decide afterwards!
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 3/7/2012(UTC) Posts: 503 Location: The Netherlands Thanks: 9 times
Was thanked: 45 time(s) in 39 post(s)
|
hello Harty,
the way I do this is to bypass the main camera and feed this directly to the projector and then into vmix for the mixing recording and streaming.
i use a yuan capture card for that that also hace a sdi throughput.
this way you have zero latency on the screen and the camera operator will know what to do because he can see his feed on the screen.
if you want to mix you have to use a hardware mixer that will switch inputs directly, but i find that a main camea on the screen is ok.
see a screen grab here for the result
stefan
|
1 user thanked h2video.nl for this useful post.
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 3/7/2012(UTC) Posts: 503 Location: The Netherlands Thanks: 9 times
Was thanked: 45 time(s) in 39 post(s)
|
[img][/img] h2video.nl attached the following image(s):  image.png (893kb) downloaded 55 time(s).You cannot view/download attachments. Try to login or register.
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 12/21/2016(UTC)
Posts: 39
Location: Germany
Thanks: 23 times
Was thanked: 10 time(s) in 7 post(s)
|
H2video.nl: interesting idea, but in my case I'll have 6 cameras running so splitting them all up and get them on the beamer would include a second switcher and another person that operates it... plus the sdi distribution... maybe I'll go the other way round and use an atem switcher for mixing and vmix for holding all the clips and other stuff 🤔
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 5/24/2015(UTC)
Posts: 370
Location: Netherlands
Thanks: 16 times
Was thanked: 81 time(s) in 72 post(s)
|
Also keep in mind that running vmix at lower framerates like 25/30 will have more latency than running at 50/60 fps.
|
1 user thanked Mathijs for this useful post.
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 2/18/2014(UTC)
Posts: 1,837
Location: Stockholm
Thanks: 144 times
Was thanked: 297 time(s) in 250 post(s)
|
I just checked the Knowledge base on Latency, and it needs updating (from 2013 ;) In addition to what has already been said. In particular - forget "External" and use "Fullscreen" (secondary output/s) from the graphics card. You can enable "Low Latency Capture" under Settings/Performance. In the Help described as "This will reduce the latency of Camera inputs by approximately one frame. This will substantially increase the load on the graphics card and may lead to dropped frames under heavy load, so it should used with caution". We once did what Stefan suggested by duplicating the incoming camera cables, and mixed the secondary production using an ATEM switcher, but the gain was marginal. Another alternative might be to send all cameras out from the production via NDI, and mix the secondary on another PC with vMix there taking in the NDI signals.
|
1 user thanked richardgatarski for this useful post.
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 12/21/2016(UTC)
Posts: 39
Location: Germany
Thanks: 23 times
Was thanked: 10 time(s) in 7 post(s)
|
I did some comparisons today and can tell that the atem has a little less latency. BUT I think that my main problem lies in the whole chain. I send every HD-SDI Out to an atomos samurai, then take the output and send this to vmix/atem. Here lies the biggest "latency source"... when connecting a signal direct I can reduce the latency quite heavily... so it's not a vMix problem... btw, my vMix is quite stable around 10ms render time which is fine for me. "low latency" mode is on... so I'll just use the "external" for my program recordings... and use the main DVI of the GTX as MAIN output to the videowall...
|
1 user thanked Harty for this useful post.
|
|
|
|
Rank: Advanced Member
Groups: Registered
Joined: 5/24/2015(UTC)
Posts: 370
Location: Netherlands
Thanks: 16 times
Was thanked: 81 time(s) in 72 post(s)
|
Keep everything as direct as possible, instead of sending the signal trough the Samurai first, better split the SDI signal and send one to the Samurai and the other to vMix or the Atem. A simple videohub is also perfect to route this without adding latency.
In my research, a setup with ATEM had 120msec latency and vMix 160msec with same camera's and projector connected. So there are 2 frames to win at 50fps with everything freerunning.
It also depends on the size of the venue. At a big venue you can better aim to be in sync at half of the audience area. If it is sync on the first row, it will be way off at the back otherwise.
|
1 user thanked Mathijs for this useful post.
|
|
|
|
Rank: Newbie
Groups: Registered
Joined: 5/7/2020(UTC) Posts: 5  Location: Köln Thanks: 3 times
|
Just to let you know.
I had a noticable latency aswell.
Just as OP described it.
My routing chain was:
5 cameras (Sony PXW-FX9) via SDI into BMD ATEM Television Studio Pro 4K, than PGM from the ATEM into vMix nad then output via Decklink to an Atomos SUMO.
It all went to no noticeable latency at all when I synched two ambient lockits and tentacles e synch together.
On each camera I put a synched Tentalce and connected it to the TC Input of each camera and one to the Sound Devices audio mixer. (So 6 Tentacles)
Then I connected an ambient lockit to the ATEM Ref Input and antoher lockit to the Ref input of the Decklink card.
As the lockit can output TC and Gen Lock.
So each device is in the chain is frame synched.
That gave me none noticeable latency at each point of the chain.
It was so good, that even the sound which was put on to PA boxes (and into the Atomos SUMO via XLR) for the audience was in snych with the PGM recorded on the end of the chain (Atomos SUMO).
The production had some tentacles lying around and our audio engineer brought the lockits and more tentalces to the production with him.
|
|
|
|
|
|
Forum Jump
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.
Important Information:
The vMix Forums uses cookies. By continuing to browse this site, you are agreeing to our use of cookies.
More Details
Close
