Latency
- Thread starter rabbit
- Start date
Zane Condren
Senior Integrated Production Specialist
Here you go https://www.youtube.com/watch?v=T-eBCescEtk.
Red Dog Records
Member
Harmut,
We are putting something together to do a more realistic latency test.
Give me a day or so to post the video.
We are comparing:
I think you will be surprised.
video uploaded soon...
Andrew
We are putting something together to do a more realistic latency test.
Give me a day or so to post the video.
We are comparing:
- SDI input to Tricaster output
- SDI camera vs NDI camera (NewTek PTZ1)
- and NewTek NDI to NewTek SDI (Same Camera)
I think you will be surprised.
video uploaded soon...
Andrew
Red Dog Records
Member
Ganz toll!
Wir haben ein paar Veranstaltungen in dieser Woche, aber ich sollte in der Lage sein, um es hochgeladen am Freitag oder Samstag.
Entschuldige mein Deutsch. Meine Oma sprach nur zu Hause deutsch, aber ich bekomme nicht mehr viel Übung.
Andrew
Wir haben ein paar Veranstaltungen in dieser Woche, aber ich sollte in der Lage sein, um es hochgeladen am Freitag oder Samstag.
Entschuldige mein Deutsch. Meine Oma sprach nur zu Hause deutsch, aber ich bekomme nicht mehr viel Übung.
Andrew
Hallo Andrew,
mein englisch ist ja auch nicht so toll. Schön, dass hier auch mal deutsch gesprochen - geschrieben - wird. Ich bin sehr gespannt auf das Video. Danach will ich die Kamera bestellen, wenn alles gut ist.
(my english is not so great. Nice, that is also times spoken in German - written. I'm very excited about the video. Then I will order the camera, if everything is good.)
Hartmut
mein englisch ist ja auch nicht so toll. Schön, dass hier auch mal deutsch gesprochen - geschrieben - wird. Ich bin sehr gespannt auf das Video. Danach will ich die Kamera bestellen, wenn alles gut ist.
(my english is not so great. Nice, that is also times spoken in German - written. I'm very excited about the video. Then I will order the camera, if everything is good.)
Hartmut
Red Dog Records
Member
OK, here we go.
Since we do quite a bit of IMAG at our shows, we are always trying to reduce latency from camera to screen by as much as possible. As we see it (and correct me if I am wrong) there are three main points of latency:
(Feel free to add more points if you use D/As and converters.)
This initial test focused on the first two.
Part one focuses on the camera.
We had been told that although the Tricaster can adjust almost any incoming video signal for use, it is most efficient to set the camera's output to match the session. We were also told that the Tricaster latency happens based on number of frames, so more frames = less latency.
With those things in mind, our first test session was at 720/60p.
The results show the Camera 1 has the least amount of latency, with camera 2 about 1/50th of a second behind (~1 frame). The PTZ1 was about 1/20 of a second slower (~3 frames).
After finding out the camera with the least amount of latency, we ran test 2.
You can see the results in the video, but basically we were looking at about .11 seconds of latency from live to program monitor. At a 60p session, this equates to about 6-7 frames or so.
What we found interesting was that the 1080i camera performed better even though the session was at 720p. This could be because the camera has to do more processing to deliver a 720p image.
Our next steps will be to try this with a 1080i session and set all camera to 1080i.
We did a quick test with these settings and actually found that the PTZ1's NDI output was .01 second FASTER than the SDI output on the same camera!
Also , we want to test the speed of the NDI output of the tricaster.
The first step, however, will be to get a better timecode generator! this one is horrible and very hard to read. So sorry about that.
I hope this helps some folks.
Video is here: https://vimeo.com/236308108/3a1201396f
You can download the original 60p file to have better control over pausing and checking out the numbers.
Andrew
PS -We just got off a show where we used the PTZ1 for the first time. I think there is something wrong with the lens on our unit. We had a panel discussion on stage and the PTZ was doing a wide shot of the whole table. The center was in focus and the outer panelists were soft. There also seemed to be some 'halos' around the bright edges of the image. The lens was cleaned (on the outside) but we will do more testing to try to figure out what is going on. Everything is packed in a truck for Tuesday's show so no testing until later this week.
Since we do quite a bit of IMAG at our shows, we are always trying to reduce latency from camera to screen by as much as possible. As we see it (and correct me if I am wrong) there are three main points of latency:
- Camera
- Switcher
- Projector
(Feel free to add more points if you use D/As and converters.)
This initial test focused on the first two.
Part one focuses on the camera.
We had been told that although the Tricaster can adjust almost any incoming video signal for use, it is most efficient to set the camera's output to match the session. We were also told that the Tricaster latency happens based on number of frames, so more frames = less latency.
With those things in mind, our first test session was at 720/60p.
- Camera one - Sony NX5 was set to 1080/60i (connected through SDI)
- Camera two - Sony Z150 was set to 720/60p (SDI)
- Camera three - NewTek PTZ1 - 1080/60i (NDI)
- Camera four - same NewTek PTZ1 - 1080/60i (SDI)
The results show the Camera 1 has the least amount of latency, with camera 2 about 1/50th of a second behind (~1 frame). The PTZ1 was about 1/20 of a second slower (~3 frames).
After finding out the camera with the least amount of latency, we ran test 2.
You can see the results in the video, but basically we were looking at about .11 seconds of latency from live to program monitor. At a 60p session, this equates to about 6-7 frames or so.
What we found interesting was that the 1080i camera performed better even though the session was at 720p. This could be because the camera has to do more processing to deliver a 720p image.
Our next steps will be to try this with a 1080i session and set all camera to 1080i.
We did a quick test with these settings and actually found that the PTZ1's NDI output was .01 second FASTER than the SDI output on the same camera!
Also , we want to test the speed of the NDI output of the tricaster.
The first step, however, will be to get a better timecode generator! this one is horrible and very hard to read. So sorry about that.
I hope this helps some folks.
Video is here: https://vimeo.com/236308108/3a1201396f
You can download the original 60p file to have better control over pausing and checking out the numbers.
Andrew
PS -We just got off a show where we used the PTZ1 for the first time. I think there is something wrong with the lens on our unit. We had a panel discussion on stage and the PTZ was doing a wide shot of the whole table. The center was in focus and the outer panelists were soft. There also seemed to be some 'halos' around the bright edges of the image. The lens was cleaned (on the outside) but we will do more testing to try to figure out what is going on. Everything is packed in a truck for Tuesday's show so no testing until later this week.
Last edited: