It is currently Sat Jul 22, 2017 12:36 pm Advanced search

Audio Delay Tests

QTAKE general discussion

Audio Delay Tests

Postby bbvcbob » Sat Jul 15, 2017 1:06 pm

In response to a recent discussion with other Qtake Operators re Audio Delays, I ran some tests to try and quantify the delays and to guide the Operators in the best application of the built in options in Qtake. Here are my results:

Qtake Audio Delay Tests

The issue of unavoidable Audio Delays through any HD video system is of particular concern in the case of Qtake, where the delays are both inherent and deliberately applied to match the latency of the video signal from Digital Movie cameras.

The most common complaints are, either, that the video image on the monitor is out of sync with the audio through the headphones or that the headphone wearer can hear the dialog both directly and, delayed, through the ‘cans’ causing an echo effect.

In order to quantify the problems and try to set out the best settings to apply to Qtake to overcome them, I ran some tests on one of my Qtake rigs with, I think, a standard hardware configuration.

My rig setup is:
• Mac Pro 'Trashcan’
• 2x Kona 3G Cards
• Focusrite Scarlett 2i2 USB Audio Interface
• BMD Smart VideoHub
• Matrox DH2G
• Gefen Dual DVI-SDI
• Mac OS 10.5.5
• Qtake v1.3 App
• AJA 12.5.1 Drivers

Output monitor: JVC DT-V24G1

Evidenced by my earlier tests on the video latency of individual monitor models and those on the delays of the various outputs available from Qtake – Live vs. Processed, Video Card vs. GPU/QOD, the actual delay required for an individual Operator’s set-up is likely to change but, hopefully, my tests can act as a guide.

I have assumed a 1.5 to 2 frame latency in the camera output* and used 2 frames as my reference.

The first test I ran was to find any delay, input to output, introduced by the Scarlett and/or any processing within the MacPro. With all of the Qtake settings at ‘0’ this delay was 28mS (0.6 Frames)

My preference is to feed live, unprocessed images to the monitors during Idle & Record modes and the output of the Kona Cards during playback**. This allows for the least delay between the live audio and the video output of the camera when in Idle & Record.

There are 3 Audio delay settings in the Qtake software:

1) In the Project settings. This value determines the delay applied to the recorded files and, when set correctly, syncs the video and audio on the recorded files. This setting is important if the files are likely to be transferred for use by, say, the Cutting Rooms. The value should be set to match the latency of the incoming image e.g. 2 frames if hardwired to a camera but more if using a Wireless Video link with it’s own latency.

2) In the AUDIO menu (use the >> arrows no navigate to the second page of the menu): The ‘Live Audio Output Device’ setting determines the delay applied to the audio throughput during Idle and Record modes. Logically the ‘Live Audio Output’ setting should be set to 2 frames to match the camera latency but that produces the ‘echo’ effect between the actual sound and that through the cans. If it is set to zero then, logically, the audio sounds correct but the lip-sync is off on the monitors. As the human brain has it’s own set of inherent latencies in both sight and hearing, it can happily correct small discrepancies and, so, around a single frame out of sync is not usually observable. On this basis, and taking into account the 0.6 frame delay introduced by the Scarlett, I do, in fact, leave this setting at ‘0’. The 0.6 frame audio delay does not sound like a distinct echo, rather more like a wide stereo image, and the now reduced lip-sync error (2 – 0.6 = 1.4 frame) on the monitor becomes hardly noticeable. I have found that, as long as the viewer or listener is unaware of the compromise and doesn’t go searching for the errors, this arrangement works very well.

3) The ‘Disk Audio Output Device’ setting determines the delay applied to the audio output during Playback. I presume that this setting is to correct for any delay in the monitoring chain, perhaps if streaming to a remote location where the video requires compressing with consequent latency from that process. For most on-set situations this can be left at ‘0’. It does, however, have some interactivity with the Project audio delay value in that, if the Project window setting is ‘0’, the recorded file will be out of sync and so the ‘Disk Audio Output Device’ setting will have to be altered to correct the playback sync.

* It is possible that the D.I.T. may introduce additional delay through their LUT processing.

** I have recently starting using the ‘Live Pass’ Video Hub routing control in Qtake, hence my choice of these outputs.
bbvcbob
 
Posts: 151
Joined: Tue Apr 19, 2011 8:07 am

Return to General

Who is online

Users browsing this forum: No registered users and 1 guest

cron