RTAV - delay when starting an audio recording

[hema1]

Hi,

I'm currently testing audio recording in an VMware View environment using the Real-Time Audio-Video functionality, and I wonder if you could help me with the following:

when I start recording, there is a delay of 0.6 seconds until the recording effectively starts, easily reproducible with the Audacity recording tool:

I've tested with 3.5mm headsets as well as USB microphones, the result is always the same.

Any idea how to get rid of this delay, maybe there is a way to configure the size of the audio recording buffer?

Setup details:

VMware Horizon View 5.3
Host: Win7x64SP1

- Client 2.3

Virtualized Desktop: Win7x64SP1
- Agent 5.3
- Remote Experience Agent 5.3 (1.0.4364)

[peterbrown05]

hello there

so - network latency is one part of the 0.6seconds you see. The larger your network latency then the larger that delay is going to be.

In addition, there is some "fixed" latency, due to the buffering that we do to try and ensure we don't get audio jitter from buffer under-runs etc and also some latency due to the time it takes for us to encode and decode the audio/video. Most of this is "fixed". I believe there might be a few things that can be tweaked to reduce the audio buffer on the agent, but vmware does not test the non-default values of that, nor would I recomend it because it is likely to lead to audio quality issues if your network has a few glitches.

Is the 0.6seconds at the start a major problem for your use case? Could you explain in more detail (eg do you only need the audio and not video?), and I might be able to help further.

cheers

peterB

[hema1]

Hi Peter,

thanks for the quick reply, I would need it for dictation workflow software (no video required), that's why 0.6 secs is definitely too much...

Maybe you could give me a hint which settings can be modified, so I can play around a bit with different values?

[peterbrown05]

Hi there,

i will ping you privately and perhaps we can work together to try a few things out. Once we have some results, we can post back here for the community.

fwiw; I know that Nuance recently did some integration work with their 360 Speechmagic solution to integrate with VMware Horizon View. Now this might not solve your issue (depending on the 3rd party software you are using in your guest), but certainly the Nuance Horizon View optimized solution is great for dictation type applications, because not only are Nuance in full control of the codecs (and quality) of the audio encoding, but they also deliver the encoded stream directly into their end application. In this way, it saves a decode and a subsequent encode on the guest which would add audio artifacts - and so the solution would give better quality and lower latency (as a result of saving the decode+encode cycle).

I dont have any links directly to the Nuance info, but VMware has the following KB that calls out that this solution is available for SpeechMagic SDK 7 Release 4 FixPack 3 and later.

(see VMware KB:    Nuance SpeechMagic VMware View Extension - Supported USB Devices and USB Configuration...)

cheers

peterB

[inframi]

Hi, 

we have exactly the same problem with current versions of VMWare Horizon. Delay is exactly 670 ms. This is way too much when medical records are being dictated. Has a solution been found in the meantime?

VMware-Horizon-Agent-x86_64-7.11.0-15238678
VMware-Horizon-Connection-Server-x86_64-7.11.0-15231595
VMware-Horizon-Client-5.5.0-16975072
Windows 10 Client
Blast and PCoIP

[JAGTechnologies]

Hi,

I'm experiencing the same issues - has anyone found a solution to reduce this delay ?

Cheers

John