[hpsdr] CW sidetone through speakers on PC instead of Headphone on ANAN SDR

[Scott Traurig]

Disclaimer: I am a huge evangelist of "radio virtualization", i.e. running
the radio with only an Ethernet connection for audio and PTT.

VAC works better than it ever has due to a number of improvements that have
been rolled out over the last few versions. It does take a little fussing
with the settings, though, and it is difficult to predict what the right
settings will be for any individual configuration. I generally recommend
people start with a high value for primary sample rate buffer size, a high
value for VAC buffer size, and the MME driver. First try to find a VAC
buffer size that will work with 0 buffer latency. If that fails, then fuss
with the buffer latency setting somewhat randomly until things start to
converge. Then try to shave down settings for lowest latency, working on
VAC buffer size first.

If you have an ASIO capable sound interface on the PC, PowerSDR mRX 3.3.7
will provide, with a reasonably fast PC and suitable finessing of the
buffer settings, around 40mS of latency via the VAC interface in either the
TX or RX direction with a DSP filter (passband) filter length of 2048,
which is plenty sharp. This should be adequate for semi-break-in CW with PC
attached paddle and speakers. And latency can be improved by running a
filter length of 1024 or less. However, there is no sidetone, as already
discussed, and the work around is to use a keyer with sidetone as the input
to PowerSDR mRX.

Warren has been working on some enhancements to how the DSP filter
buffering and certain other buffering is done. This includes an option to
use minimum phase FIR filtering instead of the current linear phase
filtering. Again, with ASIO sound hardware and a suitably quick PC that
allows minimum buffer sizes, this reduces latency to approx. 20mS at any
filter length setting, which is better than a KX3 and approaching both K3
and IC7300 territory.

It should be carefully noted that, with ASIO hardware and a quick PC,
latency using a Hermes architecture transceiver and PowerSDR mRX is nearly
identical whether the audio path terminates at the PC, or is obtained via
microphone and speaker connections directly to the radio. Obviously this is
because all audio processing (except sidetone) is done in PowerSDR mRX,
i.e. audio streams sourced or sinked at the radio must travel to/from the
radio CODEC chip, through the FPGA, and up/down the Ethernet, to be
processed by the software. Indeed, if it were not for the vagaries of the
Windows sound system, VAC sourced/sinked audio would enjoy less latency!

Finally, for those who wish to try some of this out, it is worth mentioning
the COM port PTT facility in PowerSDR mRX, found in Setup > CAT Control >
CAT. A dry switch circuit that shorts Pins 7 and 8 of a nine pin COM port
works perfectly with "RTS" selected as the PTT signaling method. Note that
this is a bit of a misnomer in the software, as in reality it is CTS being
asserted that makes this work.

73!

Scott/w-u-2-o






On Thu, May 5, 2016 at 12:00 AM, Joe Martin <k5so at k5so.com> wrote:

> Well, Scott, after examining the current HPSDR PowerSDR software I don’t
> see that option either.  In fact, I’m having difficulty getting the VAC1 to
> even work on my machine with the current software/firmware.  I remember
> being able to monitor CW via VAC1 in earlier versions but perhaps that is
> no longer possible.  I know there has been a lot of work on the VAC code in
> HPSDR PowerSDR since I used it in previous versions of HPSDR PowerSDR.
>
> The ability to disable internal (to the FPGA) CW generation does exist in
> the presently not-released new protocol via the Transmitter-Specific port
> command packet protocol (byte 5, bit 1), so I seem to have been confused
> about that also being in the current-protocol when I sent my comment.
> Confusing new and old protocol features is a hazard of working on both
> simultaneously, I apologize for that.
>
> As I stated earlier though, and it’s still true, VAC is not a viable
> option for monitoring CW with HPSDR PowerSDR due to the inherent latency of
> the VAC paths.
>
> 73, Joe K5SO
>
>
> On May 4, 2016, at 8:22 PM, Scott Traurig wrote:
>
> > Where is that option in the setup, BTW?
> >
> > 73,
> >
> > Scott
> >
> > On Wed, May 4, 2016 at 10:12 PM, Joe Martin <k5so at k5so.com> wrote:
> > It can be done if you disable FPGA CW generation (i.e., select the
> option to generate the CW in the PC rather than in the FPGA) but as a
> practical matter it’s a moot point due to the latency involved with VAC.
> >
> > 73, Joe K5SO
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20160505/9f25978c/attachment.htm>