Leif

Hi Clive!

Since the same thing happens with Spotify, at least we can rule out the CDs themselves as the cause.

What’s the CPU in your machine? Have you checked CPU usage when running?

I can’t think of any reason why it would stutter with buffers that big, especially not with kernel streaming, unless it’s running out of CPU completely, which is unlikely — even a 5 year old laptop easily handles breakaway using only a fraction of the available cpu power.

I developed a tool lately which may help us get to the bottom of the problem. Please check this thread: viewtopic.php?f=5&t=728

That way we should be able to narrow down whether the problem is on the input or output side of breakaway.

Transferring to a new computer is just a matter of installing and inputting the key on the new one.

68F is nice for this time of year, although where I’m at (Thailand) it’s 90F as usual. 🙂

///Leif

This might be a good time to mention that there is an undocumented digital MPX output in BBP. It’s always fed before tilt/eq calibration, and can work in parallel with the standard MPX output — but you have to hack the ini file to turn it on.

I’ve used it to feed a pipeline with digital mpx, while feeding a sound card with tilt-corrected mpx, so that i could look at it in mpxtool.

To use it:
* Open the I/O settings
* Select the device you want to use for DMPX in the normal MPX output. Press TEST, and then stop the test.
* Don’t close the I/O settings.
* Edit breakawaybroadcast.ini with notepad
* Search for TestMpxOutput
* Rename TestMpxOutput to DMpxOutput
* Make a copy of the DMpxOutput section and rename it TestDMpxOutput
* Save the file
* Go back to the I/O settings, select your normal MPX output.
* Press TEST again. You should see them both working.

If you want to turn off the DMPX output again, edit the ini file, and rename the DMpxOutput and TestDMpxOutput sections to something else, for example NoDMpxOutput, NoTestDMpxOutput .

Please note that the ONLY difference between MPX and DMPX outputs in BBP is that DMPX is unaffected by Tilt and PEQ!

Also please note that this undocumented feature is not an official feature, is not part of BBP ASIO, and is subject to change at any time.

Best,
///Leif

Actually Realtek HD Audio on-board can be excellent. It really depends on the motherboard.

I have a couple of cheap Asrock Dual-VSTA motherboards with Realtek 883, with DC COUPLED outputs (since they are so cheap they omitted the coupling capacitors), and as a result they’re EXCELLENT for mpx, pristinely clean.

You do need a scope to find out whether your board is good or not.
All the ones in the list are usable, but the only ones I really recommend are Marian and Juli@.

If you HAVE to use a laptop, which I don’t recommend to begin with, EMU 0202 is the best choice.

///Leif

Hi Clive!

Man, with those big buffering settings, it really shouldn’t be stuttering.

Please try the following: Disable "Audio Realtime Priority", and see if it makes a difference.

If not, I’ll have to think of something else.

Best,
///Leif

[quote author=”Sparky”]

The answer to this is yes… and no, it can be both. The design of DDS circuits are largely governed by the availability of the silicon that can clock at the appropriate rates for the final frequency of interest. Also factor in tuning resolution, cost, and power consumption will ultimately shape the overall design topology.[/quote]

Sparky, thank you, but.. that wasn’t all that clear, really. How about a description of the terminology as you understand it?

For example, what would you call a direct-to-channel exciter to differentiate it from those that use some kind of frequency doubling? For example, I’m pretty sure Nautel’s M50 is completely free from doubling, and if we’re aiming for maximum accuracy and transparency, this would appear to be an important factor.

///Leif

[quote author=”Sparky”]ILeif, DDS synthesis does not use a processor running an "algorithm". It’s essentially a clocked state machine that uses a 1/4 sine wave sine look-up table to drive a D/A converter. This converter output through smoothing filters reconstructs the FM signal (much like the sound card makes the MPX signal). By taking a numeric constant (modulation point) and adding it to the internal phase accumulators will generate a phase modulated output signal (FM). Because all the modulation is done in the digital domian, it requires a CPU to prep the audio sample point (plus any MPX signal data) into the proper numeric value prior to sending it on to the DDS chip.
[/quote]

Interesting!

This is basically how I wrote the virtual FM exciter in MpxTool (used for the Stokkemasker display). I was lazy though and used a full sine wave look-up table, and then also interpolation of the output values for increased resolution. Mine only runs at 1.536 MHz sampling rate though.

But hey, I was under the impression that DDS meant there’d be no upconversion at all, that the signal was generated at the correct frequency already to begin with.

Or, perhaps that’s what "direct to channel" means?

I’m confused. Could someone clear up the terminology for me? 🙂

///Leif

Sigmacom, that’s still kickass. Must be a very optimized algorithm, considering you’d have to have a sampling rate of at least 250 MHz or so (right?), that’s only a single cpu cycle per sample! Very impressive.

Are you selling this solution? This is extremely interesting.

///Leif