AFMG Network Forum

Niick · Member

I recently have begun the transition from Smaart v7 to SysTune, as I feel that Smaart was kinda like the on-ramp and SysTune is the Highway. However, I ran into something I dont understand. Using a Focusrite Saffire 6 USB 2 ch. interface, with the ASIO driver, sending pink noise into both channels thru a Y-splitter, i get the expected perfectly flat phase and magnitude response, but with a 3.2 msec. delay. With Smaart the delay is 0 msec, as it seems it would be. I have many other interfaces, they all show similar results. For some reason SysTune finds a not insignificant delay between the measure and reference channels when both are fed thru a Y splitter. Can anyone shed some light on this?

Waldemar · Member

Hi,
the delay is due to the electrical runtime. It is the time between the start of the feed of bits to the DA to the output and back from input over the AD to the driver. This internal delay is in principle known in the ASIO driver, but some drivers give wrong delays to the program.
We use these numbers of the driver and some drivers give almost zero delay. Examples are the "AUBION X8" and most of the RME devices.
Relative bad compensation can be done from Wave / Direct Sound drivers.
To get rid of this additional electrical delay, we recommend to use the Y-cable. One end feeds the loudspeaker and the other goes to the input 2 (for instance). The Mic is on Input 1.
The Reference in SysTune is to switch to Input 2 then.
To check the zero delay connect the other end (from the loudspeaker) of the Y-cable to input 1 and the impulse response will be at 0 ms.

tuuca · Member

Hi there,

I have problems with time delay between channels too. I supplied my customer with Aubion X.8 and Systune back in 2012. The application is to measure transfer function (Magnitude and phase chart) between two microphones and outher two microphones. Four channels in total.

First you measure the transfer function, then you swap the microphones and measure the transfer function again so you can compensate for phase mismatch of microphones (it is a precise measurement microphone pair, so the mismatch is neglectable, but we do it this way to be shure everything works fine).

Few years everything was working fine with no phase shift on any of four channels. Recently a linear phase shift is observed between channels 1 and 2 (channels 3 and 4 work perfectly). The phase shift corresponds with time delay of units of ms (like cca 1 ms).

I understand it could be an ASIO issue. But I dont understand that it worked well first and started behave strangely after few years. Of course. I dont operate it. My customer does, so anything could happen there.

Anyways. Here are my questions:
1) is there a setting option where my customer could accidentaly set delay to one of mic channels?
2) how do I restore factory settings to see if the problém can be removed with this.

thanks
tuuca

Niick · Member

Tucca,

Since the time I posted this question about a seemingly inexplicable delay between individual channels of a sound card/audio interface, I have gained quite a lot of experience with SysTune, and dual FFT in general.

With SysTune, there most certainly IS a way that the user could accidentally set a reference/receive delay differently for one channel as compared to another.

In the upper left corner of the screen you have the buttons that correspond to the number of available inputs (up to 8 ) of your interface.

If, for example, you select input button number one, with reference channel 4 (let's say you're using a 4 channel interface), and you set a receive delay of 2ms. You can then select button/input number 2 with the same reference (4), and set a completely different receive delay for channel 2 than you set for channel 1.

Also remember that if you're using a hardware input channel as you're refence, whatever receive delay has been set for that channel will be still applied to it whether it's being used as a reference or as an input.

I have noticed a bug in SysTune regarding phase measurements though (and this happens on 2 different computers, as I sometimes need to go mobile with SysTune so I move my license onto another, more portable setup)

The phase bug is this:

The phase trace doesn't update properly. If you're trying to watch the phase in real time, like when measuring the output of a DSP processor and making changes in the DSP, the phase trace will not update to track the changes (it will move, but it will be all wacky) until you select a different smoothing.

Simply select a different smoothing (or the same one, just re-select it) and the phase trace will instantly "correct itself" and "jump into position" where it should be.

Post back and let us know if any of this was helpful or if your customer was able to solve the issue.

Good day.

Nick

John W · Member

Hello Nick

I did a quick set up on Sunday night and didn't noticed the phase bug. I was having issue with Systune being wacky with other things which might be due to running sys tune with Meyer Sound Compass. I will be doing a set up on Friday and will try to the phase trace to go wacky.

Let you know.

tuuca · Member

Hi Niick!

I think you've got the right point. I was looking everywhere in settings and missed this option "on the desktop".

John W · Member

Has anyone also seen Systune do this phase bug? My machine does it kills my work flow. Is there a fix for it? Are you still have the same issue Nick?

Thanks

John

The phase bug is this:

The phase trace doesn't update properly. If you're trying to watch the phase in real time, like when measuring the output of a DSP processor and making changes in the DSP, the phase trace will not update to track the changes (it will move, but it will be all wacky) until you select a different smoothing.

Simply select a different smoothing (or the same one, just re-select it) and the phase trace will instantly "correct itself" and "jump into position" where it should be

Niick · Member

John,

Yes, this phase bug, as I understand it, plagues all copies of SysTune v1. 3.7

I have tested it on a number of different machines, windows 7 and 10, as well as brought it up at a recent training with Charles Hughes and Bruce Olson. They are aware of the bug and I have faith that a future version of SysTune will fix this issue.

So long as you're aware it's happening, it's not a big deal, at least for me.

I do remember when I was showing the bug to Charlie that he noticed that in addition to re-selecting your soothing, you can also edit a text field....maybe it was the dB offset feild in the upper left (right above delay offset), I think you can also re-select your FFT size.... probably a few other ways as well. Which leads me to believe that there may be a quick workaround using a "hot key" keyboard command. Not totally sure of this, but I'll check it out when I get back to work and report back.