Bug? SPDIF latency compared to cable

[Riccardo Ros]

I was testing the latency of a wireless device for the guitar and... I found out that the SPDIF has more latency than the cable... it makes no sense (less D/A conversion) but maybe I get it wrong.

Steps:

1. Created a preset with a synth and volume (to have an intermittent start stop)
2. Record the synth in reaper using:
   1. stereo spdif
   2. cable on left out2
   3. wireless on right out2
3. Measure the relative latency of the wireless compared to the cable

I measured ~4ms for the wireless system (not ideal but not bad) but also a little more than 4ms for the SPDIF!

Is this supposed to be correct?

The preset


The measurement

 

[Riccardo Ros]

I tried to create an aggregated device and testing again CABLE vs SPDIF vs USB...
Assuming the cable at 0, I have:

• USB ~1ms
• SPDIF ~5ms

I'm a little bit confused

 

[GlennO]

SPDIF itself does not have any latency. But the device the SPDIF signal goes to will have some latency in its connection to your computer. That latency can be affected by various factors, most significantly by the i/o buffer size.

 

[Riccardo Ros]

SPDIF itself does not have any latency. But the device the SPDIF signal goes to will have some latency in its connection to your computer. That latency can be affected by various factors, most significantly by the i/o buffer size.

Thanks for the explanation... but I was assuming that the latency in the daw impacted the input (cable or spdif) in the same way.
Obviously I was wrong.. in my test the block size is 16, so the 4ms is the worst latency I may have.

The part I don't get is what's the benefit of using the digital signal if I should compensate in the DAW for it... and if I had the FM9 with SPDIF input and output it becomes even harder.

 

[jeffgoobs]

The part I don't get is what's the benefit of using the digital signal if I should compensate in the DAW for it... and if I had the FM9 with SPDIF input and output it becomes even harder.

Fewer stages of conversion is one reason people might use S/PDIF. Others may prefer the DAC on their interfaces when it comes to monitoring.

 

[Riccardo Ros]

Fewer stages of conversion is one reason people might use S/PDIF. Others may prefer the DAC on their interfaces when it comes to monitoring.

Thanks, that's my main reason! It was more of a complaint on the latency.

I found some threads on other guitar devices and it's common to have a different latency analog vs digital.
I assume 4ms are not a deal breaker and now that I know the drill I will change my workflow accordingly: analog out in live, both in recording (analog for monitoring, digital for recording)

 

[GlennO]

Thanks for the explanation... but I was assuming that the latency in the daw impacted the input (cable or spdif) in the same way.
Obviously I was wrong.. in my test the block size is 16, so the 4ms is the worst latency I may have.

The part I don't get is what's the benefit of using the digital signal if I should compensate in the DAW for it... and if I had the FM9 with SPDIF input and output it becomes even harder.

Keep in mind that, in a properly configured rig, i/o latency is compensated for in your DAW. That means that, assuming you're monitoring direct, the i/o latency is largely irrelevant. As you know though, the FM3 has a latency reporting bug that requires a workaround. There are many benefits to keeping your signal digital, but as explained in the recording guide there are pros and cons to using digital SPDIF vs. digital USB.