Firewire connections and sound interfaces

[Moon]

OK, here's a puzzle:

I've been experimenting with my Yamaha 01X and i88x. They both make usage of the mlan drivers using the firewire port. I initially had an S/N ratio of +/- 76 db, which is not bad. After some optimalisations, I managed to obtain an S/N ratio of 80 db which I consider to be a good result. I still have to replace some cables to balanced cables, so I hope to gain an even extra db's. Note, when turning the volume knobs down on both the 01x and i88x, I obtain an S/N of 86 db (measured in Sonar).

Now the puzzle: when connecting the setup to my laptop (a 3 year old Acer), and using the same driver and software, I measure an S/N of about 69 db

Apparently, the firewire connection is a determined factor for the performance of the sound interface ?

I can't explain this, but according to my measurements, a sound interface will perform differently depending on the PC you are connecting it too.

My advise: don't stare at the technical specs of a soundcard. They don't mean anything. The actual setup will determine how well the sound interface behaves. It is nice to read that the DAC or ADC have an S/N of 110 db, but the whole system will give you very different results.

Still, if someone has made the some observations as I did, or if anyone is interested to explain the phenomenal, I'll be very interested to hear about it.

Moon

[folderol]

Pure guesswork, but maybe the laptop has poor filtering on the power supply and you are picking up noise from that, either radiated, or sent down the power chord.

[Moon]

Pure guesswork, but maybe the laptop has poor filtering on the power supply and you are picking up noise from that, either radiated, or sent down the power chord.

Good thinking, and there's a way to test it: I'll do a recording session on battery power. This should rule out the power-grid noise.

I'll keep you posted.

Moon

[Marc JX8P]

Still, that would only matter to the audio cables going into the firewire interface, wouldn't it? Since the connection by fw cable to the laptop is pure digital any interference there wouldn't affect the audio (it could of course cause drop outs in very strong cases but shouldn't influence the audio quality since that's encoded).

What I am curious about is whether both systems use the sampe sample/bitrate for the audio since I think something like the bit resolution could be of influence on the S/N ratio.

[Moon]

Exactly my thoughts Marc. I did check the bidrate and sample frequence. Both were set identical leaving me with this puzzle.

However, if the FW cables mass wire is connected to the power supply mass cable, we could have the possible explanation.

I will redo the test, with and without the power supply...

Moon

[Marc JX8P]

On another note, how do you measure that S/N ratio?

[folderol]

On another note, how do you measure that S/N ratio?

Very carefully

[Moon]

On another note, how do you measure that S/N ratio?

That's easy: record the track in Sonar with the volume of the keyboard on recording level and the input of the interface to normal. Normal means you record the track on a near 0 dB level so a minimal on normalisation is needed afterwards. The recorded 'noise' can then be analysed in any wave editor or you just read out the peak level that Sonar is indicating. I think this is the correct way of doing it: jusrt recording without playing anything on the levels you would while playing.

You do record the noise level of the plugged in synth (in my case, a Korg M3 and a Roland X6 which are giving very similar results).

Moon

[folderol]

That does leave out one source of noise (although I don't know how relevant it is these days) and that is digital clock jitter. The apparent noise level is related to the signal amplitude, theoretical signal waveshape, signal frequency & clock frequency {phew}. If there is no signal being generated, there will be no noise from jitter.

[Moon]

The mysterie has been solved: unplugging the power supply removed all the noise. Grrrr... ground loop through the firewire port. Removing the audio cables from my Korg also removed the ground loop. I've ordered a pair of balenced cables. These should resolve the ground loop...

Moon