Converting 828mkii's output +4dBu to -10dBV

[pagauthier]

Hello,

I am interested in building a voltage divider so that the +4dBu outputs of the 828mkii can be converted to -10dBV to use with an loudspeaker array (16) driven by car amplifiers.

So, what is the output impedence of the 828mkii ... on the analog outs of course!

Otherwise, is their any soft/front panel option to set the 828mkii's outs to line level?

Many thanks!

[oldguitars]

just shave 14db off the faders....

[pagauthier]

I would prefer to keep the DAC's full dynamic range active ... otherwise, using numerical gain reduction, I would then move from 16 bits to 14 bits (1 bit lost per divide by two) used resolution.

I plan to make something as described there:
http://www.audiotic.com/Tools/VDC.htm and
there:
http://www.audiotic.com/DIY/DIYFS.htm

As creating a voltage divider just implies two resistances, it worth the few seconds spent on soldering. I just want to avoid the straight forward design like 60kOhm + 20kOhm gives a 1/4 divider ... and choose the resistance values to fit the output impendance of the card and the input impedance of the amp.

So its why I would like to know the output imp of the MOTU 828mkii analog outs ...

Thanks for the reply!

[HCMarkus]

If I remember correctly, there is a software switch in the control panel for the 828mkII that allows you to switch analog operating levels from -10 to +4 and vice-versa. Also, the converters are 24 bit, not 16 bit.

[pagauthier]

Right, but the soft switch (also accesible through the front panel) only controls the inputs (same thing for my 24IO) ... Although I never checked if this had any effect on the outputs. I should.

It is written very small in the manual that the outputs are only +4 ... while the inputs are changeable between +4 and -10.

And my software (Pure-Data) and sound files are all 16bits, so whatever the converter bit-depth is, the DAC is working with 16 bits data uselessly pushed to higher resolution ... then reducing by 75% my soft signals still reduce my res to 14 bits (divide by four, two bits gone ).

It is why I am so much interested in building attenuator-cables from +4 to -10.

Many thanks!

[HCMarkus]

Dude, if you care about sound quality (as it appears you certainly do based on your level matching concerns) you should be recording at 24 bits. It saves so much headache and potential loss of fidelity even when projects are ultimately destined for a 16-bit world... the headroom makes level setting much less critical, and avoiding the dreaded digital over is easy when one has 48 dB to play within!

If you are on a Mac, you might even want to try DP.

Good luck, and have fun!

[pagauthier]

You are more than right ... but sadly I am working with Pure-Data in which we created our own signal processing patch for a 16-loudspeaker cluster to control some acoustic beams ... and Pure-Data is not dealing with 24bits ... oops ... but I should check, since it would effectively be cool and simple, my concern would then be CPU usage .... since we always deal with 16 channels of output, with lot of DSP (an algorithmic composition triggered by sensors in a public space, ptich shifting, distortions, filters, etc.) ...

Thanks again!