I've been using the MotU 896 to make sound recordings for about a year, but
I've always listened to the final output on CD/DVD, not off the MotU itself.

However, recently I connected the outputs of the MotU 896 to our mixing
board and have noticed that it's output is substantially noisier than other
devices like computer sound cards and synths.

The mic preamps are phenomenally quiet, but the playback output amps are
phenomenally noisy. How can this be? Has anyone else noticed this anomaly? I
don't particularly worry about it, since my clients aren't listening to the
output of the MotU, but to the recordings made with the MotU. For that, it's
very much noiseless. But I was rather shocked and surprised that the D/A
playback amplifier circuits are this noisy. Probably not much better than
65dB, unweighted.

Anyone else find that their MotU playback jacks are a little noisy? Probably
nice and quiet on the AES/EBU outs, but I haven't tried them yet.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

Have you noticed a drop in level also?? Is it only noisy?

"ivis" wrote in message
oups.com...
Have you noticed a drop in level also?? Is it only noisy?


The level seems fine. It's only noisy. Rather surprising, since the input
mic preamps are incredibly free of noise. One would think the quality would
be consistent, input to output.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

Mark & Mary Ann Weiss wrote:
*snip*
Anyone else find that their MotU playback jacks are a little noisy? Probably
nice and quiet on the AES/EBU outs, but I haven't tried them yet.
*snip*

Absolutely not. I've regularly rigged up the inputs and outputs like a
patch bay so i can use hardware as inserts in Cubase. That would certain
be the litmus test to see whether the signal being sent out is returned
noisier than how it left. This is not the case, the original is idential
to the newly bounced/effected signal.

Have you tested all of the analog outputs on the motu? or just the main
outs? I had a problem with an 896 right out of the box where the left
channel was noisy (white noise masking the sound), and the right channel
was much quieter and even noisier. I returned it for a brand new one.
That my only encounter with a noise related issue and that was an out of
box factory defect.

Roach

Mike Rocha wrote:
Mark & Mary Ann Weiss wrote:
*snip*

Anyone else find that their MotU playback jacks are a little noisy?
Probably
nice and quiet on the AES/EBU outs, but I haven't tried them yet.

*snip*

Absolutely not. I've regularly rigged up the inputs and outputs like a
patch bay so i can use hardware as inserts in Cubase. That would certain
be the litmus test to see whether the signal being sent out is returned
noisier than how it left. This is not the case, the original is idential
to the newly bounced/effected signal.

Have you tested all of the analog outputs on the motu? or just the main
outs? I had a problem with an 896 right out of the box where the left
channel was noisy (white noise masking the sound), and the right channel
was much quieter and even noisier. I returned it for a brand new one.
That my only encounter with a noise related issue and that was an out of
box factory defect.

Roach
I really like the 896, but in my experience it hasn't been the most
reliable box. I'd get it checked by MOTU. I had to get the inputs of
mine fixed...

Absolutely not. I've regularly rigged up the inputs and outputs like a
patch bay so i can use hardware as inserts in Cubase. That would certain
be the litmus test to see whether the signal being sent out is returned
noisier than how it left. This is not the case, the original is idential
to the newly bounced/effected signal.

Have you tested all of the analog outputs on the motu? or just the main
outs? I had a problem with an 896 right out of the box where the left
channel was noisy (white noise masking the sound), and the right channel
was much quieter and even noisier. I returned it for a brand new one.
That my only encounter with a noise related issue and that was an out of
box factory defect.

Roach

I just moved the XLR connectors around to various outputs, not just the
stereo main outs... the hiss is the same on ALL outputs. Considerably
noisier than my Turtle Beach PC sound card and a lot noisier than my
synthesizers.

I threw this question to MotU tech support, but have not received a reply in
three weeks.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

I really like the 896, but in my experience it hasn't been the most
reliable box. I'd get it checked by MOTU. I had to get the inputs of
mine fixed...

The inputs on mine are great. I have no issues there. I can live with the
noisy outputs, since I'm not mastering off of them (analog), but it does irk
me that a $49 PC sound card is much quieter than this $1249 unit.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

"Mark & Mary Ann Weiss" wrote in
message
nk.net


I just moved the XLR connectors around to various
outputs, not just the stereo main outs... the hiss is the
same on ALL outputs. Considerably noisier than my Turtle
Beach PC sound card and a lot noisier than my
synthesizers.

Not a fair comparison. The maximum undistorted output of a
typical consumer sound card is 1.0 volts, per the AC97 spec,

Most audio production audio interfaces have maximum
undistorted outputs in the 3-6 volt range. Therefore their
residual noise can be 10-16 dB greater, and they will still
have the same or greater dynamic range.

What happens when you try to run the RMAA 5.5 audio
benchmark program on your Motu? It's a free download from

http://audio.rightmark.org/download.shtml

In article et writes:

I just moved the XLR connectors around to various outputs, not just the
stereo main outs... the hiss is the same on ALL outputs. Considerably
noisier than my Turtle Beach PC sound card and a lot noisier than my
synthesizers.

Have you measured signal-to-noise? How much signal do you get out for
a full scale digital recording? Maybe all you need to do is turn your
playback gain down about 20 dB. Sorry if this is too obvious, but I
just pop in and out of this thread and I don't know how much more you
know (or don't) about the gain staging.



--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me he double-m-eleven-double-zero at yahoo

Not a fair comparison. The maximum undistorted output of a
typical consumer sound card is 1.0 volts, per the AC97 spec,

I realize that, which is why I also compared against my Kurzweil K2600RS
rack sampler's balanced TRS outs. The Kurzweil is almost 20dB quieter.
And the fact that I have a temporary unbalanced cable connecting the MotU to
the mixer, it should be -6dB with reference to the Kurzweil, if the noise
levels are the same in both systems.


Most audio production audio interfaces have maximum
undistorted outputs in the 3-6 volt range. Therefore their
residual noise can be 10-16 dB greater, and they will still
have the same or greater dynamic range.



What happens when you try to run the RMAA 5.5 audio
benchmark program on your Motu? It's a free download from

http://audio.rightmark.org/download.shtml


This is a very interesting and useful test program. First of all, thanks for
bringing it to my attention. I wish I had found this last December when I
was aligning our house sound system with a new crossover loudspeaker
management system. It would have made the process much quicker.

Now about the test, I did a loopback with 25' ProCo microphone cables from
the stereo outputs to the 7-8 channel inputs.
The test graph shows most of the noise to be in the -128dB range with a few
spikes pushing toward -100dB.

I find the numbers hard to believe because this is measuring input to output
s/n ratio and this unit is audibly noisier than the Kurzweil, also a
balanced XLR output that can swing about 10V RMS undistorted.

Given that all the inputs to the board have the same gain structure, it
would suggest that the MotU is producing more quiescent noise than the
Kurzweil.
To be sure it wasn't a difference in input gain, I used input 1 on the board
and did the following:

With K2600RS plugged in, potted 1 up to max (+10dB setting). A barely
audible hiss was present.

Potted back to 0dB and plugged the MotU into that input. More hiss was
audible at 0dB than with the Kurzweil at +10db. When this input was potted
up to +10dB, the hiss was very loud.

So here we have two devices, both have balanced outs, but both are connected
via unbalanced cables (until we have the budget to get some new balanced
cables) connecting them to the board. One device is about 20dB noisier than
the other.

This test software is telling me that over most of the frequency range, the
s/n is about 128dB. If that were so, then I guess the Kurzweil would measure
148dB or so on s/n, but I don't know of any op-amps in existence today that
approach 148dB s/n ratio.

I doubt that the MotU can produce more voltage swing than that Kurzweil, as
both have similar output devices and voltage rail potentials.

Interestingly, the Kurzweil measures -72dB unweighted on my H/P 400E AC
Voltmeter, across pins 2 & 3 of the output.
The MotU measures -54dB across pins 2 & 3.
Yet the test software claims miraculous results. I'm having trouble
reconciling these results.
How can the MotU be that much noisier, yet post such astonishing test data?

We have everything trimmed on the board so that a 0dB "maximum" recording
level produces a uniform level from all devices with the board faders at
0dB. Some devices are just much noisier, and that's what prompted this
inquiry.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

"Mark & Mary Ann Weiss" wrote in
message
k.net

Not a fair comparison. The maximum undistorted output of
a typical consumer sound card is 1.0 volts, per the AC97
spec,

I realize that, which is why I also compared against my
Kurzweil K2600RS rack sampler's balanced TRS outs. The
Kurzweil is almost 20dB quieter.

The Kurzweil may have a muting circuit.

The Kurzweil may have an analog-domain gain control with an
exceptional lot of dynamic range.


Most audio production audio interfaces have maximum
undistorted outputs in the 3-6 volt range. Therefore
their residual noise can be 10-16 dB greater, and they
will still have the same or greater dynamic range.

What happens when you try to run the RMAA 5.5 audio
benchmark program on your Motu? It's a free download
from

http://audio.rightmark.org/download.shtml

This is a very interesting and useful test program. First
of all, thanks for bringing it to my attention. I wish I
had found this last December when I was aligning our
house sound system with a new crossover loudspeaker
management system. It would have made the process much
quicker.

20-20 hindsight. BTW I find the loudspeaker testing facility
to be very useful for live sound.

Now about the test, I did a loopback with 25' ProCo
microphone cables from the stereo outputs to the 7-8
channel inputs.
The test graph shows most of the noise to be in the
-128dB range with a few spikes pushing toward -100dB.

That is characteristic of all but the very best audio
interfaces. It's better than an Audiophile 24/96, more like
a Card Deluxe or Audiophile 24/192, but short of a LynxTWO.

I find the numbers hard to believe because this is
measuring input to output s/n ratio and this unit is
audibly noisier than the Kurzweil, also a balanced XLR
output that can swing about 10V RMS undistorted.

Like I said the Kurweil may be a little more complex than
just a sound card which is typically a DAC chip driving a
buffer amp that is hooke directly to the output terminals.

Given that all the inputs to the board have the same gain
structure, it would suggest that the MotU is producing
more quiescent noise than the Kurzweil.

Agreed. BTW, you can look at the output of the Kurzweil
using the FFT analysis feature of RMA55

To be sure it wasn't a difference in input gain, I used
input 1 on the board and did the following:

With K2600RS plugged in, potted 1 up to max (+10dB
setting). A barely audible hiss was present.

There still might be a muting circuit in the Kurzweil.

Potted back to 0dB and plugged the MotU into that input.
More hiss was audible at 0dB than with the Kurzweil at
+10db. When this input was potted up to +10dB, the hiss
was very loud.

So here we have two devices, both have balanced outs, but
both are connected via unbalanced cables (until we have
the budget to get some new balanced cables) connecting
them to the board. One device is about 20dB noisier than
the other.

But, they are very different devices. The MOTU is performing
about the way I would expect a device like it to perform. It
costs like a LynxTWO but it has a ton more channels.

This test software is telling me that over most of the
frequency range, the s/n is about 128dB.

Is that the numeric number reported by RMA55?

If that were so,
then I guess the Kurzweil would measure 148dB or so on
s/n, but I don't know of any op-amps in existence today
that approach 148dB s/n ratio.

I suspect that the Kurzweil is more complex than the MOTU.

I doubt that the MotU can produce more voltage swing than
that Kurzweil, as both have similar output devices and
voltage rail potentials.

I don't know if the MOTU can put out 10 volts at digital FS.
Got a DVM?

Interestingly, the Kurzweil measures -72dB unweighted on
my H/P 400E AC Voltmeter, across pins 2 & 3 of the output.

That's wideband noise. Noise measurements are only proper
if done over a stated bandwidth, like 20Hz-20 KHz.

The MotU measures -54dB across pins 2 & 3.

So, its 18 dB noisier in this questionable but somehow
indicative broadband noise test.

Yet the test software claims miraculous results. I'm
having trouble reconciling these results.
How can the MotU be that much noisier, yet post such
astonishing test data?

I think I need to see some numbers from RMA55.

We have everything trimmed on the board so that a 0dB
"maximum" recording level produces a uniform level from
all devices with the board faders at 0dB. Some devices
are just much noisier, and that's what prompted this
inquiry.

Well, some devices *are* noisier than others. Got any mics?
;-)

In article et,
(Mark & Mary Ann Weiss) wrote:

I just moved the XLR connectors around to various outputs, not just the
stereo main outs... the hiss is the same on ALL outputs. Considerably
noisier than my Turtle Beach PC sound card and a lot noisier than my
synthesizers.

I threw this question to MotU tech support, but have not received a
reply in
three weeks.

Have you tried calling them? I see none of these problems on my MOTU896HD,
and I strongly suspect yours is broken. It seems a shame that you're
wasting so much time and energy on it.

In article t writes:

Interestingly, the Kurzweil measures -72dB unweighted on my H/P 400E AC
Voltmeter, across pins 2 & 3 of the output.
The MotU measures -54dB across pins 2 & 3.
Yet the test software claims miraculous results. I'm having trouble
reconciling these results.

I can't help but think that you're seeing some of the input coming out
the output of the MOTU. What voltage do you read out of it if you
connect a shorted plug to the input of the channel you're reading?



--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me he double-m-eleven-double-zero at yahoo

I realize that, which is why I also compared against my
Kurzweil K2600RS rack sampler's balanced TRS outs. The
Kurzweil is almost 20dB quieter.

The Kurzweil may have a muting circuit.

The Kurzweil may have an analog-domain gain control with an
exceptional lot of dynamic range.

I strongly doubt the presence of a muting circuit. I would hear a 'gating'
effect on soft passages if that were the case. In fact, I consider the
Kurzweil to be noisy, compared to my Denon CD player. The Denon's output
noise playing a test CD with all digital 0's is below the noise floor of my
mixer (the noise is inaudible with the mixer fader at +10dB and the monitor
volume at 10 o'clock where I have it for the other tests. If I max out the
monitor volume, then the noise level is barely different with the CD player
connected or not connected.
A group of musicians who use Kurzweil K series samplers have also raised the
issue about the hiss. We don't consider that a $7995 sampler should be
noisier than a $350 CD player as being reasonable for the money. And along
comes the MotU, and it's 18dB noisier still. No, the Kurzweil definately
does not mute. If it did, we wouldn't hear any hiss on it's outputs.

http://audio.rightmark.org/download.shtml

This is a very interesting and useful test program. First
of all, thanks for bringing it to my attention. I wish I
had found this last December when I was aligning our
house sound system with a new crossover loudspeaker
management system. It would have made the process much
quicker.

20-20 hindsight. BTW I find the loudspeaker testing facility
to be very useful for live sound.

It does have issues with the 24/96 mode though. I can't get it to calibrate
at all. Works okay in 16/44.1 mode.
Did some measurements on the house sound system and got some odd results: no
matter where I placed the mics, the response curve was the same--even if I
moved them from center between the array to a few inches from a subwoofer.
The latter position should have showed a radical roll off after 130hz, but
it didn't. In fact, I could hardly tell the two curves apart from the
wideband test over the whole frequency range.
When I test with white noise, and an FFT analyzer, mic placement makes very
obvious differences in response curve.


The test graph shows most of the noise to be in the
-128dB range with a few spikes pushing toward -100dB.

That is characteristic of all but the very best audio
interfaces. It's better than an Audiophile 24/96, more like
a Card Deluxe or Audiophile 24/192, but short of a LynxTWO.

The inputs were certainly quiet enough, but I don't believe these numbers,
since they include the output + input noise combined. My only complaint was
the -.6dB rolloff at 20KHz, increasing to -3.6dB at 40KHz. But since I can't
hear anything up there, it's not much of an issue.


I find the numbers hard to believe because this is
measuring input to output s/n ratio and this unit is
audibly noisier than the Kurzweil, also a balanced XLR
output that can swing about 10V RMS undistorted.

Like I said the Kurweil may be a little more complex than
just a sound card which is typically a DAC chip driving a
buffer amp that is hooke directly to the output terminals.

The Kurzweil has ten outputs channels in 5 groups. 1 group is Mix outputs
and they are noisier by almost ten dB. The individual bus outputs are the
quietest, but I suspect these are direct off the DACs, whereas the Mix have
an additional mixer stage. But I do not consider the Kurzweil to be quiet.
My Denon CD player is much quieter than this.



Agreed. BTW, you can look at the output of the Kurzweil
using the FFT analysis feature of RMA55

I suppose I could. I'll have to fashion a test setup. Of course, it will
involve the MotU, but I trust the input stages at least.


To be sure it wasn't a difference in input gain, I used
input 1 on the board and did the following:

With K2600RS plugged in, potted 1 up to max (+10dB
setting). A barely audible hiss was present.

There still might be a muting circuit in the Kurzweil.

Strongly doubt it. It's too noisy to have a muting circuit and I don't hear
any evidence of it coming out of a muted condition when signal is ramped up.


Potted back to 0dB and plugged the MotU into that input.
More hiss was audible at 0dB than with the Kurzweil at
+10db. When this input was potted up to +10dB, the hiss
was very loud.

So here we have two devices, both have balanced outs, but
both are connected via unbalanced cables (until we have
the budget to get some new balanced cables) connecting
them to the board. One device is about 20dB noisier than
the other.

But, they are very different devices. The MOTU is performing
about the way I would expect a device like it to perform. It
costs like a LynxTWO but it has a ton more channels.

I will have to do more experiments, using normalized audio to max out the
MotU output levels and do some listening tests.


This test software is telling me that over most of the
frequency range, the s/n is about 128dB.

Is that the numeric number reported by RMA55?

That is the graph chart I was describing.
This is the text output:

Frequency response (from 40 Hz to 15 kHz), dB: +0.05, -0.45 Good
Noise level, dB (A): -103.0 Excellent
Dynamic range, dB (A): 103.0 Excellent
THD, %: 0.0008 Excellent
IMD + Noise, %: 0.0035 Excellent
Stereo crosstalk, dB: -100.7 Excellent
IMD at 10 kHz, %: 0.0037 Excellent





If that were so,
then I guess the Kurzweil would measure 148dB or so on
s/n, but I don't know of any op-amps in existence today
that approach 148dB s/n ratio.

I suspect that the Kurzweil is more complex than the MOTU.

Not from the output stage, but from the fact that it is much more than a
DAC, being a sampler, synth, ROM playback device with sophisticated control
capability, but no, the output configuration, aside from having ten
channels, is pretty conventional.


I doubt that the MotU can produce more voltage swing than
that Kurzweil, as both have similar output devices and
voltage rail potentials.

I don't know if the MOTU can put out 10 volts at digital FS.
Got a DVM?

I have an RMS voltmeter. Only challenge is getting a sine tone out of the
Kurz. I think there is a diagnostic mode that does this, but I have to dump
all my sample memory (requires reset) to get to it and I don't want to do
that right now.
If you figure the rail voltage of +/- 15VDC, the op amps will typically
swing +/- 13.5 and .707 of that is 9.54 vrms.


Interestingly, the Kurzweil measures -72dB unweighted on
my H/P 400E AC Voltmeter, across pins 2 & 3 of the output.

That's wideband noise. Noise measurements are only proper
if done over a stated bandwidth, like 20Hz-20 KHz.

Yes, but as a relative test, it confirms what my ears tell me. I guestimated
close to 20dB noiser than the K2600RS. The meter told me it was 18dB
noisier.


The MotU measures -54dB across pins 2 & 3.

So, its 18 dB noisier in this questionable but somehow
indicative broadband noise test.

Relative comparison, backing up what I heard.


Yet the test software claims miraculous results. I'm
having trouble reconciling these results.
How can the MotU be that much noisier, yet post such
astonishing test data?

I think I need to see some numbers from RMA55.

Well now you've got them.


We have everything trimmed on the board so that a 0dB
"maximum" recording level produces a uniform level from
all devices with the board faders at 0dB. Some devices
are just much noisier, and that's what prompted this
inquiry.

Well, some devices *are* noisier than others. Got any mics?
;-)

Yeah, but there isn't an acoustic space quiet enough to reveal the self
noise in the mics or the MotU preamps.
I can only verify that by digitally normalizing the recording. Failing to do
that, and relying on the MotU output stage gain, boosting the level in the
analog domain only brings up a lot of hiss. But in the digital domain,
normalizing a recording whose peak level was -54dB to 0dB in SoundForge,
results in pickup of a lot of ambient noise, but still no audible hiss.
That's why I find the MotU outputs to be noisy. Relative to the inputs +
mics, that is.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

I can't help but think that you're seeing some of the input coming out
the output of the MOTU. What voltage do you read out of it if you
connect a shorted plug to the input of the channel you're reading?

Doesn't matter if the inputs are turned all the way down. In fact, for my
test, the inputs were muted. Monitor mode is only invoked when recording
software is running and in record mode. But during this test, nothing
connected to the inputs was heard on the outputs.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

Have you tried calling them? I see none of these problems on my MOTU896HD,
and I strongly suspect yours is broken. It seems a shame that you're
wasting so much time and energy on it.


I finally did get a response via e-mail. However, they want to register the
unit and require a sales receipt before they'll issue and RMA. I don't
consider the issue to be one that warrants sending the unit out (although
the $45 flat repair fee is incredibly reasonable), only to be told that
there is nothing wrong with it. The RightMark Audio Analyzer's spectrum
graph shows that the total loop output to input s/n ratio is about 128dB
over most of the frequency range, with a few spikes dipping toward 100dB. If
MotU uses a similar test, then they will say that it's meeting specs.
I'm trying to understand why it's so much noisier than any other piece of
equipment in this studio, when all devices have been trimmed in the mixing
desk for 0dB with digital full scale signals.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

Have you measured signal-to-noise? How much signal do you get out for
a full scale digital recording? Maybe all you need to do is turn your
playback gain down about 20 dB. Sorry if this is too obvious, but I
just pop in and out of this thread and I don't know how much more you
know (or don't) about the gain staging.

Well this is interesting...

I generated a 0dB sine wave in SoundForge and output it to the MotU channels
1-2. The MotU meters are registering -6dB, not even -1dB, the last segment
before [overload]. So the Analog Outs are not even being driven to full
output. Very curious indeed.... now wondering if these meters respond the
same as a broadcast transmitter modulation monitor, that being, you can
never get 100% modulation with a sinusoidal wave...
Nope... just tried a square wave and got the same -6dB reading on the Analog
Out meters.
Quite a mystery.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

I generated a 0dB sine wave in SoundForge and output it to the MotU
channels
1-2. The MotU meters are registering -6dB, not even -1dB, the last segment
before [overload]. So the Analog Outs are not even being driven to full
output. Very curious indeed.... now wondering if these meters respond the
same as a broadcast transmitter modulation monitor, that being, you can
never get 100% modulation with a sinusoidal wave...
Nope... just tried a square wave and got the same -6dB reading on the
Analog
Out meters.
Quite a mystery.


One mystery is solved:

I did some poking around in Windows and found that I could use the Volume
Control app to go to advanced properties and select something other than the
internal PC sound card, so I selected each of the MotU outputs and set the
volume from halfway to maximum.
This gives me 6dB more output, so now it's 6dB hotter than the Turtle Beach
card (as trimmed on the board), but it still has higher hiss levels overall.
6dB is a long way from where things need to be.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

"Mark & Mary Ann Weiss" wrote in
message
k.net
I realize that, which is why I also compared against my
Kurzweil K2600RS rack sampler's balanced TRS outs. The
Kurzweil is almost 20dB quieter.

The Kurzweil may have a muting circuit.

The Kurzweil may have an analog-domain gain control with
an exceptional lot of dynamic range.

I strongly doubt the presence of a muting circuit. I
would hear a 'gating' effect on soft passages if that
were the case. In fact, I consider the Kurzweil to be
noisy, compared to my Denon CD player. The Denon's output
noise playing a test CD with all digital 0's is below the
noise floor of my mixer (the noise is inaudible with the
mixer fader at +10dB and the monitor volume at 10 o'clock
where I have it for the other tests. If I max out the
monitor volume, then the noise level is barely different
with the CD player connected or not connected.
A group of musicians who use Kurzweil K series samplers
have also raised the issue about the hiss. We don't
consider that a $7995 sampler should be noisier than a
$350 CD player as being reasonable for the money. And
along comes the MotU, and it's 18dB noisier still. No,
the Kurzweil definately does not mute. If it did, we
wouldn't hear any hiss on it's outputs.

http://audio.rightmark.org/download.shtml

This is a very interesting and useful test program.
First of all, thanks for bringing it to my attention. I
wish I had found this last December when I was aligning
our house sound system with a new crossover loudspeaker
management system. It would have made the process much
quicker.

20-20 hindsight. BTW I find the loudspeaker testing
facility to be very useful for live sound.

It does have issues with the 24/96 mode though. I can't
get it to calibrate at all. Works okay in 16/44.1 mode.
Did some measurements on the house sound system and got
some odd results: no matter where I placed the mics, the
response curve was the same--even if I moved them from
center between the array to a few inches from a
subwoofer. The latter position should have showed a
radical roll off after 130hz, but it didn't. In fact, I
could hardly tell the two curves apart from the wideband
test over the whole frequency range.
When I test with white noise, and an FFT analyzer, mic
placement makes very obvious differences in response
curve.

Thats not how it has worked for me.

The test graph shows most of the noise to be in the
-128dB range with a few spikes pushing toward -100dB.

That is characteristic of all but the very best audio
interfaces. It's better than an Audiophile 24/96, more
like a Card Deluxe or Audiophile 24/192, but short of a
LynxTWO.

The inputs were certainly quiet enough, but I don't
believe these numbers, since they include the output +
input noise combined.

Right, but if the input is *that* (i.e., 10 dB) much more
quiet, then its contribution to the output noise numbers is
close to zip.

My only complaint was the -.6dB
rolloff at 20KHz, increasing to -3.6dB at 40KHz. But
since I can't hear anything up there, it's not much of an
issue.

Agreed. This looks like a typical 24/96 curve.

I find the numbers hard to believe because this is
measuring input to output s/n ratio and this unit is
audibly noisier than the Kurzweil, also a balanced XLR
output that can swing about 10V RMS undistorted.

Like I said the Kurweil may be a little more complex than
just a sound card which is typically a DAC chip driving a
buffer amp that is hooke directly to the output
terminals.

The Kurzweil has ten outputs channels in 5 groups. 1
group is Mix outputs and they are noisier by almost ten
dB.

Consistent with being the sum of 5 other things, each
contributing a bit of noise.

The individual bus outputs are the quietest, but I
suspect these are direct off the DACs, whereas the Mix
have an additional mixer stage. But I do not consider the
Kurzweil to be quiet. My Denon CD player is much quieter
than this.

A CD player can't have more than about 98 dB dynamic range -
that comes with the 16 bits. It can mute and be quieter, and
indeed most do. But that isn't really the same as having
more dynamic range for playing music.


Agreed. BTW, you can look at the output of the Kurzweil
using the FFT analysis feature of RMA55

I suppose I could. I'll have to fashion a test setup. Of
course, it will involve the MotU, but I trust the input
stages at least.

To be sure it wasn't a difference in input gain, I used
input 1 on the board and did the following:

With K2600RS plugged in, potted 1 up to max (+10dB
setting). A barely audible hiss was present.

There still might be a muting circuit in the Kurzweil.

Strongly doubt it. It's too noisy to have a muting
circuit and I don't hear any evidence of it coming out of
a muted condition when signal is ramped up.

I'm thinking of a muting circuit that knows when there is no
signal. Any signal, and its off.

Potted back to 0dB and plugged the MotU into that input.
More hiss was audible at 0dB than with the Kurzweil at
+10db. When this input was potted up to +10dB, the hiss
was very loud.

So here we have two devices, both have balanced outs,
but both are connected via unbalanced cables (until we
have the budget to get some new balanced cables)
connecting them to the board. One device is about 20dB
noisier than the other.

But, they are very different devices. The MOTU is
performing about the way I would expect a device like it
to perform. It costs like a LynxTWO but it has a ton
more channels.

I will have to do more experiments, using normalized
audio to max out the MotU output levels and do some
listening tests.

I'm quite sure that if you set levels with music playing
just below clipping, and then remove the music, you aren't
going to hear any hiss from the MOTU unless you do something
crazy like put your ear next to the speaker.

This test software is telling me that over most of the
frequency range, the s/n is about 128dB.

No, its telling you that the noise in each FFT bucket
is -128 dB. Sum them all up over a reasonable range to get
the true noise level.

Is that the numeric number reported by RMA55?

That is the graph chart I was describing.
This is the text output:

Frequency response (from 40 Hz to 15 kHz), dB:
+0.05, -0.45 Good Noise level, dB (A): -103.0
Excellent Dynamic range, dB (A): 103.0 Excellent
THD, %: 0.0008 Excellent
IMD + Noise, %: 0.0035 Excellent
Stereo crosstalk, dB: -100.7 Excellent
IMD at 10 kHz, %: 0.0037 Excellent

Looks fine to me.

If that were so,
then I guess the Kurzweil would measure 148dB or so on
s/n, but I don't know of any op-amps in existence today
that approach 148dB s/n ratio.

I suspect that the Kurzweil is more complex than the
MOTU.

Not from the output stage, but from the fact that it is
much more than a DAC, being a sampler, synth, ROM
playback device with sophisticated control capability,
but no, the output configuration, aside from having ten
channels, is pretty conventional.

I'll be confident of what it really has inside when I see a
schematic (which I expect never to see!) ;-)

I doubt that the MotU can produce more voltage swing
than that Kurzweil, as both have similar output devices
and voltage rail potentials.

I don't know if the MOTU can put out 10 volts at digital
FS. Got a DVM?

I have an RMS voltmeter. Only challenge is getting a sine
tone out of the Kurz. I think there is a diagnostic mode
that does this, but I have to dump all my sample memory
(requires reset) to get to it and I don't want to do that
right now.
If you figure the rail voltage of +/- 15VDC, the op amps
will typically swing +/- 13.5 and .707 of that is 9.54
vrms.

Right, about +22 dBm

Interestingly, the Kurzweil measures -72dB unweighted on
my H/P 400E AC Voltmeter, across pins 2 & 3 of the
output.

That's wideband noise. Noise measurements are only
proper if done over a stated bandwidth, like 20Hz-20 KHz.

Yes, but as a relative test, it confirms what my ears
tell me. I guestimated close to 20dB noiser than the
K2600RS. The meter told me it was 18dB noisier.

You lucked out.

The MotU measures -54dB across pins 2 & 3.

So, its 18 dB noisier in this questionable but somehow
indicative broadband noise test.

Relative comparison, backing up what I heard.

Yet the test software claims miraculous results. I'm
having trouble reconciling these results.
How can the MotU be that much noisier, yet post such
astonishing test data?

I think I need to see some numbers from RMA55.

Well now you've got them.

They show relatively good performance for a mid-priced audio
interface.

We have everything trimmed on the board so that a 0dB
"maximum" recording level produces a uniform level from
all devices with the board faders at 0dB. Some devices
are just much noisier, and that's what prompted this
inquiry.

Well, some devices *are* noisier than others. Got any
mics? ;-)

Yeah, but there isn't an acoustic space quiet enough to
reveal the self noise in the mics or the MotU preamps.

Exactly.

I can only verify that by digitally normalizing the
recording. Failing to do that, and relying on the MotU
output stage gain, boosting the level in the analog
domain only brings up a lot of hiss. But in the digital
domain, normalizing a recording whose peak level was
-54dB to 0dB in SoundForge, results in pickup of a lot of
ambient noise, but still no audible hiss. That's why I
find the MotU outputs to be noisy. Relative to the inputs
+ mics, that is.

I doubt that will be true in an apples-to-apples comparison.

One thing about the MOTU inputs which must have gain
controls if they have mic preamps. What are the input gain
controls set to?

Mark & Mary Ann Weiss wrote:
Have you tried calling them? I see none of these problems on my MOTU896HD,
and I strongly suspect yours is broken. It seems a shame that you're
wasting so much time and energy on it.



I finally did get a response via e-mail. However, they want to register the
unit and require a sales receipt before they'll issue and RMA. I don't
consider the issue to be one that warrants sending the unit out (although
the $45 flat repair fee is incredibly reasonable), only to be told that
there is nothing wrong with it. The RightMark Audio Analyzer's spectrum
graph shows that the total loop output to input s/n ratio is about 128dB
over most of the frequency range, with a few spikes dipping toward 100dB. If
MotU uses a similar test, then they will say that it's meeting specs.
I'm trying to understand why it's so much noisier than any other piece of
equipment in this studio, when all devices have been trimmed in the mixing
desk for 0dB with digital full scale signals.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-



Hmmmm
I would just send it in. Get it looked at for $45. If it comes back and
they say "no, that's how it is", put it on ebay and get something which
doesn't bug you so much.
If it comes back and there was a problem, it will be solved.

In article t writes:

I generated a 0dB sine wave in SoundForge and output it to the MotU channels
1-2. The MotU meters are registering -6dB, not even -1dB, the last segment
before [overload]. So the Analog Outs are not even being driven to full
output. Very curious indeed....

This may be only a difference in interpretation of how 0 dBFS is
calibrated. I think the that current version of the definition is that
the meters should read 0 when the peak of a sine wave is at full
scale, but the meters may be indicating something else. If they read
-3, that would be easier to explain. Maybe Sound Forge and MOTU are
each compensating in the same direction.

Nope... just tried a square wave and got the same -6dB reading on the Analog
Out meters.

OK, so they're reading the average value of a square wave. It must
have something to do with the response time.

But the real question is, with your monitor gain set to its normal
position and with that FS 1 kHz tone playing, did you blow out your
speakers? Could you stay in the same room?

How many analog volts come out when playing the FS tone?

--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me he double-m-eleven-double-zero at yahoo

"Mark & Mary Ann Weiss" wrote in message
k.net...

Have you measured signal-to-noise? How much signal do you get out for
a full scale digital recording? Maybe all you need to do is turn your
playback gain down about 20 dB. Sorry if this is too obvious, but I
just pop in and out of this thread and I don't know how much more you
know (or don't) about the gain staging.

Well this is interesting...

I generated a 0dB sine wave in SoundForge and output it to the MotU
channels
1-2. The MotU meters are registering -6dB, not even -1dB, the last segment
before [overload]. So the Analog Outs are not even being driven to full
output. Very curious indeed.... now wondering if these meters respond the
same as a broadcast transmitter modulation monitor, that being, you can
never get 100% modulation with a sinusoidal wave...
Nope... just tried a square wave and got the same -6dB reading on the
Analog
Out meters.
Quite a mystery.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-



Are the Motu meters properly calibrated?

Mikey
Nova Music Productions

Hmmmm
I would just send it in. Get it looked at for $45. If it comes back and
they say "no, that's how it is", put it on ebay and get something which
doesn't bug you so much.
If it comes back and there was a problem, it will be solved.


It is tempting to send it in, but since the hiss is on the playback end and
doesn't appear on the SACDs that I produce, it's not a serious issue. It's
more of a curiousity than anything else as to why a piece of gear with such
emphasis on quiet preamps would have so much noisier output amps.
Given the results that RightMark analyzer reported, I suppose I have to
conclude that there is nothing wrong with the unit.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

Are the Motu meters properly calibrated?

Mikey
Nova Music Productions



I believe so. Why? Because, after discovering a volume control app in
Windows that could be set to control the MotU outputs and setting it to full
scale, the meters now display full scale with a 0dB digital sine wave from
SoundForge. So I am satisfied that the output level is correct.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

"Mark & Mary Ann Weiss" wrote in
message
ink.net
Are the Motu meters properly calibrated?

Mikey
Nova Music Productions



I believe so. Why? Because, after discovering a volume
control app in Windows that could be set to control the
MotU outputs and setting it to full scale, the meters now
display full scale with a 0dB digital sine wave from
SoundForge. So I am satisfied that the output level is
correct.

What effect does this change have on the observed signal and
noise measurements, both with the meter and also with RMA55?

I generated a 0dB sine wave in SoundForge and output it to the MotU
channels
1-2. The MotU meters are registering -6dB, not even -1dB, the last
segment
before [overload]. So the Analog Outs are not even being driven to full
output. Very curious indeed....

This may be only a difference in interpretation of how 0 dBFS is
calibrated. I think the that current version of the definition is that
the meters should read 0 when the peak of a sine wave is at full
scale, but the meters may be indicating something else. If they read
-3, that would be easier to explain. Maybe Sound Forge and MOTU are
each compensating in the same direction.

Nope. It's a senseless design flaw in Windows which causes the MotU output
levels on all 8 channels to default to -6dB, requiring that I open the
Control Panel, click on sound devices and select each pair of MotU channels,
one at a time, and open the Volume Control app for each pair and set the
volume from half back to maximum. Then it outputs 0dB.

Nope... just tried a square wave and got the same -6dB reading on the
Analog
Out meters.

OK, so they're reading the average value of a square wave. It must
have something to do with the response time.

But the real question is, with your monitor gain set to its normal
position and with that FS 1 kHz tone playing, did you blow out your
speakers? Could you stay in the same room?

Nope. I have the studio monitors turned WAY down, for it is like the whine
of a 50HP siren if they are up too loud. I read the levels on the board.
Musically speaking, I CAN hear hiss when the house system is set to a level
where, let's say, rock music peaks at 118dB at listener position. I'd
estimate the hiss is producing an SPL of 25dB under those conditions.


How many analog volts come out when playing the FS tone?

Under this condition, the RMS voltmeter reads:

+18dB
6.25vrms

Incidentally, my K2600RS reads:

+21dB
8.7vrms



--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

"Mark & Mary Ann Weiss" wrote in
message
nk.net

Nope. It's a senseless design flaw in Windows which
causes the MotU output levels on all 8 channels to
default to -6dB, requiring that I open the Control Panel,
click on sound devices and select each pair of MotU
channels, one at a time, and open the Volume Control app
for each pair and set the volume from half back to
maximum. Then it outputs 0dB.

The default settings for the windows volume control are set
by the device driver installation process. IOW, blame MOTU.

Nope... just tried a square wave and got the same -6dB
reading on the Analog Out meters.

OK, so they're reading the average value of a square
wave. It must have something to do with the response
time.

But the real question is, with your monitor gain set to
its normal position and with that FS 1 kHz tone playing,
did you blow out your speakers? Could you stay in the
same room?

Nope. I have the studio monitors turned WAY down, for it
is like the whine of a 50HP siren if they are up too
loud. I read the levels on the board. Musically speaking,
I CAN hear hiss when the house system is set to a level
where, let's say, rock music peaks at 118dB at listener
position. I'd estimate the hiss is producing an SPL of
25dB under those conditions.

If you hear a 25 dB SPL noise in a room, that's really a
pretty fine room.

I suspect that reality might be a bit more like the 30s.

93 dB dynamic range which is what your example works out to,
is really pretty fair dynamic range for an audio interface.
I suspect that now that you found another 6 dB of output,
you'll be back to the 98 dB dynamic range that RMAA found.

If you want appreciably better dynamic range, rip out that
MOTU and put in a pair of LynxTWOs. They will cost you
almost $2k, but they will be quieter. They slave together
very nicely to provide 8 channels.

How many analog volts come out when playing the FS tone?

Under this condition, the RMS voltmeter reads:

+18dB
6.25vrms

Very typical for a audio production computer audio
interface.

Incidentally, my K2600RS reads:

+21dB 8.7vrms

Pretty typical, as well.

The default settings for the windows volume control are set
by the device driver installation process. IOW, blame MOTU.

I'll have to register a software update request/feature change with them, if
that's the case. I shouldn't have to set the levels every time I turn on the
audio interface. They should remain where I left them last session.


Nope. I have the studio monitors turned WAY down, for it
is like the whine of a 50HP siren if they are up too
loud. I read the levels on the board. Musically speaking,
I CAN hear hiss when the house system is set to a level
where, let's say, rock music peaks at 118dB at listener
position. I'd estimate the hiss is producing an SPL of
25dB under those conditions.

If you hear a 25 dB SPL noise in a room, that's really a
pretty fine room.

We're in rural country. With the computers off, all you hear is the blood
rushing through your eardrums. It is DEAD quiet in here. Our custom-built
10kW sound system has a dynamic range better than 130dB and even the 17dB
hum/hiss from the amplifiers is audible, if within 3' of the array. Not bad
though, for a system capable of 800 acoustical watts. We spent a lot of time
last winter, tuning it for quietest noise floor through gain staging and
redressing of wiring (using a mic planted in front of each driver, one at a
time, driving a headphone amp and listening on headphones to the noise
floor, amplified so that one could hear changes as adjustments were made to
the wiring.) Prior to that, for the longest time, the system emitted a 40dB
hum.

I suspect that reality might be a bit more like the 30s.

93 dB dynamic range which is what your example works out to,
is really pretty fair dynamic range for an audio interface.
I suspect that now that you found another 6 dB of output,
you'll be back to the 98 dB dynamic range that RMAA found.

What I find curious then is why my Denon CD player is quieter *with my
recordings* when I play it from the CD. It seems that when I need to hear
the cleanest reproduction, I have to burn a CD and go play it in the Denon.
Same SPL with music, but no hiss during quiet passages, for instance, when
the conductor raises his baton, just a second before the orchestra plays the
first note is often the quietest point in my recordings. Played on the MotU,
there's a small amount of hiss audible. Take the digital file, transfer to a
CD and play on the Denon and it's all there, but without the hiss. Same SPL
at crescendo levels. Very odd indeed.


If you want appreciably better dynamic range, rip out that
MOTU and put in a pair of LynxTWOs. They will cost you
almost $2k, but they will be quieter. They slave together
very nicely to provide 8 channels.

It's not worth it, since we only care about the record side. The playback
side is more of a curiousity, rather than a problem that we can't deal with.
Since it does what we need with a laptop PC in the field, it's getting a
fine job done. It was only recently that we even bothered to hook the MotU
outputs up to our mixing desk. And that's when everyone said "hey, this box
is noisy!"


How many analog volts come out when playing the FS tone?

Under this condition, the RMS voltmeter reads:

+18dB
6.25vrms

Very typical for a audio production computer audio
interface.

Incidentally, my K2600RS reads:

+21dB 8.7vrms

Pretty typical, as well.

I used to design broadcast equipment whose audio would swing 12vrms. It
seems that these devices are running at a much lower rail voltage than I
estimated, or they are not making use of the full dynamic range of the
op-amps employed.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

"Mark & Mary Ann Weiss" wrote in
message
k.net

The default settings for the windows volume control are
set by the device driver installation process. IOW,
blame MOTU.

I'll have to register a software update request/feature
change with them, if that's the case. I shouldn't have to
set the levels every time I turn on the audio interface.

Agreed. The usual deal is to remember the last settings.

They should remain where I left them last session.

Agreed.

Nope. I have the studio monitors turned WAY down, for it
is like the whine of a 50HP siren if they are up too
loud. I read the levels on the board. Musically
speaking, I CAN hear hiss when the house system is set
to a level where, let's say, rock music peaks at 118dB
at listener position. I'd estimate the hiss is
producing an SPL of 25dB under those conditions.

If you hear a 25 dB SPL noise in a room, that's really a
pretty fine room.

We're in rural country. With the computers off, all you
hear is the blood rushing through your eardrums. It is
DEAD quiet in here.

Lucky you.

Our custom-built 10kW sound system
has a dynamic range better than 130dB and even the 17dB
hum/hiss from the amplifiers is audible, if within 3' of
the array.

Well that's really wonderful, but with such a remarkable
tool, you still have to fit in with the rest of the world.

Not bad though, for a system capable of 800
acoustical watts. We spent a lot of time last winter,
tuning it for quietest noise floor through gain staging
and redressing of wiring (using a mic planted in front of
each driver, one at a time, driving a headphone amp and
listening on headphones to the noise floor, amplified so
that one could hear changes as adjustments were made to
the wiring.) Prior to that, for the longest time, the
system emitted a 40dB hum.

Ironically, 40 dB of hum in a system that is capable of 130
dB is 90 dB dynamic range which is generally really pretty
good.


93 dB dynamic range which is what your example works out
to, is really pretty fair dynamic range for an audio
interface. I suspect that now that you found another 6
dB of output, you'll be back to the 98 dB dynamic range
that RMAA found.

What I find curious then is why my Denon CD player is
quieter *with my recordings* when I play it from the CD.

The laws of physics say that 16 bits can't do much more than
96-98 dB dynamic range, and on a good day a decent 24 bit
audio interface should be able to do that.

It seems that when I need to hear the cleanest
reproduction, I have to burn a CD and go play it in the
Denon. Same SPL with music, but no hiss during quiet
passages, for instance, when the conductor raises his
baton, just a second before the orchestra plays the first
note is often the quietest point in my recordings. Played
on the MotU, there's a small amount of hiss audible. Take
the digital file, transfer to a CD and play on the Denon
and it's all there, but without the hiss. Same SPL at
crescendo levels. Very odd indeed.

I think you need to look at gain staging from the MOTU. Like
I said, we just found a *free* 6 dB worth of dynamic range
in the windows mixerf, and there might more lying around,
hiding some place.

If you want appreciably better dynamic range, rip out
that MOTU and put in a pair of LynxTWOs. They will cost
you almost $2k, but they will be quieter. They slave
together very nicely to provide 8 channels.

It's not worth it, since we only care about the record
side.

Thatr begs the question why all the worrying about the Motu?
Obviously you care about it, at least a little.

The playback side is more of a curiousity, rather
than a problem that we can't deal with. Since it does
what we need with a laptop PC in the field, it's getting
a fine job done. It was only recently that we even
bothered to hook the MotU outputs up to our mixing desk.
And that's when everyone said "hey, this box is noisy!"

How many analog volts come out when playing the FS
tone?

Under this condition, the RMS voltmeter reads:

+18dB
6.25vrms

Very typical for a audio production computer audio
interface.

Incidentally, my K2600RS reads:

+21dB 8.7vrms

Pretty typical, as well.

I used to design broadcast equipment whose audio would
swing 12vrms. It seems that these devices are running at
a much lower rail voltage than I estimated, or they are
not making use of the full dynamic range of the op-amps
employed.

I think you already did the math for +15/-15 power supplies.
Most op amps are ok with +18/-18, and the OPA604 and 2604
are rated for more like +/- 24.

However, maximum outputs in the 6-7 volt range are pretty
typical for computer audio interfaces.

Mark & Mary Ann Weiss wrote:

I used to design broadcast equipment whose audio would swing 12vrms. It
seems that these devices are running at a much lower rail voltage than I
estimated, or they are not making use of the full dynamic range of the
op-amps employed.

Sadly this is typical today. Everyone is obsessed with lower rail voltages,
lower power consumption, and going to unipolar supplies to save power supply
costs.

My console, on the other hand, can swing more than 70V output, from the
+/-48V rails.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."

I believe so. Why? Because, after discovering a volume
control app in Windows that could be set to control the
MotU outputs and setting it to full scale, the meters now
display full scale with a 0dB digital sine wave from
SoundForge. So I am satisfied that the output level is
correct.

What effect does this change have on the observed signal and
noise measurements, both with the meter and also with RMA55?



Didn't retest RMA55, but if that test suite doesn't have control of the
volume (override), then maybe I should retest with that program and see if
the results increase by 6dB.
It has no effect on the static noise level at the outputs however, as
measured with a H/P 400E.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

Agreed. The usual deal is to remember the last settings.

They should remain where I left them last session.

Agreed.

Since that is what every other sound device I have in our DAWs
does--remember the settings.


We're in rural country. With the computers off, all you
hear is the blood rushing through your eardrums. It is
DEAD quiet in here.

Lucky you.

No. A matter of choice. It was initially cheaper to come out here 40 years
ago.


Our custom-built 10kW sound system
has a dynamic range better than 130dB and even the 17dB
hum/hiss from the amplifiers is audible, if within 3' of
the array.

Well that's really wonderful, but with such a remarkable
tool, you still have to fit in with the rest of the world.

That's another reason why we're out here, and not someplace with neighbors.
They don't call me "Bass Pig for nothing. :-)


Not bad though, for a system capable of 800
acoustical watts. We spent a lot of time last winter,
tuning it for quietest noise floor through gain staging
and redressing of wiring (using a mic planted in front of
each driver, one at a time, driving a headphone amp and
listening on headphones to the noise floor, amplified so
that one could hear changes as adjustments were made to
the wiring.) Prior to that, for the longest time, the
system emitted a 40dB hum.

Ironically, 40 dB of hum in a system that is capable of 130
dB is 90 dB dynamic range which is generally really pretty
good.

Yes, the numbers look good, but when you're trying to listen to a soft
pianissimo passage in, say, Stravinsky's Firebird Suite, that hum and hiss
are pretty distracting.


93 dB dynamic range which is what your example works out
to, is really pretty fair dynamic range for an audio
interface. I suspect that now that you found another 6
dB of output, you'll be back to the 98 dB dynamic range
that RMAA found.

What I find curious then is why my Denon CD player is
quieter *with my recordings* when I play it from the CD.

The laws of physics say that 16 bits can't do much more than
96-98 dB dynamic range, and on a good day a decent 24 bit
audio interface should be able to do that.

I'm fully aware of these laws, and that's why I find it a matter of academic
curiousity why my 24-bit interface generates so much more noise than my
16-bit Denon CD player.




I think you need to look at gain staging from the MOTU. Like
I said, we just found a *free* 6 dB worth of dynamic range
in the windows mixerf, and there might more lying around,
hiding some place.

Things have gotten better with the additional 6dB, but I am hard--pressed to
find any remaining unused dynamic range. Any thoughts or suggestions would
be welcome.


It's not worth it, since we only care about the record
side.

Thatr begs the question why all the worrying about the Motu?
Obviously you care about it, at least a little.

Academic curiousity. When you buy a somewhat expensive audio interface and
go a year without ever critically listening to it's output (because you hear
the results only through other audio playback D/A systems, and one day you
connect it to your board and notice that it has about 20dB more hiss but is
not 20dB louder than your other devices on competing channels, you kinda'
wonder about that..


I used to design broadcast equipment whose audio would
swing 12vrms. It seems that these devices are running at
a much lower rail voltage than I estimated, or they are
not making use of the full dynamic range of the op-amps
employed.

I think you already did the math for +15/-15 power supplies.
Most op amps are ok with +18/-18, and the OPA604 and 2604
are rated for more like +/- 24.

However, maximum outputs in the 6-7 volt range are pretty
typical for computer audio interfaces.

And PC sound cards max out at less than 2vrms.
If the hiss were proportionately smaller, the s/n ratio should remain
reasonably good.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

I used to design broadcast equipment whose audio would swing 12vrms. It
seems that these devices are running at a much lower rail voltage than I
estimated, or they are not making use of the full dynamic range of the
op-amps employed.

Sadly this is typical today. Everyone is obsessed with lower rail
voltages,
lower power consumption, and going to unipolar supplies to save power
supply
costs.

My console, on the other hand, can swing more than 70V output, from the
+/-48V rails.
--scott


You mean this is all because of the 'green' movement?
It seems a small thing to ask of MotU, whose main power is consumed by the
digital electronics, not the op amps.


--
Best Regards,

Mark A. Weiss, P.E.
www.mwcomms.com
-

"Mark & Mary Ann Weiss" wrote in
message
nk.net

93 dB dynamic range which is what your example works
out to, is really pretty fair dynamic range for an
audio interface. I suspect that now that you found
another 6 dB of output, you'll be back to the 98 dB
dynamic range that RMAA found.

What I find curious then is why my Denon CD player is
quieter *with my recordings* when I play it from the CD.

Compared to the Motu, its maximum output is more like 2
volts than 7. That is another approx 10 dB differnce.

The laws of physics say that 16 bits can't do much more
than 96-98 dB dynamic range, and on a good day a decent
24 bit audio interface should be able to do that.

I'm fully aware of these laws, and that's why I find it a
matter of academic curiousity why my 24-bit interface
generates so much more noise than my 16-bit Denon CD
player.

It looks to me like the Motu isn't enough better than the
Denon in terms of actual dynamic range, to overcome the
approx 10 dB difference in maximum output.

I think you already did the math for +15/-15 power
supplies. Most op amps are ok with +18/-18, and the
OPA604 and 2604 are rated for more like +/- 24.

However, maximum outputs in the 6-7 volt range are pretty
typical for computer audio interfaces.

And PC sound cards max out at less than 2vrms.
If the hiss were proportionately smaller, the s/n ratio
should remain reasonably good.

It isn't. IME the very best consumer sound cards (example
M-Audio Revolution 7.1) have a little worse dynamic range
than the Motu.

"Mark & Mary Ann Weiss" wrote in
message
ink.net
I believe so. Why? Because, after discovering a volume
control app in Windows that could be set to control the
MotU outputs and setting it to full scale, the meters
now display full scale with a 0dB digital sine wave from
SoundForge. So I am satisfied that the output level is
correct.

What effect does this change have on the observed signal
and noise measurements, both with the meter and also
with RMA55?



Didn't retest RMA55, but if that test suite doesn't have
control of the volume (override),

It doesn't.

then maybe I should
retest with that program and see if the results increase
by 6dB.

Agreed.

It has no effect on the static noise level at the outputs
however, as measured with a H/P 400E.

That would be a good thing. This time we're trying to
improve dynamic range by increasing the maximum output.

I'm wondering if the .wav files you play through the Motu
have good, high peak levels (i.e., within a dB or so of FS)