Hey, Group-

I haven't posted here in eons, but am doing some consulting / setup
work for a voice talent, and so far am perplexed by this problem, which
has been exhibited by 3 portable PC audio interfaces from 2 different
manufacturers so far...

In each case, the portable mic pre interface (M Audio MobilePre USB and
FastTrack, and Focusrite Saffire) receives maximum level from the mic.
The interface's software shows that we're hitting 0dBFS, and it sounds
that way. But our recording software shows the signal level to be
-5.5dBFS (Adobe Audition) or -6dBFS (Sound Forge). An attempt to
increase the gain simply slams the interface harder into distortion,
and the recording software never gets above the -5.5 or -6 mark,
appearing clipped at that point. A glance at the waveform (and a
listen to it!) confirms that it really did record at the lower level,
not all the way up to 0 as the audio interfaces indicated.

Strange thing #1... M Audio tech said this was a known issue with the
Mobile Pre USB and FastTrack, but it's also happening with the Saffire.

Strange thing #2... Only tested with the Saffire so far, but Cubase LE
shows the levels just fine! AudioTX ("voiceover over IP" software)
displayed and passed the levels accurately, too.

So for some reason, on my client's laptop using both Sound Forge and
Adobe Audition, the levels get reduced by either 5.5 or 6 dBFS.

My next step is obviously to move the whole setup onto another computer
and try it there, to see if maybe we need to do a Windows update, or a
fresh install. We'll also be talking to Sony and Adobe tech support.
But I wanted to float this out there in case anyone has been here
before.

Has anyone ever experienced anything like this?

Thanks for any help you may have!

-Mark

Meters are different. Some read peak level, others read RMS or some
other average level. Some are intentionally calibrated so that if you
heed what they're telling you you won't clip.

Don't change computers or worry about a Windows update - that won't
solve this perceived problem. Learn how to interpret your meters, and
use them as a guide, not as a lab test to see how you or your equpment
is doing. And don't worry about being 5 dB lower than full scale.
That's easy to fix. Clipping isn't easy to fix.

Mike, thanks for the response! I've enjoyed reading your posts here
for years, and have come to greatly respect your opinion.

I understand that clipping is harder to fix than a low level. I also
get the concept of peak vs. RMS and other "average level" metering, but
it doesn't seem to me like that's what's happening. If I increase the
output of the audio interface, it just slams up against 0 harder, and
the recording software slams harder up against its "adjusted 0" of -5.5
or -6. If it were simply a matter of my recording software showing
RMS, I would expect to see some peaks in the resulting waveform going
higher and not getting brickwall limited at -5.5 or -6 as if it were 0.

What's bugging me is that one program (Cubase LE) shows the audio going
all the way up to 0 when the audio interface shows it, too. Why don't
the others? One thing I haven't done yet is to record a file in Cubase
LE, then open it in Sound Forge to see if the waveform displays the
same level as Cubase. If it does, then maybe Cubase really is
recording it at the proper level and the others are somehow digitally
attenuating it. If it appears in Sound Forge at a different level than
in Cubase, then it's just a display issue, right?

But if it is, yikes! These programs have meters with numbers on them.
They all appear to be referencing 0dBFS (or can be set to do so)... and
if one says I'm at -3.8dBFS and another -9.8dBFS, then some are right
and some are wrong... right?

Please forgive what may seem like a tirade-like cry for help out of
naivete or ignorance, but I just want the best for my client, and am
sincerely having a hard time seeing how this could really just be no
problem. Of course I can show him how to maximize or normalize after
the fact. I just hate the idea that he might not be getting the
resolution he could be getting.

If you think it's best to just let this go, I'll seriously consider it.
It's just that I always thought I could trust my software's digital
meters to display -2.3 dBFS if that's where the audio actually is at
that moment.

If I'm missing a huge, obvious point, please feel free to ram it down
my throat.

mjpatey wrote:

What's bugging me is that one program (Cubase LE) shows the audio going
all the way up to 0 when the audio interface shows it, too. Why don't
the others? One thing I haven't done yet is to record a file in Cubase
LE, then open it in Sound Forge to see if the waveform displays the
same level as Cubase.

I was going to suggest something like that - use a common reference.
When you say "shows it" are you talking about a meter indication or how
far the waveform graphic goes toward the limits? A recording at -6 dBFS
only fills up half the waveform space and that's pretty obvious.

If you think it's best to just let this go, I'll seriously consider it.
It's just that I always thought I could trust my software's digital
meters to display -2.3 dBFS if that's where the audio actually is at
that moment.

I think it would be interesting to figure out what's going on so you
know know how to interpret the meter readings on whatever program you
happen to be recording on (that's really the only time you need to
worry about it) but I also don't think you should agonize over it.
What's best for your client is that he knows that he's not going to try
to overdrive the A/D converters, or, equally important, clip the mic
preamp or analog side of the A/D converter.

You should check your drivers and their control panels carefully. It's
possible that there's some digital attenuation set on the input that
you didn't realize. That could explain everything.

Mike,

Thanks again. When I've referred to levels in the recording apps, I
mean the actual size of the waveform, and the reading alongside it in
dBFS. The -6dB limit is extremely obvious visually, I agree.

I have tried looking through control panels, but have found no option
that would cause the discrepancy I'm getting. I even checked the
Windows Mixer, but its "record level" side is disabled by the audio
interfaces' control software.

I've spent some time on the phone with Focusrite's USA tech support
(who, btw, are WONDERFUL), and they have never experienced what I'm
describing in the apps they use (though no word from anyone using the 2
apps I'm having trouble with!)... which is why I'm inclined to think
it's something in the way the OS on this computer is passing data from
the audio interface to the recording apps. It's also why I want to try
installing everything on another system, and see if I can replicate the
phenomenon.

My next step is a call to Sony and Adobe, just so see if they have any
clue.

If you can think of any other things to try, I'm all ears!

Thanks again for your help.

-Mark

burn the wav file to a CD and check it on a calibrated CD player...

Mark

mjpatey wrote:

I've spent some time on the phone with Focusrite's USA tech support
(who, btw, are WONDERFUL), and they have never experienced what I'm
describing in the apps they use (though no word from anyone using the 2
apps I'm having trouble with!)...

Focusrite doesn't have experience with their products used with Sound
Forge and Audition? Oh, well, they're an Avid company now, and I guess
they only bother to test their hardware with ProTools.

I really don't have anything else to say about this other than to quit
worrying about it and start making music.

Mark wrote:
burn the wav file to a CD and check it on a calibrated CD player...

Can you name a calibrated CD player? I can't.

Mike Rivers wrote:

Mark wrote:
burn the wav file to a CD and check it on a calibrated CD player...

Can you name a calibrated CD player? I can't.

I thought it sounded like an odd comment at the time. Don't most play at
roughly the same level to their fixed output sockets ? 2V and all that.

Graham

"Mike Rivers" wrote in news:1136033477.283701.33110
@g14g2000cwa.googlegroups.com:


mjpatey wrote:

I've spent some time on the phone with Focusrite's USA tech support
(who, btw, are WONDERFUL), and they have never experienced what I'm
describing in the apps they use (though no word from anyone using the 2
apps I'm having trouble with!)...

Focusrite doesn't have experience with their products used with Sound
Forge and Audition? Oh, well, they're an Avid company now, and I guess
they only bother to test their hardware with ProTools.

I really don't have anything else to say about this other than to quit
worrying about it and start making music.



I've been trying to find out where I'd heard this problem before but can't
seem to find the answer (or rather the correct search string) in google
groups. It came up on rec.audio.pro about a year ago & I think it was in
relation to a MOTU interface, maybe an 896.

I seem to remember the problem was put down to the wdm drivers being in
stereo pairs and applying some kind of pan law to the inputs when recording
to a mono file. Thus when you hit full scale on the converters, the
recording software doesn't recieve full scale, rather -6db or there abouts.
The ASIO drivers don't exhibit this problem.

Maybe try recording a stereo file in wavelab/audition and see what happens
then.

Pooh Bear wrote:
Mike Rivers wrote:

Can you name a calibrated CD player? I can't.

I thought it sounded like an odd comment at the time. Don't most play at
roughly the same level to their fixed output sockets ? 2V and all that.

Maybe, maybe not, but "roughly" isn't calibrated.

Hi Mark,

Audition has correctly calibrated meters, so that is not the problem. The
fact that there is a 6db error indicates that there is a scaling problem at
the driver level. 6db equals 1 bit in the digital world. Sounds to me like
the driver packs the data into the 16 bit word incorrectly. Have you tried
recording at different bit resolutions (32,24)? I believe Cubase is working
properly because it is using ASIO drivers for your card rather than windows
MME or WDM. I would also try alternate drivers in windows. For example, if
you are testing with WDM, try the older MME drivers. It would also be
interesting to see what happens in Cubase if you select the generic MME or
DirectX drivers. Finally, is it possible that there is mixer applet for your
audio card that has a record level adjustment? This applet might only
affect the windows driver not the ASIO.

Mark R.

"mjpatey" wrote in message
ups.com...
Hey, Group-

I haven't posted here in eons, but am doing some consulting / setup
work for a voice talent, and so far am perplexed by this problem, which
has been exhibited by 3 portable PC audio interfaces from 2 different
manufacturers so far...

In each case, the portable mic pre interface (M Audio MobilePre USB and
FastTrack, and Focusrite Saffire) receives maximum level from the mic.
The interface's software shows that we're hitting 0dBFS, and it sounds
that way. But our recording software shows the signal level to be
-5.5dBFS (Adobe Audition) or -6dBFS (Sound Forge). An attempt to
increase the gain simply slams the interface harder into distortion,
and the recording software never gets above the -5.5 or -6 mark,
appearing clipped at that point. A glance at the waveform (and a
listen to it!) confirms that it really did record at the lower level,
not all the way up to 0 as the audio interfaces indicated.

Strange thing #1... M Audio tech said this was a known issue with the
Mobile Pre USB and FastTrack, but it's also happening with the Saffire.

Strange thing #2... Only tested with the Saffire so far, but Cubase LE
shows the levels just fine! AudioTX ("voiceover over IP" software)
displayed and passed the levels accurately, too.

So for some reason, on my client's laptop using both Sound Forge and
Adobe Audition, the levels get reduced by either 5.5 or 6 dBFS.

My next step is obviously to move the whole setup onto another computer
and try it there, to see if maybe we need to do a Windows update, or a
fresh install. We'll also be talking to Sony and Adobe tech support.
But I wanted to float this out there in case anyone has been here
before.

Has anyone ever experienced anything like this?

Thanks for any help you may have!

-Mark

"mjpatey" wrote:

Please forgive what may seem like a tirade-like cry for help out of
naivete or ignorance


I don't think your remarks are being perceived that way. It's more that
there are two separate issues here.

The first is that you're not getting the same feedback from app to app
and interface to interface. I would be annoyed and confused by that
too. The cause of the discrepancy is the issue.

That's separate from the second issue, which is the affect on the
resulting recording. In that light, it doesn't really matter. I track
with 0VU referenced to -20dBFS. The resolution of even a 16 bit system
(never mind 20 or 24) is so high that risking damage by tracking up near
the top is not only unnecessary but irresponsible. You wouldn't dangle
your client's instrument by the cable; neither should you so recklessly
risk their recording.

So, while I don't know what's causing your weird readings either, I
wouldn't worry too much. Turn down the input and get some work done.

--
"It CAN'T be too loud... some of the red lights aren't even on yet!"
- Lorin David Schultz
in the control room
making even bad news sound good

(Remove spamblock to reply)

"Pooh Bear" wrote:

Don't most play at roughly the same level to their fixed output
sockets ? 2V and all that.


My Tascam CD601 had internal trim pots for analog output level that had
about a 20dB range. Of course, it also had a dedicated controller with
scrub wheel and RAM buffers and all manner of other broadcast related
features...

--
"It CAN'T be too loud... some of the red lights aren't even on yet!"
- Lorin David Schultz
in the control room
making even bad news sound good

(Remove spamblock to reply)