Why does Audition impose a -6 dB limit on mono files?

Maybe I'm missing something fundamental about the nature of audio files here, but I've noticed some strange behavior when I record to a single mono track in Edit mode, OR convert a two-channel file to mono, OR bounce a stereo track in Multitrack mode.

Here's what happens . . .

1. When I record to a mono file (any sampling rate), the signal is hard limited at -6 dB in either Edit or Multitrack mode -- cut off cleanly at -6 dB. This has nothing to do with outboard gear; I've swapped it out and it doesn't make a difference. But when recording the same signal to stereo tracks, the signal is NOT limited -- it takes up the entire amplitude range from infinity to 0 dB, as it should.

2. If I convert any true stereo file to mono (say, ripped from a CD), there is no problem with the levels. But when the source is a two mono signals in the left and right channels (let's say the host and guest audio on a talk show), they record properly, but the -6 dB limit DOES happen when converting to Mono (in Edit mode). It occurs both in Audition and also in another program I tried. Why would the converted mono signal not have a complete amplitude range?

3. Lastly, when a stereo track is bounced to a new track in Multitrack mode, it's limited at -3 dB -- inconsistent with the above behavior and still not full-range.

Does anyone know what's going on?

Is this a settings issue? Something else?

This is probably the most frequently asked question these days on the Adobe site but I'm not convinced that the answer has really been pinned down - otherwise there would be an FAQ for it.

OK, that's a way of saying "would someone please post something here or there that I can make an FAQ out of?"

I'll see if I can figure out what's really happening here later. Bit busy for a few hours...

Great!

Let me know if I can help with any testing . . . this question does seem to come up, but I have not found a definitive answer so far. Yes, an FAQ is probably in order.

I doubt if it's a bug in Audition, more likely to be a "behavior" based on a technical phenomenon.

Hi all, I've been a CEP and AA user since way before there was "Pro" in the name. It is in fact my primary audio editor and recorder. I have just upgraded to 2.0 because I was of the opinion that it wasn't broken so it didn't neeed fixin'. Yes ASIO and a windowed functionality is nice but my workflow isn't really changed much.

I  thought I would post something here to bring this back to the top. I just had a very frustrating experience with this exact issue. And I'm curious if you have had a chance to look at this Steve.

I can confirm that Edit view does impose a 6dB loss when recording a mono file but not a stereo one. I have tested this with a Sound devices USB Pre using the WDM driver and also with a MOTU Traveller using the ASIO driver. I can say that both devices will record a mono track fine in Multritrack view, and in Tracktion 2.whatever. Because I have metering on both of my devices I injected a a 1kHz tone into both devices and raised the input gain until the outputs were clipping. The mono file recorded in Edit view was beautifully clipped but the max level was at -6dB. When this file was saved and closed it could be opened in Tracktion but was a clipped file at -6dB. This was also the case when the file was opened on another machine in Pro Tools. It actually looks like in Edit view when recording a Mono track AA is dropping a bit or something.

It is certainly possible to do all my recording in MT view but that is moderately annoying when all I needed was a quick 30 second mono 44.1kHz / 16 bit voice over. This may be and "Undocumented Feature" but it looks more like a bug.

Any headway explaining this?

Thanks,

Brian

Any headway explaining this?

Slight problem - the machine I normally do this sort of investigation on has only just come back recently from H-P, and I have to reload all the software from scratch - so none of the things I was going to do (including this one) are up to date at all.

I'm getting there gradually...

Thanks for keeping this question alive . . . I'm sure many users would like to know the answer. So far, I've just been living with it, not really knowing if the -6 dB liimit is somehow intentional (a necessary technical limit),  or a bug.

Just wanted to keep this topic alive.

It actually looks like in Edit view when recording a Mono track AA is dropping a bit or something.

Any headway explaining this?

A little - although this may cause more confusion/consternation, at least initially. I decided in the end, because of the nature of the problem to use the entirely calibrated E-Mu system to check this out - because it can generate tones to precisely specified levels, and more importantly, generate them internally at more than 0dB in the box. In case you are not aware, the E-Mu 1820M soundcard with breakout box is basically a large, complex virtual mixer that does some neat tricks entirely in the PC, and turns out to be a good way to check out at least some of the reported bahaviour...

Initially, I found that yes, in EV, Audition is indeed under-recording by 6dB the tone that I set up to come out of the channel, and that I applied directly to the input in EV. But, and please note this, only when the Audition Windows Driver is used.

I checked the same behaviour by switching to another input on the E-Mu with the same tone level applied, and resetting the Audition driver to its ASIO mode. Now the results were perfect - everything records at the correct levels, mono and stereo. Please note that this is without stopping Audition and reloading it, or doing anything at all to the E-Mu CueMix - I've just changed the input assignment in Audition.

I'm not sure how this helps though - I'm not even sure that there's anything wrong with the driver per se. I tried the same thing using the standard E-mu Windows driver using AA1.5 and got the same under-reading - this appears to lose 6dB as well. And it really does appear to be an error - if you increase the output by 6dB in the mixer, then both versions record at 0dB perfectly in mono.

What's slightly confusing is what happens when you record in stereo using the Audition Windows Driver in AA2.0, but use a mono source. If you swing the panpot on the mixer, you can pan the mono channel between the two stereo channels being recorded - and end up with a 6dB drop in the centre, but a 0dB signal in one channel when the panpot is swung right over to the other side. AFAIK this should only drop 3dB in the centre, not 6dB - unless a very strange law is being used! It should only be 3dB because it's signal levels we are shifting - if we double the intensity in the centre pan position, then the SPL will increase by 3dB. The implemented version appears to be suggesting that we have doubled the SPL in the centre and got a 6dB increase - which we haven't done, of course.

The ASIO driver behaves completely differently - no amount of manipulation of the pan control alters anything to do with the two input channels, and even recording in mono makes no difference.

There appears to be no distortion of any kind - if you get the E-Mu to produce tone at +6dB Audition records without distortion in EV using the Audition Windows Driver - everything behaves as it should, certainly with the E-Mu system anyway.

So, my initial question is "what driver have you been using?" Regardless of the answer though, I think that some more investigation is required...

Maybe a test using ASIO4all would be interesting, too.

Paul

maybe a test using ASIO4all would be interesting, too.

I think so too - but it's presently not installed on that system. I will install it, and try again later. What I hope would happen is that it mimics the ASIO results - which is of course what it should do... but hey, some of this behaviour is strange, so all bets are off until I've tried it.

When I first found this issue I tested with a USB Pre from Sound devices and a MOTU Traveller. For compatability with another program, I do not have the ASIO driver for the USB Pre installed, so it was using the Audition Windows driver. I thought when I tested with the Traveller I was fairly certain I was using the ASIO driver. I will also revisit this this week. Thanks for looking into this Steve.

I'll look at it with the Traveler as well - I've installed the driver for it on the test rig temporarily, so that's not going to be a problem at all. It will be interesting to see if there are any differerences in perfromance...

Whatever happens, this whole scenario doesn't seem to be quite the way it should be, so the developers will get an extended note about it anyway.

I'm experiencing the same thing with an M-Audio Revolution 7.1

Here are screenshots of the phenomenon (versions 1.5 and 2.0, respectively).

http://www.finalsmoke.com/audition/aa15.gif

http://www.finalsmoke.com/audition/aa20.gif

Hopefully this can be solved soon.