Topic: WAV - squared off - what happened here?

[vanark]

I got a recording from a taper (the raw WAV file, nothing done to it) and started to master it today.  I immediately noticed that the WAV was squared off through much of the recording, but it is not clipping - it is ~8 dB below zero.  When I listen, there is no obvious distortion.  I'm wondering what may have happened during the recording to cause this.  All I know about the equipment is Soundfield ST250 > AV340.  Here is a link to a 1 min. mp3 sample at the beginning which coincides with the screenshot.

http://www.sendspace.com/file/5yt632

[bgalizio]

Zoom in all the way. Is it still squared off then? If so, it's probably brickwalling.

[vanark]

Does brickwalling always cause distortion? This "sounds" okay.

Here are two more screenshots.  The first is zoomed all the way out so you can see the effect on the entire recording.  The second is zoomed in on a 5 sec. segment so you can see it squared off.

[Petrus]

Limiter?

[Chilly Brioschi]

If there is no audible distortion, then that is compression (soft clipping) , and lots of it!

Here's a great intro to limiting / compression:
http://en.wikipedia.org/wiki/Audio_level_compression
The article is a Wiki, and not 100% accurate.
Probably due to the differences in terminology within the various parts of the audio and electronics industry

What was the source?  SF > Archos only???
Was it a recording made with the with ALC?

Was this thing in the signal chain?!!!:

               The "Holy Grail" Fairchild 670

[intpseeker]

limiter or compression of a high volume track???

[vanark]

I'm not sure what else was in the signal chain.  I have a message out to the taper but don't expect to hear back until Wednesday. I see some notes from after the show where both bass and attentuation were set to "on" for the ST250.

[DSatz]

vanark, first off I'm sorry, but I don't have the answer to your question. The waveforms look to me as if some analog circuitry was clipping before the signal was fed into the A/D converters. If these are the non-encoded waveforms, then given the very slight unevenness of their peak amplitudes they don't look as if they were clipped digitally and then reduced in level, which sometimes occurs.

On the other hand if the waveforms in the pictures were taken from MP3 files, then all bets are off because the MP3 encoding process plays around with all sorts of things in a signal, and in that case, I doubt that any of us here can reliably tell you what caused this. You'll have to find out how the recording was made, and find the weak link in the chain (the circuit stage that overloads first) step by step. Let us know if you need help.

Your message said "... but it is not clipping - it is ~8 dB below zero." If, however, the microphones and/or the inputs to the microphone preamps were being pushed into overload, you could get waveforms like this even though the eventual peak recorded level was well below full scale. The record level control is typically situated in the circuit of a recorder (particularly if it has both microphone and line inputs) at a point which (in terms of signal flow) is "after" the recorder's internal mike preamps. Thus it merely controls the gain with which the preamps' output is fed into the A/D converter. If the preamps' output is already clipped because of input overload, you could record those clipped signals at low levels, medium levels, high levels--any levels you want, really--but they will still be clipped.

When you use a given type of microphone with a certain powering arrangement and you connect that to a given preamp or recorder, there is a maximum sound pressure level which is the highest that the combination can handle without overload. It isn't a digital thing--it's purely analog, and believe me, the problem was just as prevalent back in the analog era. At some certain sound pressure level either the microphone itself will begin to put out distorted signals because it can't handle the SPLs (perhaps as a result of improper powering and/or loading), or the preamp/recorder input will start to distort because the voltages from the microphone are too strong for it. The level at which the resulting waveforms are recorded is a completely separate issue in most cases.

--best regards

[vanark]

DSatz, thank you for the detailed response.  I think I understand a little more now.     I sure would like to figure this out so I can provide some feedback to the taper.

Here is a WAV sample - 15 sec.  In case anyone wants to take a look at the uncompressed file.

http://www.sendspace.com/file/ppadwg

+T's all around for the feedback so far.

[SmokinJoe]

There is some clipping.  Although it's not widespread.  If it sounds OK, then it IS OK 

Nothing that can be done to fix this show, just try to avoid it next show.

[vanark]

Thanks, Joe.  I guess I understood it was clipping, but why it was clipping so far below 0dB was what I was hoping to find out.

+T

[cleantone]

It is because something early in the chain was maxed out gainwise. It could be the mic, it could be preamp input stage. I am not too familiar with that rig. I think a PAD was needed somewhere though. Doesn't matter if the meter was hitting 0 or -15 it was getting maxed out before that. At least that is what I am assuming.

[vanark]

I can't get much more info from the taper, so we'll have to leave it with our guesses, unfortunately. If it was my recording, I'd be able to provide more info and maybe we could figure out the source/solution.

Thanks for all your help.  +T's to those I haven't given them to already. 

[dean]

I'm having a squaring off issue on one part (the bottom) of my wav.     Mr. Skalinder is looking into this but I'm hoping to hear from DSatz and others on this as well.

Here's the deal:
"left" mic directly between two cellos with their amps right behind them.  "right" mic situated in the tablas (with tabla amp just  behind him) and about 3' from the drum kit.  A guitar, amped, was well off to the left and really wouldn't come into play here.  He only sat in on a few tunes.

Rig is the AKG 414 lineage below, both mics set on omni.  It was a split omni pull split about 3 feet.  The mic stands were only up 2 & 3 feet, respectively, over a wood floor (an issue I hadn't considered before).  It was my first recording with the 414's and I used no attenuation whatsoever.  While I've seen this once or twice with the MBHOs I never thought much about it until I saw this thread.

Any assistance is much appreciated - thanks!

[Chilly Brioschi]

Dean-

Reflections?         

Since it seems to be happening with both sets of mics... perhaps the Preamp has something odd going on?
Have you tried the same mic arrangement with a different Pre/AD ?
Do you use the same cables with both mics? Perhaps (this is a real longshot here) a crystalized solder joint causing a diode-like effect? (it can happen, but is super-rare with modern solder and good technique)
Try any other set of cables.
Can you duplicate this in a test environment?

Have you touched base with whoever modded the Pre, so that they can see the waveform?

Vanark-

I recently recorded a show which had a Meyer reinforcement system set with some form of limiting on it.
I noticed about two songs into the show that I could crank the levels to peak at -3db and never topped 0bd.
The limiting / compression was on the source, in this case.
My peaks could have been -3db or -20db, depending on where I set the levels.
If there was strong soft clipping at the source, it would shown at -3db or -20db, depending on where my levels where.
This idea applies to each stage of the input chain.
The mic's built-in amplifier, the Pre, the A/D, and any digital (ie.- DSP) processing could each limit peaks and the levels applied after would simply "contain" them at the level recorded.

Here's a way to determine if the source may have been MP3 or mini-disk, although that doesn't really apply to your case, since you have a pretty solid idea as to the lineage:
http://ca.geocities.com/dkleined@rogers.com/audio/compression/Spectral.htm

No matter what, it doesn't sound bad.
Normalize volume to the level to a peak of -3db for louder playback and you are good to go!