VO-BB - 20 YEARS OLD!

Frank F · Posted: Tue Apr 10, 2007 2:31 pm Post subject:

My solid peaks are at -2 db or 80 to 90% input; with a noise floor of -69 db.

This is a typical screenshot, after recording and before processing:

Minimum input to produce quality audio should be 65 to 75% or -9 to -6db for optimum levels. Spikes should be near 100%, too many spikes above 100% result in clipping. Clipping means audio problems and artifacts... which are not good.

Check how your inputs are driven and the output power of your soundcard if these typcial results are not achieved. Many soundcards or pre's are very low level and force using "normalization" or "compression" to achieve the audio level desired. Both of these processes may result in undesired noise and/or artifacts.

Toodles

Frank F
_________________
Be thankful for the bad things in life. They opened your eyes to the good things you weren't paying attention to before. email: thevoice@usa.com

Hart · Posted: Tue Apr 10, 2007 3:27 pm Post subject:

Now Frank that brings up a side point. It was my understanding that while the occasional peak over 100% was acceptable in the analog world it's a big time no no in the digital realm because there is no headroom like in analog.

Sometimes I really miss the Otari 8 track. Rust I can understand but ones and zeros make my head spin.

EDIT. So I've gone digging in the Audition help manual.

First, Adobe's definition of clipping: In digital audio, distortion that occurs when the amplitude of a signal exceeds the maximum level for the current bit resolution (for example, 256 in 8-bit audio). Visually, clipped audio produces broad flat areas at the top of a waveform.

Adobe says, "To get the best sounding results, you should record audio as loud as possible without clipping. Try to keep the loudest peak somewhere between –2 dB and 0 dB when setting recording levels."

And finally in the Level meters overview section, Adobe says, "the meters show signal levels in dBFS (decibels below full scale), where a level of 0 dB is the maximum amplitude possible before clipping occurs. Yellow peak indicators remain for 1.5 seconds so that you can easily determine peak amplitude. If amplitude is too low, sound quality is reduced; if amplitude is too high, clipping occurs and produces distortion. The red clip-indicator to the right of the meters lights up and remains lit when levels exceed the maximum of 0 dB."
_________________
Hart Voice Overs Blog
Brian Hart Productions

Frank F · Posted: Tue Apr 10, 2007 3:38 pm Post subject:

So true, but you may often find a spike NEAR 100% (example shown, shows a spike at 99%). Over 102% will clip. The tolerences of digital are not as forgiving as analog.

If you are spiking NEAR 100%, that is good. You can always take the spike down 3 db and get a non-clipped sound.

F2
_________________
Be thankful for the bad things in life. They opened your eyes to the good things you weren't paying attention to before. email: thevoice@usa.com

Hart · Posted: Tue Apr 10, 2007 3:39 pm Post subject:

gotcha. Thanks.
_________________
Hart Voice Overs Blog
Brian Hart Productions

Jeffrey Kafer · Posted: Tue Apr 10, 2007 4:00 pm Post subject:

Frank F · Posted: Tue Apr 10, 2007 4:44 pm Post subject:

Most "normailizers" have a factory default of just under 100%.

Personally, I have a pre-set of 85%. This setting allows for analog headroom during playback. When any file has been 'mastered' in my studio, the audio leaves my studio at a maximum of 85%.

The pre-set above allows for a warmer, more comfortable, less sterile feel when listening. The pre-set also allows for adjustments made during radio or TV processing. I have noticed my commercials; heard on a highly processed radio station, sound as loud, but cleaner and richer than those of a studio who uses major proportions of processing (compression, limiting, etc.), and whose files are notoriously at a full 100% level.

Processing can actually work against you if the BFO is not set the same as the broadcast facilities processors. Over compressed audio will be limited down to compensate for the previous audio in the chain. Sometimes the offset of compression and limiting can sound so much louder than the program audio the processing creates an un-natural sound which will ultimately be heard as 'artifcts' or noise by the listener/viewer. Most often this is due to compression on the original file by the talent and the broadcast stations compression to fit within the small spectrum allowed to be sent through the transmitter.

OK, more than enough information. Suffice it to say, 85% is the audio level at which your files should be finalized/mastered. Your recording level should be between 55 and 80% with some transient spikes reaching as high as 99.5%. (The 'transient spikes mentioned above are usually "S's"; hard K's, C's, T's, sibillant Q's, clicks, or low loud bass; like a kick drum or hard processed bass.

Too low of an audio input level and your mastered file will sound thin. Too high of an input level and your audio will sound clipped or have artifacts.

Toodles

Frank F

Think: Bunny Slippers.
_________________
Be thankful for the bad things in life. They opened your eyes to the good things you weren't paying attention to before. email: thevoice@usa.com

SoundsGreat-Elaine Singer · Posted: Tue Apr 10, 2007 4:59 pm Post subject:

Hart · Posted: Tue Apr 10, 2007 5:15 pm Post subject:

Chuck Davis · Posted: Tue Apr 10, 2007 6:38 pm Post subject:

This is an interesting thread. Especially when it comes to specialized apps like radio imaging..."compression/limiting city". I've been producing rock radio imaging for a long time now. It takes awhile to find out how many stages of gain reduction you can "get away with". In the end it almost comes down to "squashing it" enough that the airchain can't change things much on you.

Personally, for straight-up vo..dry stuff...I try to send things out peaking
at -3. In Pro-Tools the "clip light" comes on well before the audio actually clips..so the occational red light..as long I'm not actually hearing clipping...makes it thru.
_________________
Wicked huge.....in India.
www.chuckdaviscreative.com

Yoda117 · Posted: Tue Apr 10, 2007 7:14 pm Post subject:

Jeffrey Kafer · Posted: Tue Apr 10, 2007 7:16 pm Post subject:

TheVoiceOfBob · Posted: Wed Apr 11, 2007 5:18 pm Post subject:

sdelgo · Posted: Wed Apr 11, 2007 6:08 pm Post subject:

Proper recording level is -18dbFS which equals 0dbVU.
Your signal should average 0dbVU from your preamp which is -18dbFS (digital) on your AA program.
Heck, recording in 24bit you can even achieve higher resolution than cd quality by recording material that peaks at -47dbFS.

looks like Amy wins.

Steve
_________________
you'll always have something on your plate... if you keep your bearings straight.

www.steviedproductions.com