Most people normalize their waveforms before inserting them into a playback system. I've heard of different normalization standards set by different companies. What level do you normalize at, and why?
-
Get involved.
We want your input!
Apply for Membership and join the conversations about everything related to broadcasting.
After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
If you use a disposable or false email address, your registration will be rejected.
After your membership is approved, please take a minute to tell us a little bit about yourself.
https://www.radiodiscussions.com/forums/introduce-yourself.1088/
Thanks in advance and have fun!
RadioDiscussions Administrators
Normalization
- Thread starter SirRoxalot
- Start date
Normalization alone can be a pretty useless tool. If you are assembling a compilation of tracks from CDs, then what I am about to say about spikes and short-burst peaks may not apply.
I edit religious material that has been recorded live without compression. Some presenters have these little "lip-smacks" that are off the Richter Scale. Plosives can over-power a recording. Just arbitrarily normalizing without adjusting these aberrations is less than satisfactory.
Once your program material has been "rationalized" (I made that term up ;D ) Then you are ready to begin a normalization process. If there are multiple tracks or segments, I determine the RMS average level for each track and then adjust them to the same level. Then I normalize the whole thing so as your listed to the entire suite, all the tracks sound about the same in level.
Now, to finally get around to answering your question: I normalize at - 0.667 for CDs and Podcasts. I want the program material as far above the residual noise level that may exist in inexpensive playback devices. People don't tend to give Grandma at the nursing home an expensive high-quality playback machine in a location where good stuff can grow legs and walk off.
I suspect for material going in broadcast automation machines most operators will establish a lower value. Maybe -3.0 or -6.0
I edit religious material that has been recorded live without compression. Some presenters have these little "lip-smacks" that are off the Richter Scale. Plosives can over-power a recording. Just arbitrarily normalizing without adjusting these aberrations is less than satisfactory.
Once your program material has been "rationalized" (I made that term up ;D ) Then you are ready to begin a normalization process. If there are multiple tracks or segments, I determine the RMS average level for each track and then adjust them to the same level. Then I normalize the whole thing so as your listed to the entire suite, all the tracks sound about the same in level.
Now, to finally get around to answering your question: I normalize at - 0.667 for CDs and Podcasts. I want the program material as far above the residual noise level that may exist in inexpensive playback devices. People don't tend to give Grandma at the nursing home an expensive high-quality playback machine in a location where good stuff can grow legs and walk off.
I suspect for material going in broadcast automation machines most operators will establish a lower value. Maybe -3.0 or -6.0
SRP said:
Help me understand your logic. What needs headroom? If the normailizing process leaves your file with the single highest level wave at minus three, what is the headroom above that designed to accomodate?
I try to adjust every file to as close to 0 db as possible on peaks, without clipping, or much.
With music, loud isn't the only thing. If files don't have any headroom during "average levels" there will be no "room" for
any "punch" or dynamics to exist, and without them, music can't "breathe" properly.
Many times I've made a file too hot and had to start all over, leaving more "space" for the music.
Compression can be good, but square waves seldom are when fed to audio.
With music, loud isn't the only thing. If files don't have any headroom during "average levels" there will be no "room" for
any "punch" or dynamics to exist, and without them, music can't "breathe" properly.
Many times I've made a file too hot and had to start all over, leaving more "space" for the music.
Compression can be good, but square waves seldom are when fed to audio.
SRP said:
That is the traditional answer and earlier versions of audio equipment may have been more prone to do that than what is available today. It is not the sound cards that I worry about so much as the playback device. But I load my recordings into a $20 MP3 player and whatever distortion may result from my -0.667 peaks is not audible.
In reviewing this thread I would like to suggest the following reason that may valid for normalizing at a lower level. Some of the audio software may not do the math correctly, or may not give an accurate visual display of what it is really doing.
I use Adobe Audition and when I "select" a very short time frame and expand it to full screen where each individual cycle can be seen and evaluated, you can see if even the one loudest cycle (hertz?) in an audio peak has a flat top caused by over-driving the match process that results when normalization is done.
I had a laptop by Compaq circa 1996 that was the most unforgiving device I have ever seen if the input to the soundcard was over-driven. It did not leave "flat topped" wave forms, but actually went nuts and reversed polarity and put a splattered, jagged flat top on the far side of zero and then returned to normal until the next over drives wave. I once recorded a teenager doing a saxophone recital where about 15% of the time the card was doing this foolishness. I could have sold the recording to duck hunters to attract ducks. ;D
It is not flat tops that create the big problem, but AD converters that convert the over-limit portion of the wave to some kind of ringing distortion that is the killer. Some time in the last couple of years I downloaded a VST Plug-in called W1 Limiter and installed it in Adobe. I am still in awe at what this plug-in can do to audio when you ask it to hard limit. It will square off the waves if necessary and it has to get pretty extreme before it becomes "audible". I suspect someone with a good distortion meter could measure what it does whether my ears can or not.
SRP is spot on about the D/A part of a LOT of sound cards. I do a lot of work for a major production house in Orlando that has had an "issue" with my louder "0" level files..therefore I send them MINUS 4 and there is no problem. When I would send them ZERO level files their high end sound cards would break up and I'd end up either redoing them, or re-setting them to -4. As a result I now send everyone -4 files and if they want to normalize to zero and it accomodates their sound cards (read DA converters) they can. Besides most of what I am doing these days gets produced and re-processed anyway..so I ship as raw as I can for them to "tweak" as they like.
VO Dood does production on some of my work, and he re-works the dynamics to suit his stations, and it comes out just amazing...Same with the TV spot for Hyundai..I send the agency really raw audio (usually AIF files) and they make those puppies rock!
VO Dood does production on some of my work, and he re-works the dynamics to suit his stations, and it comes out just amazing...Same with the TV spot for Hyundai..I send the agency really raw audio (usually AIF files) and they make those puppies rock!
- Status