Recomended Audio Level (in db) for DVD recording

[rmccann]

I did look at every post in the Audio section (that I know of) tonight before posting this.

I am wondering if anyone has a document or will share their knowlage involving the recomended recording level in "db" for Audio files recorded with video to disc

In the past I was used to using +3db used as the reference to 0db (in the days of analog audio recordig)g, but I notice that when I save files seperating the Audio and Video using ADS DVDXpress capture device, and I open that audio file with Audition, the level is at about -18db I have not yet burned a DVD or CDRW at this low of a level, just thought I would ask before I wasted a disc as the level seems so low to me.

While this may be a "pipe dream", what I would really like to see is a list of recomended Audio recording levels for DVD, SVCD, VCD, CDR (wave or any audio formats, as I would not expect to see a difference with the over all gain) I do realize that everything is relitive to some reference level, so that being on the list as well for a starting point would be great as well.

Thanks so much in advance for any and all assistance.

[SaSi]

The best way to record any digital signal is to normalize it so that the max. level is at 0dB. Any higher than that and you get clipping.

In the days of analog recording, clipping was rather soft and progressive and didn't harm too much.

With a digital format, however, clipping is as hard as an axe cutting off your neck.

Too low a MOL (Max. Output Level) and you loose dynamic range. If your mol is set at -18dB, then you lose 18dB from the dynamic range of the content.

Whatever encoding method you choose, (AC3 or MPEG1) low audio levels require higher amplification settings during playback and at low bitrates quantization noise becomes apparent if you listen to it though descent hifi speackers.

Having said that,

Most commercial DVDs that I have worked through CoolEdit PRO (audition in the pre-adobe days), show an unrealistically low recording level.

And, several commercial audio CDs (mosty dance music and hard rock) are heavily clipped beyond repair (check out the Shakira Laundry Service CD through Audition for example).

[vcdforme]

Generally I would stay around -12, some professionals prefer
-14 or -18. I just try to stay out of the red. And with many systems
it's OK to hit the red and even over OCCASIONALLY.

Sometimes, depending on the material, I'll slam red
pretty hard. Don't be afraid. Make some mistakes. Experiment.
Trust your ears, they're you're best tool.

[hiptune]

Good question.

I tried using the standard levels coming out of my VCR and LD players going into the ADVC-100 and found that mpg encoding was having it rough with levels that high. Then I lowered the levels being captured, and go cleaner audio in the end.

The issue of level is going to be different depending on final audio soundtrack type.

PCM audio can be louder and remain clean.

Mpeg audio needs to be kept rather low, as it will peak out easy.

AC-3 needs to be watched for overload, but needs more gain than mpeg, but should not be as high as pcm.

I am now using the ADVC-500 which has input level control. I find -15 db to be about right for DVDs. But I have not had enough experience with AC-3 encoding to know if any higher would over cook the AC-3. Still testing.

But I will say that I do not believe in normalizing my audio in general. I like lots of dynamics, and the least amount of resampling. So to get it captured clean and correct in peak level, then saved as a wav, and left alone until authoring is ideal if you want highest quality.

For DVDs 70min. and under, PCM is the best stereo sound you can have on a DVD. For over 70min. AC-3 is the way to go.

my .02

Jeff

[muskgrave]

am I missing something? we're all talking about digital audio correct? shouldn't you get it as close to 0db as possible? I'd amplify untill your "spikes" are almost there then use some moderate limiting @ -.01db

[hiptune]

am I missing something? we're all talking about digital audio correct? shouldn't you get it as close to 0db as possible? I'd amplify untill your "spikes" are almost there then use some moderate limiting @ -.01db

We are talking about compressed audio on dvd. That is a different thing than Digital CD levels.

I have a different way of working than you. I prefer the least amount of processing. So I would not jack up the levels, and then limit them. I would just nail it right perfec the first time around.

So no, you should not want to get as close to .01db. But if you are not listening on good playback equipment it may not matter as much.

BUt as stated above, it depends on what type of soundtrack you are encoding to on the finished dvd. You don't say what you are using.

Jeff

[rmccann]

Thank you all so much for yout time and effort in answering my questions!
Being from old school, analog 1/4 and 1/2 inch tape, and now attempting to transfer as much of those as possbable to digital format was one of the reasions I asked, also due to occational DVD's that I rip and end up noticing a -18db audio level. However as I am also a user of Cool Edit Pro, that answer was hit right on the head (thank you again!) as well as you kind folks answering a few questions that I most lkely would have asked at a later date. The community spirit here is amazing, not only have I learned so much from the tutorials, and "lists", these forums as such a wealth of infommation from the "learned" folks, who are kind enugh to pass on their knowlage. So thank you all again for sharing your knowlage, and especially to those of you who have put up this site, to assist those of us out in cyberland that do not have any other close method of finding answers and help. Cheers to all!

[FOO]

Anybody here know what dB means ? Exactly

[muskgrave]

decibels i would assume..but apparently i'm stooopid :P

and for those who just don't do things the way i do...
good luck keeping that pristine dynamic range on that COMPRESSED file
while you're at it I have some stereo files I want you to convert to 5.1


seriously though...have you seen the "professional" recordings as of late? they are way past moderate limiting...

[BJ_M]

-20 dB digital is exactly 0 dB on a analog tape machine

[FOO]

Db measurements are a RATIO not a level. You must know what the
reference level is for it to mean anything.
Nobody seems to get that.
Does anybody know what the reference is for audio ?
Radar uses 1 microwatt
Telephones , modems . etc use Dbm , which is 1 milliwatt

[muskgrave]

Nobody seems to get that.

yes and my car requires 1.21 gigawatts of electricity. What does this have to do with finding the appropriate digital audio level for a dvd? when did anyone question whether or not db was a ratio? It seems to me like someone read a book and is spouting out things that don't pertain to the original question.

..a peanut isn't really a nut..it's a legume...nobody seems to get that..

[FOO]

What does this have to do with finding the appropriate digital audio level for a dvd?

because , as I said , Specifying Db tells you NOTHING.
Somebody gives you a wire with 0 Db audio on it . How do you measure it ?
How many volts is that ? Tell us , smartass

[BJ_M]

FOO -- what is said is correct .. and if you want its dBfs , but no one really uses it that way .... BUT you are correct for things like dBu and dBV , dBA, dBC and so forth ..

here is a superb tutorial on the whole subject -- which says the same thign i did , but in a great amount of detail ..

http://www.indigipix.com/digtoanalog.htm

[dphirschler]

It's a database right?


Darryl :P

[FOO]

BJ_M , that site is pure unadulterated speculation and
bullshit.

[hiptune]

Db measurements are a RATIO not a level. You must know what the
reference level is for it to mean anything.
Nobody seems to get that.

Nobody needs to know jack about any reference or ratio, just keep an eye on the db audio levels, and capture your wav at about -15. I use the ADVC - 500, and I am getting nice clean and loud audio on my DVDs with no distortion. I love the dynamics with no limiting or normalization, as much of my material is concert performance.

I certainly get it!

Jeff

[FOO]

I give up. I won't bother you experts any more.

[CrustyCurmudgeon]

I'm with you, FOO.

If I had to guess, the 0dB reference point in digital audio would have to be the max unclipped level from the D to A converters. That's the only thing that would stay the same from PC to PC, Audio Card to Audio Card, DVD player to DVD Player. Since most PC's audio run through a digital mixer, they probably have some 'headroom' to prevent additive peak clipping. I should hook up my audio meters and see just what setting the mixer controls to midposition and max change. By ear, I think it's about 6dB on each control, so that means 12 dB for max on the master and one mixer source.

TV and Cable stations push their audio to just below the clipping point, about -3dB, since they know most analog TVs don't have super audio hum and noise levels and the listener's environment is generally noisy (kids, airconditioners, car traffic) and they really want the commercials to be heard.

Movie makers pride themselves on the super dynamic range of audio their products have, never mind that the actual movie may never have a silent spot, and decree the DVD's shall preserve that dynamic range. So their average levels are around -18dB to -24dB, that way when the bomb goes off in a scene, the audience has a collective heart attack along with the ruptured eardrums.

That being said, I hope the PC applications that are using dB settings are using 0dB as the clipping point.

[muskgrave]

What does this have to do with finding the appropriate digital audio level for a dvd?

because , as I said , Specifying Db tells you NOTHING.
Somebody gives you a wire with 0 Db audio on it . How do you measure it ?
How many volts is that ? Tell us , smartass

somebody's got a case of the MUUNDAYS
The fact remains that you are coming in here trying to flex some sort of intellectual muscle that nobody asked for or cares about. If I fart in a can whats the loudest I can make it? What does it have to do with the question? we are talking about a reference point for digital audio specifically on DVD. Does your audio editing program asks you about volts? No it doesn't. So according to your logic it must be an ignorant program. No...the program says.."here is 0db...if you go beyond it i'm gonna give you a little red flag saying don't be so loud". Perhaps this requires a lesser mind than yours.

"or perhaps..a mind less bound by the parameters of perfection"...

tee hee!

Crusty..you did help me to understand why you would want to keep things at a much lower level than 0db to make reservations for the "bigger" sounds....and hiptune you certainly also had the best way of recording things...to guarantee you don't clip you would still need some sort of compressor safeguarding the incoming feed correct?

[FOO]

Initial research reveals that audio usually
is referenced to 1 Volt RMS. 0 Db = 1 Vrms

Sorry to piss people off , but I have a long standing problem
with people bandying numbers about without having clue what they mean.

[BJ_M]

BJ_M , that site is pure unadulterated speculation and
bullshit.

what the heck you talking about ? the guy who wrote that , knows a thing or two ...

besides running the vegas forums and doing a tonne of training ..

here is a little about him -- i will not even get into my own background (i have written 2 books on digital audio theory and applications and also on compression and proofed several others on mpeg and compression)

Douglas Spotted Eagle, a talented musician; multiple Emmy and Grammy nominee and co-recipient of the 2001 Grammy for his production work on The Gathering of Nations project; new media visionary; and grandmaster of the highest order on all things SONAR.

“SONAR is absolutely essential to my production process, and I can say without exaggeration that it is my most used music tool.”

Douglas’ Native Restoration Studios, with Tom Bee, has been largely responsible for bringing Native American music, from authentic Pow Wow recordings, to experimental Modern Ethnic amalgams into the digital age, advancing the format from low-fi cassettes to state-of-the-art compact discs. Together, Spotted Eagle and Bee released the first-ever recording of Pow Wow music on compact disc, paving the way for their co-receiving the first-ever Native American Grammy Award. Along the way Douglas has fashioned a unique modern ethnic sound with his recently released VOICES on Higher Octave Music and Nino Reyos’ new CD Warrior of Time which was recorded 100% in SONAR.


Douglas has recorded 13 previous albums under his own name and 8 of those albums have achieved status on Billboard charts. He was named “Debut Jazz Artist of the Year” in 1996, and his Closer to Far Away recording was the #5 Jazz release of the same year. He collaborated with Kevin Costner on the production of 500 Nations, and Robbie Robertson on the score for The Native Americans. He has also scored or has been a featured in the scores of popular film/TV productions such as The Truman Show, Thunderheart, Millennium, The Scarlet Letter, The X-Files, City Slickers, The Way West, and Johnny Quest.

Douglas is actively involved in film and new media production. He most recently produced Toubat, a new historical documentary on the Native American Flute that features Richard W. Payne, MD, Mary Youngblood, Tom Bee, R. Carlos Nakai, Elmer Running, Jeff Calavan, Douglas Spotted Eagle himself, hundreds of flutes, and rare, never seen before video footage dating back as far as 1892. This entire documentary was scored using SONAR.

Douglas’ work in digital media has led him to be chosen as the Visionary keynote speaker for last year’s AVID World and DV Professional’s conventions. He will also be speaking at Streaming Media East. All of his talks will heavily feature SONAR as both a tool for developing streaming media and for audio content creation.

Douglas' music also drew the attention of the Olympic Committee, and he provided the theme music for the torch running through Tooele County, Utah during the recent Olympic games. Tooele County is where Douglas calls home, and it is also an area rich in Native American history.

One of Douglas’ favorite things to do is help other people. He has been a spokesperson for both the National Committee for the Prevention of Child Abuse and Straighttalkers USA, a national organization combating drug and alcohol dependency among youth. He even hosts a highly successful SONAR forum on the Internet at Creative Cow. Log on to his forum and get answers to your SONAR questions from a pro.

more here : http://www.spottedeagle.com/press.htm

[FOO]

I'm really not trying to cause trouble , and i wiil quit now after I quote
what I thought was bullshit

Unfortunately, this measurement was distorted by a number of factors, such as fluctuating changes in voltage.

What does this mean ? I think it implies foggy understanding

[muskgrave]

FOO (MAN CHOO),
I seriously doubt anyone's really ticked at you and it's your right to be somewhat annoyed by people throwing around words they don't fully understand..BUT i believe your delivery of this knowledge was the breaking point..people don't like being cornered with questions and being painted as mindless sheep.. instead of saying we "don't have a clue" perhaps we "don't fully understand all variables." There I'm happier already .

Second, however, was the fact that our understanding, although limited..was adequate for the application..i.e., the question at the top of this thread. My understanding was obviously less adequate than some of the others

but you're still invited to my birthday party. You just don't get to sit at the birthday boy table :P

[BJ_M]

Initial research reveals that audio usually
is referenced to 1 Volt RMS. 0 Db = 1 Vrms

Sorry to piss people off , but I have a long standing problem
with people bandying numbers about without having clue what they mean.

Audio equipment is often expressed in dBu - which represents the level compared to 0.775 Volts RMS with an unloaded, open circuit, source (u = unloaded).

dBm represents the power level compared to 1 mWatt. This is a level compared to 0.775 Volts RMS across a 600 Ohm load impedance. Note that this is a measurement of power, not a measurement of voltage.

therefore 0dB is not always 1Vrms , unless you are refering to dBV

What is often talked about here is digital audio - which is expressed as dBFS. The "FS" is not often wriiten out but assumed by most people.

0 dBFS represents the highest possible level in digital gear. All other measurements expressed in terms of dBFS will always be less than 0 dB (negative numbers).
0 dBFS indicates the digital number with all digits ="1", the highest possible sample.
The lowest possible sample is (for instance for 16 bit audio):
0000 0000 0000 0001, which equals -96 dBFS. Therefore the dynamic range for 16-bit systems is 96 dB. For 20-bit digital audio it is 120 dB. For 24 bit digital audio it is 144 dB.
Full-scale input level is the analog input voltage level that will cause the A/D converter to just equal full scale with no clipping on either positive or negative peaks.

Output full scale is defined as the analog output voltage produced while playing a 997 Hz digital full-scale sine wave, assuming the THD+N is less than -40 dB relative to the signal level.

This level is -20dB (FS) in a digital system which will be 0dB on a analog input.

The dynamic range of a digital system is the ratio of the full scale signal level to the RMS noise floor.

[dphirschler]

My simple approach to this question is this:

1) If the audio came from DVD, then leave it as is.

2) If I recorded the audio myself (such as a music concert), then normalize.

3) If the audio came from VHS, TV, or FM (ie. compressed), then 75% of normalized.

That's what sound right to my ears.


Darryl

[FOO]

BJ_M ,
Thanks.
I'm still not sure what the professional audio people use
is it dBu or dBV I think that's a 4 dB difference

[BJ_M]

To convert from dBV to dBu: 1dBV equals +2.2dBu.
+4dBu equals 1.23 Volts RMS.

The reference level of -10dBV (consumer) is the equivalent to a level of -7.8dBu.

+4dBu (pro audio equipment) and -10dBV systems have a level difference of 11.8 dB and not 14 dB. This is almost a voltage ratio of 4:1

why the use of an impedance of 600ohms to measure as a standard (when much of todays equipment is NOT 600ohms but high impedance)?
Its because of the largest audio system in the world - the telephone system, everything we do today in measures derives from the phone system and the transformers that were used (and still are)..

Whaty the heck is impedance you might ask - since its used a lot in speakers also, and in audio equipment (at one time) you were supposed to match impedance (days of tubes and transformer coupled connections) , but now its low into high is the rule.

i just copied the below info:

Impedance, denoted Z, is an expression of the opposition that an electronic component, circuit, or system offers to alternating and/or direct electric current. Impedance is a vector (two-dimensional)quantity consisting of two independent scalar (one-dimensional) phenomena: resistance and reactance.

Resistance, denoted R, is a measure of the extent to which a substance opposes the movement of electrons among its atoms. The more easily the atoms give up and/or accept electrons, the lower the resistance, which is expressed in positive real number ohms. Resistance is observed with alternating current (AC) and also with direct current (DC). Examples of materials with low resistance, known as electrical conductors, include copper, silver, and gold. High-resistance substances are called insulators or dielectrics, and include materials such as polyethylene, mica, and glass. A material with an intermediate levels of resistance is classified as a semiconductor. Examples are silicon, germanium, and gallium arsenide.

Reactance, denoted X, is an expression of the extent to which an electronic component, circuit, or system stores and releases energy as the current and voltage fluctuate with each AC cycle. Reactance is expressed in imaginary number ohms. It is observed for AC, but not for DC. When AC passes through a component that contains reactance, energy might be stored and released in the form of a magnetic field, in which case the reactance is inductive (denoted +jXL); or energy might be stored and released in the form of an electric field, in which case the reactance is capacitive (denoted -jXC). Reactance is conventionally multiplied by the positive square root of -1, which is the unit imaginary number called the j operator, to express Z as a complex number of the form R + jXL (when the net reactance is inductive) or R - jXC (when the net reactance is capacitive).

There is also acoustic impedance - which is a ratio of pressure to flow
(and is somewhat analogous to electrical impedance).

[FOO]

I used to design modems , so dBm was the standard.
You would think that dBu would have become the
audio standard - why make a new one - ?
and why did the digital people move it ?
All they had to do was keep it the same , with the disclaimer that
the headroom is a fixed 20 dB

its worse than cruising around the world and having to change
currency and/or metric/english all the time.

[sdsumike619]

does this whole issue have anything to do with the post I just made today in this forum? The topic is called:

? for audio gurus ! ! (recording guitar)