Effect of DC Offset Correction - Louder?

Page: 123 > Showing page 1 of 3
Author
larrymcg
Silver Member
  • Total Posts : 1299
  • Reward points : 0
  • Joined: 2003/11/05 17:00:22
  • Location: S.F. Bay Area (Silicon Valley)
  • Status: offline
2010/07/28 17:12:10 (permalink)

Effect of DC Offset Correction - Louder?

Before I started using Sonar HS7 (SHS7) I used CoolEdit2000 to normalize an audio track.  CE2K also removes DC Offset while normalizing.
Now I use SHS7's normailize function.  It doesn't include DC Offset removal as part of normalize but SHS7 provides a separate DC Offset removal function.

I happened to notice that one of my (unnormalized) audio tracks had a small DC Offset so I removed it.  That made the track sound much louder.

Is it normal that DC Offset removal will make the audio louder (even though the waveform hadn't been normalized)?

I had the audio going through a couple effects.  Could it be the combo of the effects processing and DC Offset?

--Larry

7/3/2012: Sonar X1 Studio. 10/2/2009: WinXP SP3; Intel Core 2 Duo E7500 (2.93GHz); Asus P5N7A-VM using on-board video and sound; 4GB RAM; 500GB, 250GB, 250GB HDDs; DVD writer, Midisport 4x4, midiman Fineline mixer, Roland JV-30 and Yamaha TG-55 synths...
homepage  http://lmmcgowan.home.comcast.net
#1

65 Replies Related Threads

    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 18:03:54 (permalink)
    A DC offset can be positive or negative.

    If the offset removed was negative, then a higher amplitude can result.



    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #2
    dmbaer
    Silver Member
    • Total Posts : 1770
    • Reward points : 0
    • Joined: 2008/08/04 20:10:22
    • Location: Concord CA
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 18:22:17 (permalink)
    Proper removal of DC offset should have the same effect as placing a high pass filter with a low cutoff frequency in the signal path.  Placing the filter there would not alter the volume, would it?  So, I would think something else is going on (or maybe it's just not doing it properly).
    #3
    bitflipper
    Platinum Member
    • Total Posts : 18090
    • Reward points : 0
    • Joined: 2006/09/17 11:23:23
    • Location: Everett, WA USA
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 18:40:58 (permalink)

    A DC offset can be positive or negative. If the offset removed was negative, then a higher amplitude can result.

    Sorry, no. A negative DC bias will simply shift the waveform toward the negative side of the 0 centerline. Overall amplitude remains the same.

    One possible impact of a DC offset, if it was large enough, would be to skew the threshold of a limiter. This could affect perceived volume, as well as potentially introduce harmonic distortion (and perhaps even audible aliasing) if the limiter were set to fast attack and release times.

    A DC offset could also affect normalization by causing peak values to be artificially increased in either the positive or negative excursions. It might prevent peaks from reaching the specified normalization level. That's just an educated guess on my part, since I never use normalize functions.

    Bear in mind that DC offsets are normally very small, typically so small as to not cause any problems at all. Consistent large offsets would probably indicate a defective piece of equipment.


    All else is in doubt, so this is the truth I cling to.

    My Stuff
    #4
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 18:53:36 (permalink)
    bitflipper



    A DC offset can be positive or negative. If the offset removed was negative, then a higher amplitude can result.

    Sorry, no. A negative DC bias will simply shift the waveform toward the negative side of the 0 centerline. Overall amplitude remains the same.

    One possible impact of a DC offset, if it was large enough, would be to skew the threshold of a limiter. This could affect perceived volume, as well as potentially introduce harmonic distortion (and perhaps even audible aliasing) if the limiter were set to fast attack and release times.

    A DC offset could also affect normalization by causing peak values to be artificially increased in either the positive or negative excursions. It might prevent peaks from reaching the specified normalization level. That's just an educated guess on my part, since I never use normalize functions.

    Bear in mind that DC offsets are normally very small, typically so small as to not cause any problems at all. Consistent large offsets would probably indicate a defective piece of equipment.

    The amplitude moved up, when the DC offset is removed, will result in a different RMS value, bitflipper.

    I said "higher" amplitude, meaning being upwards, my friend.

    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #5
    brundlefly
    Platinum Member
    • Total Posts : 8937
    • Reward points : 0
    • Joined: 2007/09/14 14:57:59
    • Location: Manitou Spgs, Colorado
    • Status: online
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:08:21 (permalink)
    The amplitude moved up, when the DC offset is removed, will result in a different RMS value, bitflipper.



    The direction of the offset is immaterial. RMS is not affected by displacement from 0. Bitflipper is correct.
    #6
    larrymcg
    Silver Member
    • Total Posts : 1299
    • Reward points : 0
    • Joined: 2003/11/05 17:00:22
    • Location: S.F. Bay Area (Silicon Valley)
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:13:17 (permalink)
    Well, I'm glad to see that everyone is in agreement on this! 

    I'm going to go find the original audio track and do some measurements and experiments to see what happens.  As I recall the centerline of the audio waveform was a small fraction of a db from 0db (at least according to the SHS7 display of the audio).

    --Larry


    7/3/2012: Sonar X1 Studio. 10/2/2009: WinXP SP3; Intel Core 2 Duo E7500 (2.93GHz); Asus P5N7A-VM using on-board video and sound; 4GB RAM; 500GB, 250GB, 250GB HDDs; DVD writer, Midisport 4x4, midiman Fineline mixer, Roland JV-30 and Yamaha TG-55 synths...
    homepage  http://lmmcgowan.home.comcast.net
    #7
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:22:59 (permalink)
    brundlefly



    The amplitude moved up, when the DC offset is removed, will result in a different RMS value, bitflipper.



    The direction of the offset is immaterial. RMS is not affected by displacement from 0. Bitflipper is correct.
    RMS is a total energy content of a wave. A DC offset IS PART of a wave when it is present.

    A DC offset represent a fixed content of energy and when it is present. It has to be included in the total energy of the wave.



    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #8
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:23:03 (permalink)
    NoKey


    The amplitude moved up, when the DC offset is removed, will result in a different RMS value, bitflipper.

    I said "higher" amplitude, meaning being upwards, my friend.
    no, that's incorrect.  removal of DC offset will NOT affect RMS.  there's no way to affect an AC signal's amplitude by adding or removing a DC component.


    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #9
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:31:33 (permalink)
    In a wave the DC offset is part of the wave.

    In assymetrical waves the DC offset is not plainly visible as is in a symmetrical wave.

    Sound waves are not symmetrical at all, except when using things like test tones.

    RMS is total content, it has to include DC offset..

    In a symmetrical wave you could think of offset being separate and just add it to the RMS of a symmetrcal wave, but that's a simplification. True RMS must include the total energy content of a wave.


    ....

    Now as far as the original question, I doubt that there be enough DC offset in a typical recording to make a big volume sound difference.
    post edited by NoKey - 2010/07/28 19:33:02

    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #10
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:33:19 (permalink)
    RMS is a total energy content of a wave.


    no, it is not.  RMS = Root, Mean, Square of a waveform.  for a pure sine wave, the formula is y=a*sin(2*pi*f*t) where a = peak amplitude, f = frequency and t = time.  there is no DC component in the forumula for RMS.  DC would mearly offset the crossover point up or down depending on the DC positive or negative value.  it is NOT part of the RMS value.

    http://en.wikipedia.org/wiki/Root_mean_square

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #11
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:34:26 (permalink)
    RMS is total content, it has to include DC offset..


    no it does not.  where did you get your Electrical Engineering degree?

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #12
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:51:48 (permalink)
    Beagle, you are seem to be thinking exclussively on terms of sine waves.

    Take a SQUARE WAVE and offset it.

    Or even better, take a wave or portion of a wave that does not cross the X axis, offset it can calculate its RMS and see if there is difference.

    For any wave you take the values at any point in the wave, square it, then take the mean value, then take the square root.

    The DC offset changes the value of each point high and low of the wave that you must use. You just don't exclude the DC offset, otherwise you don't get the correct energy content of the wave.

    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #13
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 19:59:49 (permalink)
    NoKey -
    I am tired of arguing with you.  just because you say it is doesn't mean you're right.  I've already figured that out from previous arguements with you.

    have YOU tried what you're saying?  no.  you have not.  RMS values will not change if you move them completely above the "x axis" or not.  the value of the ROOT MEAN SQUARE has nothing to do with DC.

    The DC offset changes the value of each point high and low of the wave that you must use. You just don't exclude the DC offset, otherwise you don't get the correct energy content of the wave.

    that statement is NOT the same thing as RMS.  you are confusing terms.

    what you are saying in that last statement is the only thing that will make sense for the OP's problem.  YES, a DC offset will affect normalization because it will bring the peaks of the waveform to a point where it can't be raised any more if the DC offset is there.

    but that is NOT the same thing as RMS.

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #14
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 20:11:48 (permalink)
    RMS is a method to calculate energy content, in contrast to simple measurements of peaks, or even just averaging.

    R stands for Root...M stands for media, and S stands for square(root).  Ironically, the term is backwards, because first you SQUARE, then you take the mean, and lastly the ROOT (square root) of any point in a wave.

    Sine waves are easy to calculate, .7 or something like that.  I honestly forgot.

    But assymetrical waves are a little more complicated.

    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #15
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 20:21:59 (permalink)
    NoKey


    RMS is a method to calculate energy content, in contrast to simple measurements of peaks, or even just averaging.

    R stands for Root...M stands for media, and S stands for square(root).  Ironically, the term is backwards, because first you SQUARE, then you take the mean, and lastly the ROOT (square root) of any point in a wave.

    Sine waves are easy to calculate, .7 or something like that.  I honestly forgot.

    But assymetrical waves are a little more complicated.

    yes, that's correct.  and I put the formula above if you choose to look at it. (which would equate to Vpeak/(square root of 2)).  I also posted a link on what RMS is.
     
    and yes assymeterical waves are more complicated (the link I posted above has more information on that as well).  but they still do NOT include DC content when calculating RMS.  you are simply confusing RMS with total power (which isn't EXACTLY correct either, but that's close enough for this conversation).
     
    for the OP removing the DC offset might increase his NORMALIZED wave because of the peaks being lowered after removing the DC offset.
     
    that is NOT THE SAME THING AS RMS.
     
    RMS does not include DC offset.  period.  if you still want to argue about it then take it up with an EE professor at a college because you have PROVED nothing and simply keep repeating the same thing over and over.

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #16
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 20:32:22 (permalink)
    Beagle


    NoKey


    RMS is a method to calculate energy content, in contrast to simple measurements of peaks, or even just averaging.

    R stands for Root...M stands for media, and S stands for square(root).  Ironically, the term is backwards, because first you SQUARE, then you take the mean, and lastly the ROOT (square root) of any point in a wave.

    Sine waves are easy to calculate, .7 or something like that.  I honestly forgot.

    But assymetrical waves are a little more complicated.

    yes, that's correct.  and I put the formula above if you choose to look at it. (which would equate to Vpeak/(square root of 2)).  I also posted a link on what RMS is.
     
    and yes assymeterical waves are more complicated (the link I posted above has more information on that as well).  but they still do NOT include DC content when calculating RMS.  you are simply confusing RMS with total power (which isn't EXACTLY correct either, but that's close enough for this conversation).
     
    for the OP removing the DC offset might increase his NORMALIZED wave because of the peaks being lowered after removing the DC offset.
     
    that is NOT THE SAME THING AS RMS.
     
    RMS does not include DC offset.  period.  if you still want to argue about it then take it up with an EE professor at a college because you have PROVED nothing and simply keep repeating the same thing over and over.
    Beagle, as I said, for a simple sine wave, there is no need even for formulas. Just a simple multiplication.

    The original RMS meaning is as I have described,  Each value of each point is the value that includes the DC offset if you want the total energy content of the wave.

    For the analysis of assymetrical waves and their energy contents, other mathematics are used that are more advanced, specially when  the systems are not one-phase but are multi-phased...Audio is  one-phase system.

    Power systems are 3 or more phases.  And , yes that's beyond the scope of this discussion...Just as your asking about my background.

    post edited by NoKey - 2010/07/28 20:34:02

    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #17
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 21:24:45 (permalink)
    The original RMS meaning is as I have described,  Each value of each point is the value that includes the DC offset if you want the total energy content of the wave.

    your argument is pointless.  prove it.

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #18
    RobertB
    Platinum Member
    • Total Posts : 10666
    • Reward points : 0
    • Joined: 2005/11/19 23:40:50
    • Location: Fort Worth, Texas
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 21:29:28 (permalink)
    Side note:
    Reece, I don't fully understand the principles involved, but this is what I was thinking of when we were looking at Larry's waveforms the other day.
    In particular, the last line of Bitflipper's post was what I had in mind.
    Larry, maybe you could post one of your screen shots here?
    I'd like to get a better understanding of what we are seeing here.




    My Soundclick Page
    X3eStudio,X1dEssential,Sonar8.3SE,SHS6XL,P5.2.5,MC5, Toshiba L505D,Vista 32 bit,AKAI EIE Pro, Edirol PCR-M50,Alesis DM6,Alesis ControlPad,Yamaha MG10/2,Alesis M1Mk2 monitors,Samson Servo300,Lava Lamp

    Shimozu-Kushiari or Bob
    #19
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 21:32:08 (permalink)
    bitflipper's entire post is correct. 

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #20
    jcatena
    Bronze Member
    • Total Posts : 404
    • Reward points : 0
    • Joined: 2009/06/09 09:34:04
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 21:47:34 (permalink)
    DC offset adds to true RMS as NoKey says, but it does not affect perceived loudness, as everyone else said.
    We don't hear DC.

    Jose Catena
    DIGIWAVES, S.L.
    #21
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 22:06:25 (permalink)
    Although audio is not my field, Jcatena, I believe that some DC offset can be slightly heard due to the fact that in some systems the DC is not too "pure" since it can come from some sort of rectification, in earlier systems that can have ripples, and now there has been a lot of improvements, but power supplies can still cause some.

    So if the DC that leaks somewhere is not pure DC, some slight hum might be heard.

    Also DC offsets do not necessarily have to be continuous, but can also vary and even pulsate, as a deffective equipment can cause,..In electronics likely maybe an old capacitor, or some other faulty component could cause a variable DC offset.

    I agree though, that DC is theoretically not audible because it has zero frequency. And I feel, and said, that in most audio recordings likely is not readily hearable. But it sure does flow through the speakers if it's there, can add to the saturation of those components that relay on coils and magnets if it were sufficiently high, affecting sound if so.

    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #22
    jcatena
    Bronze Member
    • Total Posts : 404
    • Reward points : 0
    • Joined: 2009/06/09 09:34:04
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 22:19:42 (permalink)
    1) Any ripple or variation in the DC is AC (like the audio) and could be heard. The DC itself can not be heard.
    2) The DC in digital audio recordings, usually very small, is caused by drift in the zero offset of ADC input or preamps, and normally it only varies very slowly following temperature changes.
    3) In 99.9% of audio amplifiers, DC is blocked. So normally the DC doesn't reach the speakers.
     
    post edited by jcatena - 2010/07/28 22:22:57

    Jose Catena
    DIGIWAVES, S.L.
    #23
    larrymcg
    Silver Member
    • Total Posts : 1299
    • Reward points : 0
    • Joined: 2003/11/05 17:00:22
    • Location: S.F. Bay Area (Silicon Valley)
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 22:23:41 (permalink)
    Hi!  I'm the OP.  I have found the original audio track and have done a couple experiments.
    I have to leave for a big band rehearsal now so tomorrow I will post a couple screen shots and mp3s to demonstrate the differences.



    7/3/2012: Sonar X1 Studio. 10/2/2009: WinXP SP3; Intel Core 2 Duo E7500 (2.93GHz); Asus P5N7A-VM using on-board video and sound; 4GB RAM; 500GB, 250GB, 250GB HDDs; DVD writer, Midisport 4x4, midiman Fineline mixer, Roland JV-30 and Yamaha TG-55 synths...
    homepage  http://lmmcgowan.home.comcast.net
    #24
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 22:32:44 (permalink)
    jcatena


    DC offset adds to true RMS as NoKey says, but it does not affect perceived loudness, as everyone else said.
    We don't hear DC.

    as I said to nokey, you're going to need to quote a reference to back up your statement.  if you can show me where I'm wrong with a textbook or reputable reference then I will concede, otherwise you're just heaping more opinion into the disagreement without proof.
     
    the link I gave above has the direct applications for RMS calculations and no DC values are included in those formulas.
     
    DC will affect total POWER or ENERGY, but not RMS.  RMS is calculated by peak values of the wave(s) with respect to time and frequency and do not apply only to sine waves (as is also discussed in that link above).  DC offset only adds offset to the wave, it does not add value to the RMS value.  in nokey's example above - if you have a waveform of 10Vp-p (which is calculated at 7.07Vrms) at DC offset of 0DC then you have 5Vpeak above and 5Vpeak below 0DC.  if you move that waveform up by 5VDC (offset by 5VDC), then your lowest peak is at 0Vpeak and your highest is at 10Vpeak, but your peak to peak value is STILL 10Vp-p (and thus your RMS value is also still the same).  it is not affected by the DC offset.

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #25
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/28 23:48:24 (permalink)
    Beagle


    jcatena


    DC offset adds to true RMS as NoKey says, but it does not affect perceived loudness, as everyone else said.
    We don't hear DC.

    as I said to nokey, you're going to need to quote a reference to back up your statement.  if you can show me where I'm wrong with a textbook or reputable reference then I will concede, otherwise you're just heaping more opinion into the disagreement without proof.
     
    the link I gave above has the direct applications for RMS calculations and no DC values are included in those formulas.
     
    DC will affect total POWER or ENERGY, but not RMS.  RMS is calculated by peak values of the wave(s) with respect to time and frequency and do not apply only to sine waves (as is also discussed in that link above).  DC offset only adds offset to the wave, it does not add value to the RMS value.  in nokey's example above - if you have a waveform of 10Vp-p (which is calculated at 7.07Vrms) at DC offset of 0DC then you have 5Vpeak above and 5Vpeak below 0DC.  if you move that waveform up by 5VDC (offset by 5VDC), then your lowest peak is at 0Vpeak and your highest is at 10Vpeak, but your peak to peak value is STILL 10Vp-p (and thus your RMS value is also still the same).  it is not affected by the DC offset.
    Dear Beagle,

    RMS is calculated from peak values ONLY when the wave is symmetrical, or sinusoidal.

    Common sense should tell you that when a wave is not symmetrical there is NO way that just the peak can determine its energy contents.



    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #26
    larrymcg
    Silver Member
    • Total Posts : 1299
    • Reward points : 0
    • Joined: 2003/11/05 17:00:22
    • Location: S.F. Bay Area (Silicon Valley)
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/29 03:19:59 (permalink)
    Hi!  I'm the OP.  I have found the original audio track and have done some experiments.  If you don't want to slug through all my experiments, the bottom line is that SHS7 has some kind of problem and the audio playing louder has NOTHING to do with removing DC Offset.  Of course, if you want to, read on.......   Or skip directly to experiments 3 and 4 below.

    I made sure there were no audio effects (like reverb) being used.

    The center line of the original audio is below the audio display center line.  The display center line is labeled -infinity db and the audio center line is at about -27db.

    The original audio track was in a ProAudio9 project.  I saved the PA9 wrk file as a bun and then opened the bun with SHS7.  The waveform in SHS7 looks like the one displayed by PA9 and they both show the same DC Offset.  However, with the PA9 and SHS7 volume sliders all set to 0db, the sound level is much lower in SHS7.  Seems like they should be about the same and I think they are the same for tracks that do not have DC Offset.

    In the following experiments, I always started with the unmodified audio from the bun file in SHS7.

    EXPERIMENT 1: In SHS7, I applied the remove DC Offset function.  The result is much louder.

    EXPERIMENT 2: In SHS7, I sent the unmodified audio to CoolEdit2000 where I removed the DC Offset.  When played back in CE2K, the sound level did not change.  When the modified audio was returned to SHS7, SHS7 played it back at the original softer level.
    Then I applied the SHS7 DC Offset removal to the audio from CE2K which already had the DC Offset removed.  The playback got much louder even though the waveform did not change.  Applying the DC Offset removal yet again had no additional effect.

    EXPERIMENT 3: I accidentally found that if I clicked (just once) in the unmodified audio track with the scissors (Split Tool) the audio played back much louder (just the same as when I had used DC Offset removal).  That's crazy!

    EXPERIMENT 4: I found that if I dragged the unmodified audio a bit to the right, the audio played back loudly.  That's crazy too!

    I think experiments 3 and 4 show that SHS7 has something wrong with it.  The audio playing back louder has NOTHING to do with removing DC Offset.

    Now ain't that just lovely!

    --Larry


    7/3/2012: Sonar X1 Studio. 10/2/2009: WinXP SP3; Intel Core 2 Duo E7500 (2.93GHz); Asus P5N7A-VM using on-board video and sound; 4GB RAM; 500GB, 250GB, 250GB HDDs; DVD writer, Midisport 4x4, midiman Fineline mixer, Roland JV-30 and Yamaha TG-55 synths...
    homepage  http://lmmcgowan.home.comcast.net
    #27
    NoKey
    Silver Member
    • Total Posts : 974
    • Reward points : 0
    • Joined: 2008/10/28 15:30:19
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/29 04:02:32 (permalink)
    Hi Larry...Yes sounds weird.

    Why not try the wave in a virgin project. Kill the existing one, and create a new one.

    Maybe there's some automation or something in the track that contains the wave?

    Or try another wave in the track and project as it is..Wipe the track contents and import another audio wave, and see what happens.

    Interesting..Hope you find the reason.

    Best of all.
    post edited by NoKey - 2010/07/29 04:23:12

    Songs I've produced with Sonar LE: http://www.soundclick.com...ult.cfm?bandID=1086857
    #28
    Beagle
    Platinum Member
    • Total Posts : 46335
    • Reward points : 0
    • Joined: 2006/03/29 11:03:12
    • Location: Fort Worth, TX
    • Status: offline
    Re:Effect of DC Offset Correction - Louder? 2010/07/29 07:09:44 (permalink)
    Dear Beagle,

    RMS is calculated from peak values ONLY when the wave is symmetrical, or sinusoidal.

    Common sense should tell you that when a wave is not symmetrical there is NO way that just the peak can determine its energy contents.

    all I'm asking, El, is that you show me the reference where you can find the DC value in the formula for determining the RMS value of a non-uniform waveform. 

    I have not argued anywhere that you are incorrect when you say that a complex waveform is calculated by something more than Vpeak/sqrt2.  (technically for audio waveforms the RMS value is simply calculated as the average of the peak to peak).  I have said over and over again that it does not include DC offset in the calculations and you cannot convince me that it does by saying it's "common sense."  prove it with a reference and I will concede.  saying it's "common sense" is completely unacceptable for proof.

    Larry - very sorry for sidetracking your problem.  in your experiments you say that SHS is playing them louder but not cooledit.  are you using your ears as the reference or are you looking at peak levels on your meters in SHS and cooledit and seeing that SHS has a new higher peak value?

    http://forum.christiansongwriters.org
    http://www.beaglesound.com/
    http://soundcloud.com/beaglesound/sets/featured-songs-1
    i5 2500K, 16G DDR3, Win7x64, Asus P8P67, MOTU Ultralite Hybrid MK3
    Privia PX-3, Yamaha MO6
    #29
    brundlefly
    Platinum Member
    • Total Posts : 8937
    • Reward points : 0
    • Joined: 2007/09/14 14:57:59
    • Location: Manitou Spgs, Colorado
    • Status: online
    Re:Effect of DC Offset Correction - Louder? 2010/07/29 11:49:11 (permalink)
    OMG. What have I started...?

    NoKey is technically correct about RMS, if you're talking about the electrical energy. But we're not talking about electrical energy; we're talking about acoustic energy. And a DC offset does not add to the acoustic energy of a signal coming from your monitors.

    All a DC offset does (if it even makes it to the output of your interface and through the amplifier, which is a whole other piece of the puzzle) is offset the cone of the driver in your monitor from it's normal resting point. Putting speaker mechanics and the increased possibility of clipping aside, this will not alter what you hear, because all that matters in terms of the signal that gets transferred to the air, and hence to your ears, is the excursion of the speaker cone to either side of that new resting/center point.




    #30
    Page: 123 > Showing page 1 of 3
    Jump to:
    © 2014 APG vNext Commercial Version 4.6