IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Sound Quality After File Conversion from Stereo to Mono
dspargo
post Sep 5 2012, 21:57
Post #1





Group: Members
Posts: 3
Joined: 22-June 09
Member No.: 70895



I work in a office where I can listen to music with my headphones using the left one. I want to take my stereo files and convert them to mono using dBpoweramp Music Converter. It is easy enough to do but I do have some concerns.

Originally I used an adapter to go from stereo to mono and my perception was that it muddied the sound. There are some DSP effects I can use in J. River Media Center 17 but I feel like those mess things up too. I even looked into getting a mono headphone amp but those are not easy to find and I would rather not buy a whole new amp. I am using a CEntrance DACport and love it because of the sound quality and it takes up minimal space on my desk.

I've done some research and it looks like I might need a summing box.

Basically I don't want to have to change my setup too much and would like to know if the conversion from stereo to mono is going to affect the sound quality like the adapter. I've listened to some converted files and they do seem to sound good. I do realized that I will be losing the stereo affect and the nuances with that.

Thank you in advance for your help. This has been driving me crazy finding a solution that has minimal changes to my WAV files.
Go to the top of the page
+Quote Post
DVDdoug
post Sep 5 2012, 23:17
Post #2





Group: Members
Posts: 2535
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
IOriginally I used an adapter to go from stereo to mono and my perception was that it muddied the sound.
Like a Y-adapter? Right! You can use a Y-adapter to connect two inputs together, such as "splitting" a mono signal and connecting it to the left & right inputs of a stereo amplifier, but you should never connect two outputs together! These circuits have (relatively) low output impedance, and basically each output "shorts-out" the other. You'll get distortion and it's even possible to damage the device. If you try that with a power amplifier, you most-likely will fry it.

QUOTE
I've done some research and it looks like I might need a summing box.
Yes. Or, a mixer. But, computers are pretty good at summing too! biggrin.gif I don't use dBpoweramp, but I assume it works fine.

Or you could use an audio editor. If you don't have an audio editor, Audacity is FREE!

I assume dBpoweramp takes care of this automatically, but the one thing you have to be aware of is that when you mix (sum), you are of-course, increasing the levels and you can get clipping (distortion). So, the trick is to cut the signal in half (reduce by 6dB) before mixing. Or, with most audio editors (including Audacity) you can cut the level in half after mixing, but before saving/rendering, since most audio editors work in floating-point internally and they won't clip internally. The waveform will be clipped if it goes over 0dB and you save as a regular (integer) WAV file. (Analog mixers also sum and they have input-level controls and a master volume control, and they can have "headroom" that's not always available in digital.)

P.S.
QUOTE
This has been driving me crazy finding a solution that has minimal changes to my WAV files.
Just FYI - You'll find it easier to "tag" your files (with the song title & artist, etc.) if you use a compressed format instead of WAV. The compressed formats are more "tag-friendly". Plus you'll save disc space (if that's imortant). If you want to preserve 100% of the quality, you can use a lossless format (FLAC or ALAC).

Or, high-quality MP3 or AAC files (lossy) can have very-good quality and often sound identical to the uncompressed original (in blind ABX tests). And, you've already decided to alter the original sound by going mono anyway... And, you are listening with one ear in an imperfect environment... So, I wouldn't be too quick to "write-off" lossy compression.

This post has been edited by DVDdoug: Sep 5 2012, 23:41
Go to the top of the page
+Quote Post
_if
post Sep 6 2012, 00:37
Post #3





Group: Members
Posts: 55
Joined: 18-December 10
From: Ohio
Member No.: 86607



I don't think there's any reason converting to mono with dBpoweramp or downmixing with J. River should sound any worse than using hardware to sum the signal, unless it does some other (inaccurate) sweetening. Maybe the reason it sounded bad when you tried is phase cancellation? Or mono just might sound bad for the music since it wasn't mixed for it.
Go to the top of the page
+Quote Post
greynol
post Sep 6 2012, 00:54
Post #4





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (dspargo @ Sep 5 2012, 13:57) *
I am using a CEntrance DACport and love it because of the sound quality and it takes up minimal space on my desk.

Careful!


--------------------
Placebophiles: put up or shut up!
Go to the top of the page
+Quote Post
dspargo
post Sep 6 2012, 01:00
Post #5





Group: Members
Posts: 3
Joined: 22-June 09
Member No.: 70895



QUOTE (greynol @ Sep 5 2012, 19:54) *
QUOTE (dspargo @ Sep 5 2012, 13:57) *
I am using a CEntrance DACport and love it because of the sound quality and it takes up minimal space on my desk.

Careful!


Careful?
Go to the top of the page
+Quote Post
greynol
post Sep 6 2012, 01:19
Post #6





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



That was a link. Did you click on it?

This post has been edited by greynol: Sep 6 2012, 01:19


--------------------
Placebophiles: put up or shut up!
Go to the top of the page
+Quote Post
RobWansbeck
post Sep 6 2012, 01:53
Post #7





Group: Members
Posts: 23
Joined: 24-January 10
Member No.: 77468



Unfortunately simply adding left and right leads to instruments originally in the centre sounding louder. Since bass instruments are usually placed in the centre this could account for some of the 'muddied' sound you hear.
Go to the top of the page
+Quote Post
greynol
post Sep 6 2012, 02:15
Post #8





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



That doesn't make sense to me.

If you place your stereo speakers right next to each other and stand far enough away will you get the same effect? If not then why not?


--------------------
Placebophiles: put up or shut up!
Go to the top of the page
+Quote Post
dc2bluelight
post Sep 6 2012, 04:12
Post #9





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



QUOTE (RobWansbeck @ Sep 5 2012, 19:53) *
Unfortunately simply adding left and right leads to instruments originally in the centre sounding louder. Since bass instruments are usually placed in the centre this could account for some of the 'muddied' sound you hear.

This is actually correct, and why many recordings don't quite sound right when simply summed to mono. Any sound exactly identical in both channels is hyped by addition (+6dB) while any sound panned hard to one or the other channel won't have the same additive boost. It's differnt than playing stereo from two speakers because of being in an acoustic space, and the directional components of stereo, but if two speakers were exactly coincedent in position the same problem would exist.

Here's the deal, mono from stereo is a compromise. Unless specifically mixed to work that way its going to be different than hearing the piece in stereo. The only practical way to end up at mono from stereo is a simple sum, which from what I read, can be done as a DSP effect live in dBPoweramp (so you don't have to make mono versions of all your tunes). There are ways to combine with less center buildup, but they aren't simple, and arn't found in free software. The basic idea is to introduce a 90 degree phase difference between channels pre-summing, which reduces the center buildup to a more reasonable level, but that 90 degree trick...well, it's basically a bunch of phase shift in both channels, but with one channel 90 degrees off from the other. Some might argue that all that phase shift has its own down side, I'm not going to say either way, but that's how it's done, and does work. Just ask the folks at Dolby (it's part of the encoder for LTRT tracks that we now decode with ProLogic). It's just not something you can get your mits on easily. I'd stick with the channel mapper in dBPoweramp or the like, ditch the Y cord.

So sum-up and mono-up, it's a compromise. Got to be better than total silence, right? And, may I say, a whole lot better than listening to only one channel...ever sit in a restaurant playing oldies that way? The Mamas and Papas become just the Mamas. The Beatles become an instrumental band. The Dave Clark 5 becomes the Dave Clark 2.5. You get the idea.

By the way, listening with one earphone causes its own "mask" on things that you may not like or find easy to get used to. Just another compromise, though.

Go to the top of the page
+Quote Post
2Bdecided
post Sep 6 2012, 10:25
Post #10


ReplayGain developer


Group: Developer
Posts: 5059
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



dc2bl is right - that's why dedicated mono mixes were made in the days when it was still widely used.

It's also the explanation for the weird mixing on some stereo albums in the says when many people would buy a stereo album, and play it with a stereo pickup/cartridge, but then sum the output to mono for feeding into a mono amplifier and speaker. They bought stereo albums to future-proof their record collection for the day when they could afford a stereo amplifier. So, for example, George Martin mixed Rubber Soul with nothing in the centre - to avoid the problem that anything mixed to the centre jumps up relatively in level when combined to mono.


I'm not sure I agree with DVDdoug - it's true that damage can be caused by shorting out speaker outputs, but most headphone outputs are more resilient IME. You may be lucky or unlucky - if it damages your equipment, please don't come crying to me - but sometimes it's OK. Some equipment is built to cope with (or be protected against) direct shorts, but not necessarily coupling the two channels together.

Cheers,
David.
Go to the top of the page
+Quote Post
RobWansbeck
post Sep 6 2012, 10:53
Post #11





Group: Members
Posts: 23
Joined: 24-January 10
Member No.: 77468



Power is proportional to the square of the voltage.

If 1V gave one Watt of power then an instrument to the far left or right would require 1V to be heard at 1W. An instrument in the centre would require 0.5W from each speaker to be heard at 1W overall. This would require 0.7V per channel, i.e. the square root of 0.5.

If the channels are summed the instruments at the far left or right would still be 1V giving 1W but the instrument in the centre would now be 1.4V giving 2W or +3dB.

As dc2bluelight noted, this can be corrected by putting a 90 degree phase shift between the channels so that the 2 centre signals add to 1V rather than 1.4V.

I've never dealt with this since the days of NICAM stereo when mono compatibility was an issue and don't know if anyone has produced an accurate solution.
Go to the top of the page
+Quote Post
mzil
post Sep 6 2012, 19:01
Post #12





Group: Members
Posts: 499
Joined: 5-August 07
Member No.: 45913



This Rane article describes the problem of attempting to use a Y-cord to combine the L and R of stereo line level signals and how to build a simple circuit to overcome the problem.

I disagree with the others, above, that there will be a problem such that the center stage musician(s) [let's say for instance the vocalist] will end up being too loud by simply summing the L and R together (of a stereo recording into mono). What people are forgetting is that when listening to a single channel by itself, the vocalist will seem to be too quiet. This is because the recording engineer assumes there will be another speaker in the room playing the same sound of the vocalist simultaneously, stereo that is, and the combined L and R speakers' output heard together will reproduce the vocalist at the correct level. It doesn't matter if the summing is done in the acoustical domain or the electrical domain.

This post has been edited by mzil: Sep 6 2012, 19:21
Go to the top of the page
+Quote Post
bryant
post Sep 6 2012, 19:38
Post #13


WavPack Developer


Group: Developer (Donating)
Posts: 1290
Joined: 3-January 02
From: San Francisco CA
Member No.: 900



QUOTE
It doesn't matter if the summing is done in the acoustical domain or the electrical domain.

Actually, it's a little counter-intuitive, but there is a difference. Two speakers in a room playing the identical signal are only 3 dB louder than a single speaker playing that signal. The reason is that at almost any given point the distance to the speakers is not identical (which reduces correlation) and a significant portion of the sound has been reflected off various surfaces multiple times (which further reduces correlation). If your ear was exactly the same distance between the speakers in a perfect anechoic chamber, only then you would get the full 6 dB of gain that you get from summing the signals electrically.

Imagine playing a pure continuous tone in both speakers. In this case (depending on the frequency) you would get standing waves in the room. In some places the speakers would be in phase (ignoring reflections) and the gain would be 6 dB whereas in other spots the output would cancel and there would be no signal. Again, the gain averaged over the entire room would be 3 dB.

edit: eliminated confusing parenthetical remark

This post has been edited by bryant: Sep 6 2012, 19:53
Go to the top of the page
+Quote Post
DVDdoug
post Sep 6 2012, 19:52
Post #14





Group: Members
Posts: 2535
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
Two speakers in a room playing the identical signal are only 3 dB louder than a single speaker playing that signal (even though twice the power would suggest 6 dB).
No!!!! Twice the power is +3dB. Nothing in math or physics "suggests" +6dB.
Go to the top of the page
+Quote Post
bryant
post Sep 6 2012, 19:55
Post #15


WavPack Developer


Group: Developer (Donating)
Posts: 1290
Joined: 3-January 02
From: San Francisco CA
Member No.: 900



QUOTE (DVDdoug @ Sep 6 2012, 10:52) *
QUOTE
Two speakers in a room playing the identical signal are only 3 dB louder than a single speaker playing that signal (even though twice the power would suggest 6 dB).
No!!!! Twice the power is +3dB. Nothing in math or physics "suggests" +6dB.

Yeah, you're right. I noticed that mistake about the same time you did and revised (I'm supposed to be working on something else). Thanks! smile.gif
Go to the top of the page
+Quote Post
RobWansbeck
post Sep 6 2012, 20:56
Post #16





Group: Members
Posts: 23
Joined: 24-January 10
Member No.: 77468



If you sum in the electrical domain you are summing the voltage. If you sum in the acoustic domain you are summing the power which is proportional to the square of the voltage.

Doubling the voltage gives 4 times the power, i.e. +6dB.

Doubling the power gives twice the power, i.e. +3dB.
Go to the top of the page
+Quote Post
dc2bluelight
post Sep 7 2012, 04:24
Post #17





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



QUOTE (RobWansbeck @ Sep 6 2012, 14:56) *
If you sum in the electrical domain you are summing the voltage. If you sum in the acoustic domain you are summing the power which is proportional to the square of the voltage.

Doubling the voltage gives 4 times the power, i.e. +6dB.

Doubling the power gives twice the power, i.e. +3dB.

...except... acoustic measurements are usually done by measuring SPL, where we're back to "pressure", which behaves like voltage (also a form of pressure). So no, acoustically you sum pressure, so you'd expect a theoretical 6dB. But as noted, playing two speakers with identical material won't get you 6dB gain because of, oh what should we call it...how about granularity? Simply too many reflections, variables, impossible to locate speakers that way, etc. You can't get a perfect acoustic sum unless you had two speakers in a perfectly co-incedent location, or an anechoic chamber. Not in any practical space, though. 3dB would be much more like what you'd get.

To measure acoustic power you'd have to take into account not only sound pressure but the volume of air moved, and you'd end up with a figure something like a fraction of a watt per square meter, or something. Not really practical or useful. And, from a marketing standpoint, the numbers would be really small! Speakers aren't very efficient, and once you move away from them the power per square meter is milliwatts at best.
Go to the top of the page
+Quote Post
2Bdecided
post Sep 7 2012, 10:18
Post #18


ReplayGain developer


Group: Developer
Posts: 5059
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



It's even more complicated. If L and R are perfectly out of phase, the electrical sum is zero. The real world acoustic sum is nothing like zero!

Also, see M3 vs M6 metering (which implies mixing levels) here...
http://en.wikipedia.org/wiki/Peak_programm...eo_British_PPMs

There is no single simple right answer to this. By doing the sums for a couple of in-phase sine waves you can "prove" the correct answer, but there are circumstances where it's not really applicable (including some real recordings with stereo microphones).

Cheers,
David.
Go to the top of the page
+Quote Post
dc2bluelight
post Sep 7 2012, 13:18
Post #19





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



QUOTE (2Bdecided @ Sep 6 2012, 04:25) *
... it's true that damage can be caused by shorting out speaker outputs, but most headphone outputs are more resilient IME. You may be lucky or unlucky - if it damages your equipment, please don't come crying to me - but sometimes it's OK. Some equipment is built to cope with (or be protected against) direct shorts, but not necessarily coupling the two channels together.


When you tie two channels together with a Y cord or the like, you create a strange kind of short. It looks sort of like an open circuit to each output when the two channels deliver the exact same signal (lets ignore DC coupling for now, ok?). When one channel is silent, the other drives its output into the silent one, which as a load, looks like the output Z of the channel. Could (and should) be fairly low, even for headphone amps, though may not be able handle much current. So you could end up with fairly low impedance load that way. Some headphone amps are nothing much more than a barefoot opamp output. But when both channels deliver the same signal out of phase, you have essentially a worse-than-a-short condition, one output driving into another with an equal but opposite polarity signal. That will significantly increase the current they each try to source/sink, and depending on the design, cause a problem. I saw a pair of distribution amps blow each other up that way once, but they were a really unprotected output design. Even a simple build-out resistor would help save the day.

In practice there are very few times when just tying two outputs together for a mono sum causes a problem, but there are a few. It's not really best practice, but usually works anyway.
Go to the top of the page
+Quote Post
RobWansbeck
post Sep 8 2012, 13:40
Post #20





Group: Members
Posts: 23
Joined: 24-January 10
Member No.: 77468



QUOTE (dc2bluelight @ Sep 7 2012, 04:24) *
...except... acoustic measurements are usually done by measuring SPL, where we're back to "pressure", which behaves like voltage (also a form of pressure). So no, acoustically you sum pressure, so you'd expect a theoretical 6dB.


Except ... loudspeakers don't sum that way. If they did you would have a perpetual motion machine. If you have a loudspeaker giving 1W of acoustic power and you add a second also giving 1W then according to your calculations you would have 4W of acoustic power.

I am aware that there are many other factors involved but even in the free-space low frequency condition where phase issues can be ignored things still don't add up.

If you add two voltages together you will also get twice the current hence 4 times the power. With a typical loudspeaker the acoustic load is negligible so adding a second sound field doesn't make the cone move any further.
Go to the top of the page
+Quote Post
dc2bluelight
post Sep 9 2012, 14:50
Post #21





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



QUOTE (RobWansbeck @ Sep 8 2012, 07:40) *
QUOTE (dc2bluelight @ Sep 7 2012, 04:24) *
...except... acoustic measurements are usually done by measuring SPL, where we're back to "pressure", which behaves like voltage (also a form of pressure). So no, acoustically you sum pressure, so you'd expect a theoretical 6dB.


Except ... loudspeakers don't sum that way. If they did you would have a perpetual motion machine. If you have a loudspeaker giving 1W of acoustic power and you add a second also giving 1W then according to your calculations you would have 4W of acoustic power.

I would agree with this, but for measurements we are using a sound pressure meter not a sound power meter. An SPL meter measures sound "pressure" at a theoretical point in space. If we used a sound power meter it would have to take into some unit of area or volume for a figure measured in watts (or milliwatts) per meter squared. An SPL meter is therefore more like a volt meter. Your electronic analogy would work if we could measure total radiated power, but there are no tools for that. For practical measurement we only have SPL, which is pressure at a point.

In electronic measurement you can calculate power because you have measures of total voltage and current. You can't do that once the voltage (pressure) and current (air movement) enters a three dimensional space. I'm not disagreeing with your electrical formulas, they are correct. But while you can meter total voltage and total current to get total power, once that power is translated to a radiated wave you now have power over an area, and your formulas don't take that into account, and it's not what we are measuring. It may not matter to the OP anyway, as we've moved from the practical to the theoretical, and in this case the theoretical is pretty much moot. He's looking for a mono sum in headphones.
Go to the top of the page
+Quote Post
dc2bluelight
post Sep 9 2012, 15:05
Post #22





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



Let me add this example. For an electrical sum only, if you had two perfectly identical waves, we all agree summing them results in a 6dB (slightly more) voltage gain. If we electrically sum completely uncorrelated waves, like two different noise sources, we get an approximate 3dB voltage gain. Good so far? So, if in an acoustic space we have two sources playing two uncorrelated noise sources, and measure the resulting SPL using an average of several points in space all equidistant from the radiators, we'd get a 3dB increase over a single radiator. 3dB increase in SPL, right? Now if we could (and we can't) present perfectly a perfectly correlated signal from two drivers, what would the result be? Still 3dB, or 6dB? It has to change, it's now correlated, so 6dB.
Go to the top of the page
+Quote Post
dc2bluelight
post Sep 9 2012, 17:32
Post #23





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



Here are a couple of links to SPL calculators that do SPL addition you might find relevant.

This one lets you choose between non-coherent and phase coherent sources.
http://www.doctorproaudio.com/doctor/calculadores_en.htm

This one is essentially the same, but makes note that the calculator assumes non-coherant sources.
http://www.pope-pro.com/calculator/spl_addition.htm

If you don't want to play with them, the result is two equal coherent sources add at +6dB, two non-coherent sources add at +3dB. But don't take my word for it...
Go to the top of the page
+Quote Post
RobWansbeck
post Sep 10 2012, 12:24
Post #24





Group: Members
Posts: 23
Joined: 24-January 10
Member No.: 77468



QUOTE (dc2bluelight @ Sep 9 2012, 14:50) *
Your electronic analogy would work if we could measure total radiated power, but there are no tools for that.


Actually my analogy does not work. blush.gif

We can't directly measure radiated power but we can calculate it. In the ideal case with 2 loudspeakers close together the increased area will lead to an increase in the radiating efficiency such that the radiated power will increase in line with the SPL. So, as you correctly said, the SPL will rise by 6dB and so will the radiated power.

I can only guess that with a typical stereo speaker placement you get a lesser vector addition which some people have felt the need to correct for.

I can remember working on corrective circuitry some 20 odd years ago but it now seems that I can't remember what we were correcting for unsure.gif

so apologies for taking up your time.
Go to the top of the page
+Quote Post
dc2bluelight
post Sep 10 2012, 18:13
Post #25





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



QUOTE (RobWansbeck @ Sep 10 2012, 06:24) *
..so apologies for taking up your time.

Thank you, and my compliments for your humility, a rare quality in forums. But no apologies necessary, the exercise was worth the effort and its own reward. Hopefully it all benefitted more than just the two of us.
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 26th July 2014 - 20:25