Stereo width post processing

[Dave2002]

Are there widely available (preferably free or low cost) software solutions for adjusting the perceived width of a stereo recording?

I thought Audacity might do this, but I can't find an appropriate method in Effects in my installation.

[MrGongGong]

If you are comfortable with plugins there are lots

http://www.vst4free.com/free_vst.php?id=649

http://www.kvraudio.com/product/stereo-tool-by-flux/details

or look here

http://www.vstplanet.com

http://www.vstwarehouse.com

but with free ones you will probably have to play around with a few to find the one you like

Here's how to get Audacity to work with plugins

http://wiki.audacityteam.org/wiki/VST_Plug-ins

OR if you have a single file you need narrowing or expanding send it over and i'll do it for nowt (as long as it's NOT an Elgar oratorio )

OR (if you want to narrow the width) you can do it "manually" with Audacity by separating the stereo file into 2 mono ones (one with L and one with R) then remixing them with different pan settings (I haven't had enough espresso to get my head round the exact sequence yet but it would be very simple to do)

[Dave2002]

OR (if you want to narrow the width) you can do it "manually" with Audacity by separating the stereo file into 2 mono ones (one with L and one with R) then remixing them with different pan settings (I haven't had enough espresso to get my head round the exact sequence yet but it would be very simple to do)

Thanks very much for the tips. I'll be some time!

Re the manual method, which I'll have to check for feasibility, one presumably generates S=L+R, and D=L-R and then from those tracks generate S+kL and S-kL where k is a parameter - maybe in the range -1 to 1 (0 to 1 should do, -ve values simply reverse the channels) and then assign these to L' and R' respectively. Perhaps not being able to do these easily in real time could be a pain. Should surely be possible to do in real time either in analogue, or with a fairly simple real time program. Once a suitable value for k has been found, then batch processing faster than real time would be helpful.

I haven't had much coffee today yet, so still thinking about whether k could/should be greater than 1 for widening - possibly!

[MrGongGong]

Thanks very much for the tips. I'll be some time!

Re the manual method, which I'll have to check for feasibility, one presumably generates S=L+R, and D=L-R and then from those tracks generate S+kL and S-kL where k is a parameter - maybe in the range -1 to 1 (0 to 1 should do, -ve values simply reverse the channels) and then assign these to L' and R' respectively. Perhaps not being able to do these easily in real time could be a pain. Should surely be possible to do in real time either in analogue, or with a fairly simple real time program. Once a suitable value for k has been found, then batch processing faster than real time would be helpful.

I haven't had much coffee today yet, so still thinking about whether k could/should be greater than 1 for widening - possibly!

Erm, not sure about your manual method technique (but still a bit low on the caffeine )

If you make a mono file with just the L signal from the stereo file
and another from just the R signal
put them onto 2 tracks
set the pan for each (+ or - ) then mix these down to a stereo track

You will be able to narrow things
Playing around with phase cancellation (by inverting and mixing) is a possible way of isolating the information that is only on the L or R ... BUT you are likely to get odd "artefacts" or phase issues.

Software editors like Audition, Pro-tools etc do this stuff easily and you can do batch processing etc

[Dave2002]

Erm, not sure about your manual method technique (but still a bit low on the caffeine )

If you make a mono file with just the L signal from the stereo file
and another from just the R signal
put them onto 2 tracks
set the pan for each (+ or - ) then mix these down to a stereo track

You will be able to narrow things
Playing around with phase cancellation (by inverting and mixing) is a possible way of isolating the information that is only on the L or R ... BUT you are likely to get odd "artefacts" or phase issues.

Software editors like Audition, Pro-tools etc do this stuff easily and you can do batch processing etc

I guess Audition etc. cost "real" money, so since I'm not really in the business probably not worth it for me.

I actually think on one recording at least I want to widen the stereo stage - though it's not a disaster if I leave it alone. I've not had enough experience to know, but if I use microphones with an angle setting, do I get a wider stage with a larger angle setting, e.g 150 or with a smaller one such as 90?

I can see that 150 is bigger so one might think that would be wider, but a counter argument could be that in fact the "width" of the performancers/sound source stays the same, while the microphone sets that within its own "virtual stage", so in fact would be narrower. Some microphones are using MS techniques, so presumably simply sum/difference the outputs internally in analogue circuitry, though I'm not sure that all stereo mics do things that way. If not that, what else could they do? - really don't know!

[MrGongGong]

One question is why do you want to widen the stereo image?
There is, obviously, a huge difference if you sit in front of an orchestra (for example) and are on the front row or further back.

This is a huge subject (and I usually run out of knowledge and phone someone who knows it in relation to the particular situation) but this seemed useful

https://www.soundonsound.com/sos/1997_articles/feb97/stereomiking.html

[richardfinegold]

One question is why do you want to widen the stereo image?
There is, obviously, a huge difference if you sit in front of an orchestra (for example) and are on the front row or further back.

This is a huge subject (and I usually run out of knowledge and phone someone who knows it in relation to the particular situation) but this seemed useful
https://www.soundonsound.com/sos/199...reomiking.html

Agree with MGG. If you want a bigger sound stage, the place to look is your system, not by mutating files

[Dave2002]

Agree with MGG. If you want a bigger sound stage, the place to look is your system, not by mutating files

Maybe, but I have a very specific issue right now. I made some recordings yesterday in a church, and I now want to get a better approximation to what I heard live. This may now be impossible, but I'm willing to experiment. Some things worked well, and other things not so well, so now trying to patch various audio and video files together. A redeeming feature might also be that someone else, rather unexpectedly (to me at least), turned up with a decent video camera and a decent audio recorder, and we may be able to swap files to our mutual advantage. With possibly four audio sources and two video sources for most of the event there is hope that both a video rendition and an audio version can be compiled.

There is, incidentally, an added complication with some video systems, in that orienting the combined camera/microphone system will also rotate the perceived soundstage slightly. I guess one has to decide whether the final "product" is going to be a video one, or an audio one. I don't know of any systems for microphone which would allow rotation to be done in software - except perhaps for Michael Gerzon's and Peter Fellgett's systems. MG was keen on tetrahedral microphone disposition, and worked with PF to develop Ambisonics, which also included some psycho acoustic features I think.

Apparently the Soundfield microphone was liked by some video producers because it gave a lof of flexibility in post production stages.