frequency separator required

[virendra]

I have been trying to make surround audio files in ac3 or dts for long time. I found steeve thompson plug in very useful for this. Now I want to go further and want to have separate frequencies in my front and rear channels. Normal equalisers or band pass are not giving the desired result that can satisfy me.

Is there any solution for this. I want separate sound streams from every speaker like only treeble from rears and only bass from fronts and when played together should give me the final music.

Pl guide me.

[Cornucopia]

You say you want treble in one and bass in the other, but that EQ or bandpass filters don't satisfy you, but that is precisely what those filters do - separate the frequencies. If that's not working for you, you need to be a little clearer about why they aren't working for you and about what you REALLY want.

BTW, in my experienced opinion, frequency separated fake multichannel is not even worth the effort. It "smears" the locations of objects. Listen to the Beatles' "I am the Walrus" right at the point where it goes into "sitting in an English garden...". Even they don't get it right.

Scott

[magillagorilla]

Virendra,
Like Cornucopia said, EQs, bandpass, and notch filters are the tools for the job. I've done this kind of thing but my source was 4 channel. I made 5.1 by creating a mono of the front 2 channels and cutting everything above 100k for the sub channel. The same for the center, only notched for midrange. It sounds good, but I was working with 4 channels to begin with.

You arent getting the results you want, as in frequencies you are filtering out are still audible? Maybe you aren't using the filters properly or you have a PAN problem and there is some cross talk between your channels.

[Compo]

What are you trying to do with the content and what effect are you after.
It seems to me that you are looking for something specific to create an effect of some sort?

Try a surround sound reverb. with a good one or with EQ on each output you can create high tingles and fizzes at on position and boomy ambiance at another.

We use this in alot of live theater productions and the way you describe your problem is the way that directors oftenask us to create a certain ambiance or effect.

[jman98]

You say you want treble in one and bass in the other, but that EQ or bandpass filters don't satisfy you, but that is precisely what those filters do - separate the frequencies. If that's not working for you, you need to be a little clearer about why they aren't working for you and about what you REALLY want.

BTW, in my experienced opinion, frequency separated fake multichannel is not even worth the effort. It "smears" the locations of objects. Listen to the Beatles' "I am the Walrus" right at the point where it goes into "sitting in an English garden...". Even they don't get it right.

The public finally agreed for the most part that colorizing old black and white movies was a bad idea, yet the idea persists that adding extra channels that weren't there in the beginning to audio makes it better.

Not to digress, but there is a reason for "I Am The Walrus". It was decided to mix some live radio into the song so the song was mixed live with the radio being fed into the mix. Producer George Martin mixed the song live to mono because he was most comfortable with that. The live radio can be heard best towards the end of the song when you hear some excerpts from King Lear ("Oh untimely death!" etc.). The first half of the song had no radio mixed into it so it could be mixed into stereo from the original 4 track tape. Since the live broadcast only existed on the mono mix the 2nd half of the song had to be put in "fake stereo" for the stereo mix. All I can tell you is that at the time consumers were not very educated and fake stereo was acceptable to them. The fake stereo was accomplished by taking the 2nd half of the live mono mix and using some kind of blunt and fairly primitive pass filter it sent the lower range to one side and the higher range to the other side of the stereo mix. The fake stereo was attached to the first half true stereo mix.

A few years ago on the "Love" CD release George Martin and his son Giles Martin (mostly Giles I'm sure) were able to produce a true stereo mix of "I Am The Walrus" for the first time. I believe that the original radio broadcast was recorded by someone (BBC maybe) and this recorded source enabled them to remix the song into stereo for the entire track since they could mix the entire performance into stereo and add the radio broadcast into that mix.

[Cornucopia]

Thanks, jman98! I'd like to hear that new mix.
What you said about the setup makes sense, although I think there was some fudging done. With the equipment they had available at the time, they could have mixed the mono broadcast in with the stereo tracks without resorting to all-mono-izing, just by using a pre-mixer-mixer, or by "track bouncing" (they did that all the time in those days).

BTW, you probably know that I am in a minority which thinks that "modernizing" a media piece (whether Silent ->Sound, B/W->Color, 2D->3D, Mono->Stereo, Stereo->Surround/Multch/3d/binaural) is not necessarily a bad thing at all - as long as an original is preserved,restored, and available at the time also, and as long as it's appropriate for the piece, and as long as GREAT PAINS are taken to make it natural and seamless. That's where lots of those past and present money-making propositions go sour - they just want to use a cheap sledgehammer instead of a series of expensive scalpels, and they don't do any research.

@OP,
Mono to stereo is VERY hard. Stereo to Multichannel is hard, but not as much (you've got more to work with). There are certain processes which can extract ambience, and isolate elements in the mix (isolation, while maintaining quality, is usually the hardest part). Once you extracted/isolated those extra things, a true Multichannel audio editor/mixer app is the best tool to finish with.
Haven't heard from you since I replied that 1st time...do you want to continue? We can guide better if given more info.

Scott

[magillagorilla]

Sony Sounforge has a plugin that allows you to "sample" a space and then use that sample as a stereo filter. You have to set up a stereo mic and a white noise generator in the space to be sampled and record some clips. Soundforge uses the sample as an image to generate stereo, I guess using frequency delays and magic. As I understand this can create a very convencing stereo image from a mono source. I don't have that plugin, my Soundforge is a less expensive version.

[virendra]

I would like to give little more details of my need.

When I purchased my home theatre system the dealer gave me an audio deme which was an indian classical instrumental music. I was asked to sit at the centre of the room with my eyes closed and when the music started the effect was amazing. I could clearly feel every instrument being placed at different place and being played from there.

This is something I am looking at. The dealer had no idea about technicallities of this and was pretty sure that there needs to be a special equipment to make something like this and this can not be made on home computers which I do not believe.

Only one thing is sure for me after lot many trials is that the source has to be wav file and not mp3. I do not know about FLAC

I tried some music with steeve thomson plug-in and the results were better but not near to something I had experienced.

This could be my presumption that these effects can be generated by sending separate frequencies to the fronts and rears and even to left and rights but i am looking something which gives me a better control on this .

I hace samplitude, soundforge and nuendo.

So i feel on software front I am well equipped but what i need is the guidance if I am missing something.

[magillagorilla]

Sorry, I'm still lost. This "sample" music at the store, was it played on a 7.1/5.1 system. What was the hardware/speaker arrangement? Also was the source/player an SACD or DVD Audio disc recorded in multi channel. You are trying to get some kind of surround sound from 2 channel.

I've auditioned some VERY high end 2 channel systems that sounded "3D". They filled the room with an extremely realistic sound stage. This was the result of a perfectly designed sound room and very high fidelety equipment and a pristine stereo recording. This arrangement had nothing to do with software, aside from the recording/mastering process of the CD. I could take that same CD, play it in my car and not get the same effect.

The only way to cheat 2 channel sound in to multi channel immage is to use a digital imulation like the one that comes with the expensive Sound Soundforge. When I was a musician I used to use the microphone/room simulators in Cakewalk. These tools do the same thing.

Stereo sound happens when the left and right ears perceive sound at various delays from a single source or they receive unique sounds individually. Like the bass guitar in the left ear and the drums in the right ear. In a real environment there is still crossover, as the bass guitar, even hard stage left, will still be herd, with a slight delay, by the right ear. The subtile delays between the ears is what allows us to perceive from where the sound originates.

In a perfect world it is possible to create a realistic "3D" sound stage using 2 sound sources, because afterall, we only have 2 ears. Like my super high end stereo example. Unfortunately the technical and physical limitations of sound reproduction, speakers, amplifiers, and source recordings are crude compared to real living sound and therefore register as fake to the human ear. To get around this, we have added more speakers to fill in the gap. Multi channel audio forces the spacial difference of sound.

What you are trying to do is not unlike the process of turning a 2D movie in to a 3D movie. Objects do pop out but they look like paper cutouts rather than real volumous objects. There is only so much you can do to interpolate data that is not there.

In sound imageing frequency seperation is only 1 tool. You need to consider delays between the left, right, front, rear and furthermore variable delays for specific frequencies. For instance, high frequencies can travel faster than low frequencies. If you want the sound to start at the front stage and reverberate from the backstage, the high frequencies will bounce back before the lows. The greater the delays between the frequencies may add to the illusion of the depth of the "space".

[Cornucopia]

@virendra,

(Long explanation...)

A sound such as you heard in the showroom isn't THAT hard to recreate, it just takes the right kind of equipment with the right layout, a couple of good pro-type softwares and computer, and planning and imagination and understanding of how this all works.

There are likely 2 paths that could have been taken to get you that "wonderful demo":
1. The "Realist" approach - Where the performance is recorded "as-is" by a pair of microphones (in Binaural, or one of many stereo patterns) or by a Soundfield mike array. This gets played back, with as LITTLE intervening processing/electronics as possible, directly to the ears via headphones. OR, if speakers are to be used, crosstalk cancellation processing is applied, then it's played back through the (2) speakers.

2. The "Formalist" approach - Where the performance is recorded one-to-one mike vs. instrument with multiple microphones into a multitrack recorder. Then it's mixed into either Multichannel (5.1/7.1) or Stereo (2) channel. Multichannel would be played through a Surround speaker system, Stereo would be played either through Stereo system speakers (2), or through a Multichannel system with the sound coming from LF and RF, or through Multich. with Ambience Processing to multiple speakers. For the stereo, Crosstalk Cancellation could be done here as well.

(There is a 3rd - HRTF/Binaural/Soundfield processing of Multi-mono signals to place into a virtual soundscape. A blend of the 2 above)

@magillagorilla,

I say Crosstalk Cancellation a number of times, particularly because with either method 1, or with meticulously- and imaginatively-engineered modern versions of method 2, it is possible to apply such processing that the normal interaural crosstalk, always present with stereo sound, can be reduced or (nearly) eliminated. Once that happens, it's a whole new ball game for stereo.

I've been a proponent of XtalkCancellation (XTC for short) methods for over 20 years and have even created alternate speaker layouts to enhance this effect (patent applied for). But most common public examples of this have been around for years and work in one of 2 ways: signal preprocessing (Carver Sonic Holography, SRS) or Acoustical/Placement (Polk SDA Reference speakers).

Both work by using an inverted & delayed feedback loop. The phase inversion is to do the cancelling and the delay is to align the signal up with the Xtalk signal in the opposite ear/channel.

Acoustically, it looks like this:

[-R]<.....>[L]_______________________________________[R]<.....>[-L]

in version 1
or, this:

[L-R]<.....>[L]_______________________________________[R]<.....>[R-L]

in version 2

***Note: the distance between the arrows is the distance between your ears (~21.5cm)
And both of these show a speaker array that the listener would be looking at (hopefully in the triangular/equidistant sweet spot - although the sweet spot is MUCH better with these, esp. v2)

Version 1 nearly restores the speakers back to a "headphone" effect, except with HRTF tonalities being different and a slight "spaciousness" due to the extra delayed Xtalk signal which is incorporated into the Haas effect.

Version 2 maintains some "speaker-ness" with the additional signal. Again the extra-delayed Xtalk signals are subsumed into the Haas effect and are unnoticed except as "ambience".

The difference is STRIKING! Even with just a frontal array of 2 speaker pairs, a CLEARLY DISTINGUISHABLE, SOLID PANORAMA of enveloping soundfield is available in the Front 180degree arc. In Front, Beyond, Behind, and Between the speakers, sound elements are no longer vague "phantom" signals. It's often as strong a difference from stereo as stereo is from mono.

I've had my Den setup like this for 20 years (expanded to a 9.1 surround system incorporating Center, LFSub & Side & Back surround channels ~10 years ago), and have never failed to WOW any of my guests, many of which are musicians and media people.

..................

So, it can be done with less.

But it doesn't happen by just "cutting up the frequencies" and splitting them off to different paths.

BTW, this SHOULD be done with a wave file. An MP3 or other compressed file will have removed some of the small/quiet harmonics & transients that while not necessary for regular stereo, are important for accurate localization once XTC gets applied. Compressed files could be used, but you'll get less than optimal resullts.



What you want to do is GET BACK TO A MULTRACK MASTER, if you can. Once as a multitrack master, you can "remix" to multichannel audio, or "Super"-stereo.
But, getting back to multitrack is the hardest thing to do. It requires separating the INSTRUMENTS, not just selected frequencies, as many instruments have similar, overlapping frequency ranges.



My suggestion...

If the material is only available in stereo and isn't a simple L-C-R type separation of instrument placement or contains instrument/element combinations that are similar in tonality/range, you won't be able to do ANY reasonable mutlitrack isolation.
Then, your best bet is to do (in no particular order - pick one):
A. Ambience Extraction & Enhancement (can be done automatically by most Home Theatre amps)
B. Dolby-Surround Quasi-extraction (like your Steve Thompson plugin)
C. XTC (the crosstalk cancellatin I was referring to above)
D. Leave it ALONE and live with it as is.

Scott

[virendra]

Dear Cornucopia,

You are amazing and very deep. Unfortunately I do not have as much knowledge of this area. Sound processing is my hobby and I do not try to make money by this. It is only for my own satisfaction, so I have not taken any official training on sound engineering.

I have a Yamaha A/V receiver with Yamaha speakers. I try various softwares to find what processing gives good results for me. Most of the times after doing so mauch work I get negative results. So I thought of getting some guidance.

The file I am using is a wav file from an audio cd and the demo I got from the dealer also was an stereo audio cd. He had no idea of existance of dts cd till I gave him one made by me.

I agree that the room where demo was done must have certain parameters which I was not aware of. Secondly this is difficult for me as my room is very small and I can not go for any high end system as they are damn expenssive in India. I have to manage within what I have. Unfortunately my audio system and my computer are not connected as they are at two different places. So every time I burn the cration on a RW CD or DVD and play. So there is a delay and editing becomes time consuming. That is the reason I was trying to look for something like steeve thompson plug-in.

If you can give some method to create a 5.1 mix out of a stereo wav it will be great for me. ( I have seen many guides in Doom 9 forum and not too happy about the result)

[magillagorilla]

cornucopia, I think I was saying some of the same stuff about crosstalk, only I perceived it as sound bleeding over from left to right. I didn't know the technical term thanks. For that matter, if you have the original multitrack recording of something, 1 isolated element (voice instrument) per track, the sky is the limit for mixdown.

Also, I have Sonic Holography on my Carver and I love it.

[Gavino]

A few years ago on the "Love" CD release George Martin and his son Giles Martin (mostly Giles I'm sure) were able to produce a true stereo mix of "I Am The Walrus" for the first time. I believe that the original radio broadcast was recorded by someone (BBC maybe) and this recorded source enabled them to remix the song into stereo for the entire track since they could mix the entire performance into stereo and add the radio broadcast into that mix.

Thanks, jman98! I'd like to hear that new mix.

Scott, if you haven't got around to buying the "Love" CD yet, there is a copy of the remixed "Walrus" on youtube:
http://www.youtube.com/watch?v=_Myr3P-H2QE

Audio is 112kbps AAC, but you can hear the difference from the original.

[Cornucopia]

Dear Cornucopia,

You are amazing and very deep. Unfortunately I do not have as much knowledge of this area. Sound processing is my hobby and I do not try to make money by this. It is only for my own satisfaction, so I have not taken any official training on sound engineering.

I have a Yamaha A/V receiver with Yamaha speakers. I try various softwares to find what processing gives good results for me. Most of the times after doing so mauch work I get negative results. So I thought of getting some guidance.

The file I am using is a wav file from an audio cd and the demo I got from the dealer also was an stereo audio cd. He had no idea of existance of dts cd till I gave him one made by me.

I agree that the room where demo was done must have certain parameters which I was not aware of. Secondly this is difficult for me as my room is very small and I can not go for any high end system as they are damn expenssive in India. I have to manage within what I have. Unfortunately my audio system and my computer are not connected as they are at two different places. So every time I burn the cration on a RW CD or DVD and play. So there is a delay and editing becomes time consuming. That is the reason I was trying to look for something like steeve thompson plug-in.

If you can give some method to create a 5.1 mix out of a stereo wav it will be great for me. ( I have seen many guides in Doom 9 forum and not too happy about the result)

That's why I gave that last bit about the few things you could try.

Believe it or not, I also have a Yamaha HT Amp and speakers (well, some of them anyway, my 5 Fronts are another brand Larger, full-range models). The Yamaha does have the "Ambience Extraction" facility in it's various "DSP Modes".
However, since I'm guessing you've tried that (as you own the device yourself) and since you've also tried the (pseudo) 2.0 -> 5.1 conversion (either with the HT amp's feature or with the Steve Thompson plugin), all that's left is to either apply XTC or extract the various elements back to multitrack and remix. Both of these can vary in success greatly depending upon how well or badly the original stereo was mastered. Since I don't know the material, I won't comment on it.

Be warned:
The topic of extraction of elements BACK to Multitrack contains enough material for a year or 2 of post-graduate college. And usually takes just as long to execute.

Ultimately, my guess is that if you're not already satisfied now, you may not be with XTC, so I'd leave it as is.
If you really want to try your hand at XTC or a few other sleights-of-hand, look at this link: http://www.savioursofsoul.de/Christian/vst-plugins/stereo-plugins/.

HTH,

Scott