3D imaging pre-amp

rlyach

New member
I may be getting in over my head but I was wondering if it would be possible to incorporate a 3D imager into a pre-amp circuit. I have been enjoying my Crack headphone amp, but I have noticed that because my right ear only hears the right channel and my left ear only hears the left channel, the stereo image is between my ears. My setup (Crack with RCA clear top 12AU7 and a tungsol 5998, and a pair of Tesla T1s) gives a nice large soundstage but I want the image in front of me. After a little thought I discovered the reason. So... I wanted to try an experiment. I wanted to build a preamp using a 12AU7 for each channel. One half would be for the inputting the main signal, the other half for sending the signal to a cross-feed circuit with a Bridged Transformer delay circuit tuned to about 60 degrees. Then sum the delayed right channel into the left signal by way of an attenuator and coupling capacitor. The attenuator would be stereo, one half for each cross-feed signal. The output stage would be a capacitively coupled cathode follower. With the attenuator controlling the cross-feed signal, I should be able to move the sound stage close or far away, simulating speakers. Is this a waste of time. I know there are several DSP options out there but they are all set up for surround sound processing. I simply wanted an analog cross feed solution for simple stereo.
 
PB,

Thanks so much... This is exactly what I was looking for. I knew what I wanted to do but wasn't sure where to begin. I like the passive approach since my DAC can drive pro audio levels. This will give me something fun to try.
 
Well... I did some pretty heavy research. I analyzed the classic cross-feed circuit (below) and found some fundamental issues. For instance, in order to get the proper delay (0.5ms to 1.0ms) you need to use a resistor / capacitor filter which also acts as a low pass filter with a cutoff frequency of 585 Hz. This translates to more than a 30db drop at 20Khz. Consequently, only the low to mid frequencies are cross-fed. To compensate, the designer of this particular circuit put a high pass filter on each side, boosting the primary signal. This is an artificial solution since these frequencies are not cross-fed into the opposite channel. I discovered that the real difficulty in designing this type of circuit is how to generate a long and accurate analog delay. I experimented with the shunted transformer circuit but there is a limit there as well. The longest delay you can get with a max frequency of 20Khz is 25 microseconds. This is not enough delay to even simulate the difference in sound across the separation of your ears. Rather than trying to find a circuit solution, I evaluated a digital plugin. I used HiJack Pro and Redline Monitor to simulate near field monitors on my headphones. While the sound was a little more forward, it sound more focused and less open. It also seemed that I lost some fullness and dynamic range in the low end. After all this I decided to just enjoy my Crack and Tesla T1s just as they are. If anyone has another solution for a full and open forward soundstage please let me know. I do like to experiment.
 

Attachments

  • crossfeed.jpg
    crossfeed.jpg
    116.4 KB · Views: 532
Well... I had some time on my hands so I actually built the Soundfiled Simulator from PB's link. I made some slight modifications to the circuit based upon recommendations I found. I also moved the POT to allow the actual soundstage to be adjusted. I also made use of my Bottlehead experience in designing the chassis. I was not expecting much but when I plugged it in and adjusted the soundstage, I was very surprised that to my ears the filter actually opened up the sound. It did not move the sound very far forward at all but it definitely was more pleasing. There is a 4.5db loss with this circuit so it also pads the source slightly. At this point I will be using this little box and giving it a more critical listen. Thanks PB for the link. In case anyone else wants to try this little circuit, I have posted a few picts on how to build it. I also have top and bottom plate layout files and the circuit board layout as well if anyone is interested. Enjoy!

Update: I updated the schematic to the actual values I used. I found an error in the original post.
 

Attachments

  • SF Inside.jpg
    SF Inside.jpg
    394.3 KB · Views: 500
  • SF Complete.jpg
    SF Complete.jpg
    409.5 KB · Views: 408
  • New Office Crack.jpg
    New Office Crack.jpg
    373.4 KB · Views: 692
  • crossfeed.jpg
    crossfeed.jpg
    250.8 KB · Views: 763
OK, Perhaps someone can enlighten me. My engineering analysis says this thing should be marginal at best, but it sounds fantastic. Now that I have had more time to play with it, I can honestly say I love the sound. It is indeed more forward. Not only that but there seems to be added clarity, which is exactly opposite of what I would expect. According to simple analysis and verified through simulation, this thigh should be a little darker but that is definitely not the case. I even tried it on a different amp with similar results. I had two colleagues give it a listen and now they both want one. One has already asked for the circuit and plans on building it. The engineer in me keeps looking for the reasons why. If anyone has a little time, perhaps you could reproduce this circuit and let me know what you think.
 
I decided not to question success.  When Carver Corp. came out with the Sonic Hologram I knew what it was immediately.  It feeds some L out of phase to R.  It feeds some R out of phase to L.

Years later I bought a C9 cheap on ebay.  It does all sorts of things within the soundstage.  A few live recordings have ambient sounds from way behind me.  It is not usually in my system but I drag it out when a recording strikes me that it would benefit from it.
 
Thanks for the info Granger. I seem to recall several times where PJ would mention the apparent discrepancy between engineering judgement and sonic perception. I am pretty much a skeptic on some things but I can't argue with direct experience. Your comment reminded me of one thing I forgot to mention. You stated that some recordings benefit from this little trick more than others. I have found the same thing, but there is a slight difference. I modified the circuit slightly and inserted a POT in the phase delay portion of the circuit. With this change you can adjust the delay for maximum effect. I found that some recordings like a large delay while other a smaller one. All-in-all, I like the result. Cheers.
 
I seem to recall several times where PJ would mention the apparent discrepancy between engineering judgement and sonic perception. I am pretty much a skeptic on some things but I can't argue with direct experience.

Yeah and this is not really an indictment of engineering as much as it is another example that we need to keep looking for new ways to measure these things that we are hearing. Like why different wires sound different. It makes no sense based on the way we measure things. But we hear the difference, so we need to figure out what measurement we are missing that will help us to understand what is going on.
 
Great explanation DOC. I once heard a lecture by Dr. David Menton in which he said that the human ear is so sensitive that it can detect the pressure change from simply standing on a piece of paper. Apparently our ears are more sensitive than some of the crude measurements we make, or like you say, we have yet to measure the correct thing. One more thought. I have had the pleasure of visiting a few recording studios and I was amazed at what was used to cobble together signal paths from microphones to the actual recording equipment. Not to mention the patch bays.
 
I'd like to experiment with this, too, for my SEX amp, but I can't read a schematic. How cheap is it for parts, overall?

EDIT: I noticed that there's a feature in DSP output for J.River Music console called Crossfeed that I think does exactly this. It's pretty remarkable how effective it is.
 
The cost of the parts is about $25, not including the enclosure. I used all 1% metal film resistors and Mallory Coupling caps. As always, you can upgrade the parts. Here is the BOM
 

Attachments

  • BOM.jpg
    BOM.jpg
    143.1 KB · Views: 786
At low frequencies, sound goes around the head to both ears, with a time delay difference that is perceived and with little difference in loudness. The RC lowpass filter creates the delayed cross feed.

At higher frequencies, the ear/brain becomes less sensitive to phase differences, and the head blocks more of the sound from the far side. The ear/brain at high frequencies thus relies more on loudness differences to detect location.

Different recording microphone techniques have a huge effect on how and whether this works.
 
Thanks Paul, Now I know why I was confused. I assumed that since low frequencies are omnidirectional that delaying the high frequencies is more important. With your explanation on the psychoacoustics of the stereo image, the circuit makes perfect sense. So when they say that bass is omnidirectional, does it mean that the brain does not use loudness to locate the sound like it does with higher frequencies? It is rather convenient that delayed low frequencies make this work. Hence you can use the simple RC low pass filter.
 
Yeah, you have the concept correctly.

I think it's evolution, not a coincidence. Since loudness does not work well at low frequencies, there is a survival advantage to developing phase sensitivity in that range. It's actually quite remarkable that we can detect phase effects up to 1500Hz or so - most neurons are not usually anywhere near fast enough for the task. I believe the phase/timing information is retained all the way into the brain, not decoded at the first layer or two. But I read about this a long time ago so there may be more knowledge, ort my memory may have gotten some of it wrong.
 
Here is the board layout, a picture of the placed components and the backside of the board. I inserted the components from the top, soldered them in place on the bottom, bent the leads together to form traces, and inserted one lead back through the board for a top connection on all nodes that need to connect to the RCA jacks, switch and POT. The node names match the schematic. There is also a simple instruction file. Enjoy! I will be away for a couple of weeks.
 

Attachments

  • Instructions.pdf
    Instructions.pdf
    4.7 MB · Views: 240
  • plate layout.jpg
    plate layout.jpg
    90.8 KB · Views: 169
  • board bottom.jpg
    board bottom.jpg
    278 KB · Views: 173
  • board top.jpg
    board top.jpg
    232.5 KB · Views: 181
  • Layout.png
    Layout.png
    58.4 KB · Views: 179
Finished mine today!
Caps still to burn-in (pio matched russian ones...lot of time ahead!), but hey, the effect is incredible!!
I put this little wonder in the same case of my passive BSC circuit (Paul Joppa's open source circuit found here http://www.t-linespeakers.org/tech/bafflestep/passive-BSC.html ).

A killer add-on to every headphone setup!

Bottlehead crew, what about throwing a kit for the Quickie+Quicksand users with both those simple circuits?  ;)

Thank you again for the scheme rlyach!
 

Attachments

  • WP_20151025_005.jpg
    WP_20151025_005.jpg
    364.4 KB · Views: 129
  • WP_20151025_006.jpg
    WP_20151025_006.jpg
    407.8 KB · Views: 146
Back
Top