ITC Bridge Forum: Stream Capture Filtered with EVPamp

[EVPDave]: I have been experimenting with an old filter app from early 2005 called EVPamp. It was originally designed to filter voices from noise sources. Here is an example from today using the software. The output of the filter is a bit harsh so I remix about 25% of the original file back with the filtered file. This seems to help with a more natural sounding result. EVPamp was abruptly discontinued in late 2005. You may notice some overlay of voices at times. This appears to be a result of the mixing process. The source was the current forum stream recorded directly to computer. "The children of the blind see light" Attachment: 062312thechildrenoftheblindseelight.mp3 (Downloaded 1638 times) [EVP-ITC Research and Development]: Very clear!! Looking forward to working with this program! Jayne [Deborah]: I don't hear it exactly as you, however; do like the clarity of the sound. Thanks Dave [EVPDave]: Deborah wrote: I don't hear it exactly as you, however; do like the clarity of the sound. Thanks Dave Thanks Jayne & Deborah for the feedback. Yes, the smoother sound is easy on ears for a change. Here was another from that same session. "Listen to my filter" Attachment: 062312listentomyfilter2.mp3 (Downloaded 1564 times) [EVPDave]: Here is another capture from the same two minute session. I post it more out of wonder than trying to hear any words. What I hear is a harmonic melody of sorts. I know Keith is using a blend of Spiricom tones and white and pink noise

on the stream. That being said I hear a distinct pattern that should not be derived from that type of transmission. Perhaps someone musically trained can decode the musical sequence, a hidden message in the notes? Or, am I just hearing "Night Music" the stuff of dreams? Attachment: 062312musicalwords.mp3 (Downloaded 1550 times) [EVPDave]: This was from a different stream that uses looped noise. "Can ya hear me now...Debbie" Attachment: 062612canyahearmenowdebbie.mp3 (Downloaded 1714 times) [Keith Clark]: Hi Dave, I'm pretty intrigued on what this software actually does, enjoyed listening to your files, first was the clearest. I heard it as "the children of the blind they see the light" I feel that between ("they" and "see") they were thinking "finally", the result was sort of a combination of "they finally" and "they see". Um, that was sort of an er, impression, shall we say. Looking forward to more Keith [EVPDave]: Keith Clark wrote: Hi Dave, I'm pretty intrigued on what this software actually does, enjoyed listening to your files, first was the clearest. I heard it as "the children of the blind they see the light" I feel that between ("they" and "see") they were thinking "finally", the result was sort of a combination of "they finally" and "they see". Um, that was sort of an er, impression, shall we say. Looking forward to more Keith Thanks for listening Keith. The results vary from day to day. Some days clearer then others. Here is a clip from today. I left it

long cause there is a lot of chatter. I am hoping the creator of EVPamp will come forward again. I have a few questions I didn't ask back in 2005. Dave Attachment: 070312stremchatter.mp3 (Downloaded 2583 times) [Deborah]: Hello...I think I hear, at the very beginning a question..."what's happened to the _______ ?" I do like the sound, pace and tone of this application...must be "time" to get back to the "stream"... Thanks Dave. Happy Independence Day [EVPDave]: Deborah wrote: Hello...I think I hear, at the very beginning a question..."what's happened to the _______ ?" I do like the sound, pace and tone of this application...must be "time" to get back to the "stream"... Thanks Dave. Happy Independence Day Thanks, the most I hear the voices described as is a 'robotic burble'. Today was odd. I was getting names, mention of the heatwave in the area and my first 'hate' message since I began recording in 2004. "We hate Human, which you are" Attachment: 070612wehatehumanwhichyouare.mp3 (Downloaded 1544 times) [EVPDave]: Here is a name that can be heard at the start. Other interesting chatter follows. "Dan Matthew" Attachment: 070612danmatthew.mp3 (Downloaded 1533 times) [EVPDave]: Same session as the last two posts, another name. "Ed!.....it's Ed" Attachment: 070612editsed.mp3 (Downloaded 1563 times) [Keith Clark]: I'd really like to know what this program does behind the scenes, I think we can find some other applications for it, I'm sure! In Stream chatter clip I hear the following: "What's happened today?...(unknown) ..learn to be patient.....would it.....(foreign language)....watch

I'm gonna slide my head....into the picture." Keith [EVPDave]: Keith Clark wrote: I'd really like to know what this program does behind the scenes, I think we can find some other applications for it, I'm sure! In Stream chatter clip I hear the following: "What's happened today?...(unknown) ..learn to be patient.....would it.....(foreign language)....watch I'm gonna slide my head....into the picture." Keith Hi Keith, From the few communication with the original EVPamp developer back in 2005 a few fragments of information are available. "The filter is similar to a Kalman filter or a Hidden Markov Model based filter". So it is filter based in design and not jumbling, chopping or adding additional info like other programs and 'talkers' out there. In running many tests the best response has been with using high frequency noise or even video sweeps converted to sound. The current stream here on ITC Bridge works very well. This is now what some may be calling the 'fog face' mode. Here is an example from yesterday. "A/the man has an announcement!" P.S. More info about these types of filter designs http://www.cs.unc.edu/~welch/media/pdf/kalman_intro.pdf Attachment: 070712manhasanannouncement.mp3 (Downloaded 1555 times) Last edited on Jul 8th, 2012 11:16 AM by EVPDave [Keith Clark]: First word may alo be "Japan" Well if the stream works, I think we can provide something cleaner....to be honest I'm too lazy to use a dos based program to input/output files, so I give you credit for that. I'm really into live, real-time. I'm sure we'll think of something

new to try for the input...I'm still amazed that is what you are getting out of the stream....when the original source is nothing but noise and spiricom tones that I reverse engineered.....there is no human voice possible...now if we could filter live with that program, now that would be something!! Keith [EVPDave]: Keith Clark wrote: First word may alo be "Japan" Well if the stream works, I think we can provide something cleaner....to be honest I'm too lazy to use a dos based program to input/output files, so I give you credit for that. I'm really into live, real-time. I'm sure we'll think of something new to try for the input...I'm still amazed that is what you are getting out of the stream....when the original source is nothing but noise and spiricom tones that I reverse engineered.....there is no human voice possible...now if we could filter live with that program, now that would be something!! Keith Seems to work with any noise I can generate including impulse noise from a Tesla coil. I don't know of a way to process with EVPamp live. It might be possible with a large buffer and scripts to link the files together but it would not be real time. The original developer was getting requests for a live system in 2005. He envisioned a separate DSP chip running the EVPamp algorithm. Unfortunately, that is the key that is missing right now. I may try my hand with some of the academic Kalman Filter library's

now available but this is not my forte. Another example attached of the chatter I'm getting with EVPamp filtering the current stream. I hear something about working and a second voice mentions to "tape it" towards the end. Attachment: 070712wevebeenworkingchatter.mp3 (Downloaded 1620 times) [EVPDave]: Recorded the stream for two minutes today and processed the results with EVPamp. I hear a name and more. "Hi I'm Patrick Stimple, I'm diligent" Attachment: 071012hiimpatrickstimpleimdiligent.mp3 (Downloaded 1535 times) [EVPDave]: This capture was recorded from today's stream (07-14). Out of two minutes this was the clearest (and longest) of the session processed through EVPamp. "We're pushing great pulse with the elevator" Attachment: 071412werepushinggreatpulsewiththeelevator.mp3 (Downloaded 1353 times) [RussianEVP]: Hi, Dave. I tried the program EVPamp. Most of all I like the way it handles files with recorded silence. Very interesting touch is obtained in the form of a question the answer. DimaS [EVPDave]: RussianEVP wrote: Hi, Dave. I tried the program EVPamp. Most of all I like the way it handles files with recorded silence. Very interesting touch is obtained in the form of a question the answer. DimaS Hi DimaS, Good to read that you are trying recorded silence and getting results. Yesterday I was getting good results recording the sound of wind chimes and then running the session through EVPamp. I think a live version would be very interesting to work with. I am hoping the original developer will come forward so the development of the EVPamp filter can continue. Dave