42 replies to this topic

[CaveRat2]

Andy is correct about quiet recordings and the H2.

I only specified the H2 because it is the cheapest digital to meet minimum specs. (That seems to be what most want on here.) The Sony is a better recorder, although at a higher price. Personally I still use a TASCAM cassette system over digital. (4 discrete channels) And for digital I have a Fostex FR2 (list $1199 when I bought it.) And like the comparison between the Sony and the H2, I do hear a noticable improvement between the Fostex and the H2. Of course that is to be expected. But still if price is a major issue, the H2 is satisfactory, just not great.

[Yalbi]

I have found that most of the time all that is required is a pretty basic voice recorder (as long as you are able to connect it via usb to your computer) and it will do just fine. The voices still respond directly to questions and are sometimes clearer on a cheap device. In fact try reducing the recording qualty on an expensive machine to a lower bit rate and see. It works for me. They seem to like the noise wether it be analogue or digital.

[ourobouros2k2]

IMHO the noise takes away from the data integrity. I guess it is a matter of personal choice. Belief happens on so many levels, and ultimately it comes down to what constitutes personal proof and it varies with every individual.

Personally I believe that if you do not make every possible effort to record the cleanest signal possible at the highest sample rates, then regardless of what you catch it won't hold up to scrutiny.

I know that the LPCM format at high sample rates and a/d conversion is capable of recording legitmate evp. In side by side comparisons, though, what sounds like a telephone conversation with lucifer on a cheap RCA recorder is usually something a lot more benign when comparative audio is reviewed from a linear pcm recorder at higher sample rates. I personally use 48khz at 24bit, as my ears cannot really hear much of a difference between that and 96khz. For grins and giggles, I tried to locate sample rates and frequency response of cheaper RCA units, but I cannot find those specs anywhere.

Andy

[CaveRat2]

There are proven scientific reasons why low sample rates yield more "EVPs". However these proven reasons apply to false EVPs. If one considers the Nyquist point and sample rates it becomes evident that additional artifacts are generated in the voice frequencies when sampling drops near or below the Nyquist point. Aliasing and intermodulation distortion rise because the audio being recorded approaches the sample rate resulting in the sum and difference between them being in the audio spectrum.

Consider if a sample rate of 96 KBPS is used, and the frequency (third harmonic of the highest fricatives, S and T sounds, is used for reference.) would have to respond to about 3 kkHz x 3. The audio range needed would be up to 9 kHz. So we have etablished the high frequency limit of the recorder.

Now hetrodyning effects and intermodulation is taken into account. If we use the 9 kHz as a base, and a sample rate of 16 KBPS, there will be artifacts generated at the sum and difference points. Thus artifacts will occur at 7 kHz and 25 kHz. The 25 kHz is not a problem, but clearly the 7 kHz falls well within the passband of the recorder.

It gets even worse when we consider what noise does. Noise is actually many frequencies at one time. Thus noise will create artifacts in this manner over a wide range of frequencies since the dfference varies at any given instant based on what is present in the noise. These artifacts, when they vary in frequency can imitate voice at times. Add a little audio paridolia and you have a false EVP.

Now consider a higher sample rate of 96KBPS. You still have the sum and difference factors as before. But now the sum and difference become 87 kHz and 104 kHz. Both of these are well beyond the range of hearing and will not be a problem.

Thus the reasoning behind using a high sample rate. And an explanation for those who say they get more EVPs on their cheap little recorders at a low sample rate. And a solid reason why debunkers can discount most EVPs made on cheap recorders.

[ourobouros2k2]

Jim, you are very knowledgeable on these matters, I wonder if I could pose a simple question that has never been properly explained to me.

Put simply, if the frequency of an evp is inaudible at the time of recording, why should it be any more audible on playback? Since the recorder isn't frequency shifting, it should still playback as inaudible...

Todd Bates once explained that this had to do with the nyquist point. I didn't fully understand it, but my understanding of it went that at divide the sample rate by 2 and you have the lower threshold frequency that all others were "shifted" up to? I am in danger of misstating or oversimplifying, but I figure if anyone can give me a no BS answer it's you.

Also, are you saying due to the nyquist point alone, it is more helpful to elevate sample rate to 96khz even though my hearing isn't really keen enough to detect a difference? Does this "frequency shift" lower freqs. into audible ranges, much like a ccd camera will display infrared and uv/a light visually?

thanks
Andy

[Yalbi]

I take your point and accept what you say, if you are holding your EVP's up to scrutiny. All I mean to say is, if you want to have a go you can do it with cheaper stuff and get good results.

[CaveRat2]

ourobouros2k2, on Jul 17 2010, 12:28 PM, said:

Jim, you are very knowledgeable on these matters, I wonder if I could pose a simple question that has never been properly explained to me.

Put simply, if the frequency of an evp is inaudible at the time of recording, why should it be any more audible on playback? Since the recorder isn't frequency shifting, it should still playback as inaudible...

Todd Bates once explained that this had to do with the nyquist point. I didn't fully understand it, but my understanding of it went that at divide the sample rate by 2 and you have the lower threshold frequency that all others were "shifted" up to? I am in danger of misstating or oversimplifying, but I figure if anyone can give me a no BS answer it's you.

Also, are you saying due to the nyquist point alone, it is more helpful to elevate sample rate to 96khz even though my hearing isn't really keen enough to detect a difference? Does this "frequency shift" lower freqs. into audible ranges, much like a ccd camera will display infrared and uv/a light visually?

thanks
Andy

Not quite sure wat you mean as an "inaudible frequency". If you are refering to out of band, i.e. below 20 Hz or above 18 kHz then it would be inaudible either way unless shifted. You also need to be awaare that equipment may impose limitations as well, cheap recorders often limit between 100 Hz and maybe about 4 kHz. Those would also limit your ability to record anything outside the bands they cover. Even good recorders often cut out above 15 kHz.

Now as far as intentional frequency shifting, there are ways to accomplish that, and in so doing you could actually hear something above the normal range of hearing. This can be done several ways, the prefered is by superimposing the audio on a high frequency carrier then hetrodyning it down. Filtering must be employed to prevent aliasing and artifacts but it can be done. The Nyquist point comes into play there because you are sampling a much higher frequency, that of the carrier. So obviously the device doing the conversion must sample at a rate well beyond the audio range. Recorders are not frequency shifting devices, although by changing sample rates on a fixed recording such a sample shift can be done. You would be required to push the original sample rates higher though since later slowing could affect the conversion process. In other words the Nyquist point would be established at the lowest sample rate to be used since that would be the worst case scenario.

Spectrum analyzers employ such techniques, my Agilent technologies unit samples at 500 mHz. Of course thios is so it can deal with RF as well as audio frequencies. While not mentioned in the literature, there is a Nyquist point that applies in its design as well. But I let the engineers who designed it figure that out!

[afterlife]

I did a simple test with my recorder because I was baffled as to why we can not hear some EVP's but upon playback they were there. I utilized a speaker with direct current running through it, which emits a low frequency that we can not hear. The room i used was totally isolated & quiet. I recorded for one minute without the D/C speaker turned on, then recorded for one minute with the D/C speaker turned on. In both situations I could not hear any thing with the naked ear. Upon reviewing the audio samples there was a notable hum in the audio sample with the D/C speaker turned on. As I said, I myself heard no difference while doing this but upon playing back the recorder it had picked up the hum of the D/C . My conclusion is that EVP's might come in at the lower frequency range. I beleave that would be around 20Hz or below. This would explain why we can not hear the EVP's at the time of occurrence. For the record My recorder meets all of caverats spects & is stereo not mono. I have caught many class A EVP's but Ive only heard one of them with my naked ear.

[CaveRat2]

Yalbi, on Jul 17 2010, 03:04 PM, said:

I take your point and accept what you say, if you are holding your EVP's up to scrutiny. All I mean to say is, if you want to have a go you can do it with cheaper stuff and get good results.

But the problem with cheap recorders is not that you can't get EVPs, in fact you can get them on any recorder. The problem is that the cheap recorders are prone to artifacts, and if these are present how does one determine which is the EVP and which is the artifact? Bottom line is you can't, therefore the validity of the EVP is called into question.

[ourobouros2k2]

CaveRat, on Jul 17 2010, 03:08 PM, said:

So obviously the device doing the conversion must sample at a rate well beyond the audio range. Recorders are not frequency shifting devices, although by changing sample rates on a fixed recording such a sample shift can be done.

So regardless 96khz would be the recommended rec. mode...

thanks
Andy

[Old Guy]

afterlife, on Jul 17 2010, 02:11 PM, said:

I utilized a speaker with direct current running through it, which emits a low frequency that we can not hear.

When you say, "I utilized a speaker with direct current running through it, ...", What are you really saying? Are you biasing a speaker with a DC source? What is the source? And when you say, "which emits a low frequency that we can not hear." Then how do you know it's there?

A speaker is an electromagnet. And inherently an AC device. Pass DC through it at sufficient current, and it will bang against the stops doing physical damage, and then the smoke escapes. Lower currents will drive the voice coil to certain a position and stop. That's called a linear actuator.

[CaveRat2]

afterlife, on Jul 17 2010, 04:11 PM, said:

I did a simple test with my recorder because I was baffled as to why we can not hear some EVP's but upon playback they were there. I utilized a speaker with direct current running through it, which emits a low frequency that we can not hear. The room i used was totally isolated & quiet. I recorded for one minute without the D/C speaker turned on, then recorded for one minute with the D/C speaker turned on. In both situations I could not hear any thing with the naked ear. Upon reviewing the audio samples there was a notable hum in the audio sample with the D/C speaker turned on. As I said, I myself heard no difference while doing this but upon playing back the recorder it had picked up the hum of the D/C . My conclusion is that EVP's might come in at the lower frequency range. I beleave that would be around 20Hz or below. This would explain why we can not hear the EVP's at the time of occurrence. For the record My recorder meets all of caverats spects & is stereo not mono. I have caught many class A EVP's but Ive only heard one of them with my naked ear.

There is no hum in a DC source. By nature all DC sources consist of a current flowing from neg to positive. There is no frequency to a DC source, by nature it is steady state. That said I would next question your source of DC. If you used a power converter, rectifier, then it is possible some ripple was leaking through from the transformer / rectifier. This is a common source of hum in a DC supply. Another factor would be the coupling into the recorder. If your recorder uses a direct coupled input then it is possible the DC would throw off the bias on the input stage. This could result in erratic operation.

[CaveRat2]

ourobouros2k2, on Jul 18 2010, 05:14 PM, said:

CaveRat, on Jul 17 2010, 03:08 PM, said:

So obviously the device doing the conversion must sample at a rate well beyond the audio range. Recorders are not frequency shifting devices, although by changing sample rates on a fixed recording such a sample shift can be done.

So regardless 96khz would be the recommended rec. mode...

thanks
Andy

I recommend 96 KBPS as a result of testing done in my lab. Using a Fostex FR2, (which can sample up to 192 KBPS) I tested frequency reponse and distortion at several rates. Using a scope and spectrum analyzer I could see no appreciable difference between 192 and 96 KBPS. Thus I found 96 to be satisfactory. When I dropped down to 44 KBPS you could not really hear much difference. But response began to show some deterioration on the spectrum analyzer. Some extended frequencies were attenuated, and harmonics, while not serious began to appear. Thus I determined that while usable, 48 KBPS was beginning to show slight degradation. When I dropped down to 32 KBPS, deterioration became more noticable. You could ven begin to hear it audibly. Thus anything below 32 KBPS was deemed compromised.

So to summarize I would say if you had 96 KBPS available use it. If your recorder only went to 48 KBPS then you might get away with it in a pinch, but it would be suspect if you expected to do any serious analysis. Below 44 KBPS, it's not recommended. Below 32 KBPS, time to get another recorder!