Claravox Centennial- known issues, bugs and quirks

"The “ghost frequency” in this video sticks out like a sore thumb! (video cued)."  - coalport

The video for me just starts at the beginning.  Where exactly in it are you hearing the ghost tones?  [EDIT] Never mind, I see you have it cued up on Levnet at 9:17.

I must admit that I'm one of the ones that can't hear them at all in the previous Claravox video.  I even downloaded it, extracted the audio, and examined the spectra.  But I have some tinnitus which could be masking it.

In general, I wonder if the YouTube transcoding process has anything to do with this?  They can really mangle stuff when you upload it, and altering the sample rate can be a tricky thing to not introduce audible artifacting, not to mention the compression process as well.

[EDIT] I downloaded the M4A file of the Theremini video and definitely hear aliasing or something at 9:24.  Here's an MP3 of just that: [MP3]

(Steven from the Claravox facebook group, first time posting here) In the theremini example it sounds like she is using effects with it. Delay in particular seems rather noticeable (timestamp 9:30, when she mutes there is a bit of echo going on). I know some people also use a harmonizer pedal. Could that be the source of the ghost pitch on the theremini, given the ghost pitch does seem to follow the main pitch? EDIT: Could also be the preset she is using? Later on in the video she swaps to other presets that dont seem to have the ghost pitch (at least to my ears).

Speaking of harmonizing, the Claravox does have 2 oscillators and the second one can be detuned from the first and their levels can be independently adjusted. Possible source of it on the Claravox? IDK, I couldnt hear the ghost pitch in the Claravox sample even through my studio reference headphones driven by a dac+amp combo. At -42db as per the other poster it could be that the signal to noise ratio is just on the edge of perceptible to some and not to others. EDIT: Also a potential preset thing as with the theremini?

It also could be aliasing from the transcoding process. There are so many variables involved I'd hesitate to say its 100% a flaw with the Claravox. You'd want to listen to and analyze as many samples of it as possible to make that determination.

(Steven from the Claravox facebook group, first time posting here) In the theremini example it sounds like she is using effects with it. Delay in particular seems rather noticeable (timestamp 9:30, when she mutes there is a bit of echo going on). I know some people also use a harmonizer pedal. Could that be the source of the ghost pitch on the theremini, given the ghost pitch does seem to follow the main pitch?Speaking of harmonizing, the Claravox does have 2 oscillators and the second one can be detuned from the first and their levels can be independently adjusted. Possible source of it on the Claravox? IDK, I couldnt hear the ghost pitch in the Claravox sample even through my studio reference headphones driven by a dac+amp combo. At -42db as per the other poster it could be that the signal to noise ratio is just on the edge of perceptible to some and not to others.It also could be aliasing from the transcoding process. There are so many variables involved I'd hesitate to say its 100% a flaw with the Claravox. You'd want to listen to and analyze as many samples of it as possible to make that determination.

I am able to consistently hear the "ghost pitch[es]" and am certain that they are not harmonics or something related to the transcoding process. The reason for this is that the “ghost pitches” are most prevalent when he’s sustaining a constant pitch with his right hand; a very clear ascending glissando is heard as his left hand moves away from the volume antenna and an equally clear descending glissando as his left hand moves closer. It’s as if the volume antenna is also serving as a secondary pitch antenna. 

I'm not sure why many others cannot hear it (it's very evident to me) but I wonder if it may have something to do with it being at the edge of hearing perception in many individuals. I just listed to it again, several times, using my junky internal computer speakers; surprisingly, I can still hear it. At first, I thought the inability for others to hear it may be related to high frequency hearing loss - but that's not it. There's something else about the phantom sounds - the pitch seems to be considerably lower (at least two octaves) below that controlled by the right hand. Thus, the ability to hear (or not hear) it appears to be related to the amplitude of the sound, not necessarily the frequency.

To be fair, I haven't yet heard the phantom sounds with my own Claravox. 

I am able to consistently hear the "ghost pitch[es]" and am certain that they are not harmonics or something related to the transcoding process.

Well, all I can do is speak on what I myself am hearing.

In the Claravox example I don't hear the ghost pitch at the timestamps they are reported to be - or at any point in the video really.

In the Theremini example I absolutely do hear the ghost pitch at Dewsters reported 9:24 timestamp. It absolutely is crystal clear that there is another much lower pitch there, but it is following the main pitch instead of being controlled by the volume antenna. But then later on in the video when she is changing to other presets on the Theremini I once again can't tell if there is any ghost pitching going on.

To be fair, I haven't yet heard the phantom sounds with my own Claravox.

If you haven't heard it on your own Claravox in person then I'd definitely suspect something is up with how the videos were processed. YouTube is not known for high quality audio. It could absolutely be the Claravox itself, I'm not denying that possibility, but I'd wait until people are reporting hearing it in person before saying so for certain.

If you haven't heard it on your own Claravox in person then I'd definitely suspect something is up with how the videos were processed. YouTube is not known for high quality audio. 

I'm eager to hear what others have to say too. However, the phantom glissandi very clearly follow the motion/direction of his left hand - there is a 1:1 correspondence between the two, which would be highly improbable if the ghost pitch was the result of the processing of audio by YouTube. If that were the case, the ghost pitch should follow the motion/direction of the right hand as the only function of the left hand is to control volume. I wonder how many others can hear it?

As I haven't heard it on mine, I wonder about another consideration - we know that Moog made some (apparently slight) modifications to the electronics of the instrument from the earliest versions - Dewster noted these when we opened up mine to fix it and we could compare with photos that Moog provided on their site (I believe they were dated August of this year). Gregoire's Claravox is, obviously, one of the very earliest instruments - mine is #00218. There are reports that other modifications have since been made (although serial numbers were not reported). It'll be interesting to see if others have the "ghost pitches" on their instruments and, if so, what are their serial numbers.

BTW, the "ghost pitches," which are also clearly evident, on the Theremini video are very different than those on the Claravox video.

The “ghost pitch” phenomenon is something I have  heard many times from various different theremins over the years. It is highly unlikely that it is being generated by an effect or by reformatting protocols, and just because you hear it from one CLARAVOX and not from another, does not mean that the theremin itself is not responsible for the anomaly. 

The CLARAVOX theremins, from what I have seen and read about them on the internet, is beset with random, quirky little problems. This is a result of poor quality control at the source. 

As for the fact that some people hear the phantom and others do not, hearing acuity (as well as playback technology) varies enormously between individuals. Sorry folks, I don’t mean to piss people off, but just cuz ya don’t hear it, don’t mean it ain’t there! 👂🏻 

I wasnt trying to deny the existence of a problem, nor am I pissed or angry. Just a bit of a skeptic.

If its an issue such that person A may be able to hear it while person B can't, you really should come with better evidence than a youtube link with a timestamp.

Hi Mennenth,

I’m afraid YT demos are the only “evidence” I can find for problems with the CLARAVOX - and there are plenty of people who have been disappointed enough with the performance of their new instruments that they have resorted to public forums to complain.

My remark about some people not hearing “ghost frequencies” (aka digital aliasing) was not directed at you. It is something we have observed for years in the theremin community.  

In The Interests Of Science...

I think it's more or less fair use to extract short snippets of audio from the Claravox video for purely educational purposes: [ZIP]

I noticed (heard) fairly severe audio artifacting in the 44.1kHz sample rate audio and none in the 48kHz sample rate audio, so I used the 48kHz to generate a few high bit rate MP3's.  I'm using MP3 here because most audio programs will open it, and haven't experienced any significant spectral analysis issues with it in the past (surprisingly).  The "best" combined video & audio playback selects the 44.1kHz sample rate, so maybe that's some of what folks are hearing?

Imported the 48kHz, 16 bits, 2 channels, Opus webm format in Audacity.

Nice round number snippet selections (for easy indexing during analysis & discussion) exported to 48kHz MP3 @ 320kbps:

8:30:00 to 9:14:00
10:40:00 to 10:45:00
10:53:00 to 11:04:00

For spectral analysis I've found SonicVisualiser to work pretty well: https://sonicvisualiser.org/

Have at it!

[EDIT] Looking at the video, we don't even know what mode the Claravox is in, Traditional or Modern.  I wouldn't expect ghost notes in Modern mode as they are presumably digitally generated, though I suppose there could be Nyquist aliasing if not generated correctly.

[EDIT2] Performed the same snippet extraction for the 44.1kHz audio: [ZIP].  Listen in particular to"10:53:00_to_11:04:00_44.1kHz.mp3" - during the gliss it sounds like someone is rustling a beaded curtain!  Not sure how that got in there, but I can't say that I really trust the YT audio conversion process.

Cvox Traveling Hump

Looking at the sound snippet "10:53:00_to_11:04:00.mp3" I'm seeing a tone below the fundamental frequency, and about 30dB down, clearly frequency modulated with the volume:

I believe this is what Jeff (ContraDude) was describing?  Here's a video of it: [MP4].

So I'm guessing the Claravox is in Traditional mode, and that there is some leakage or something from the volume side intermodulating with the pitch.