Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Getting strange spectrograms from my vinyl rips (Read 6364 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Getting strange spectrograms from my vinyl rips

In the last couple of days i ripped a few vinyls out of my collection. I'm new to vinyls and just wanted to fool a little bit around.
Everything went good and the recordings are all sounding fine. Sometimes i check few of my audio files with Spek a "Acoustic Spectrum Analyser".
I don't weight a spectrogram of an audio file too much, i just want to see where the cutoff frequencies are.
I then realized that the spectogram of my vinyl recordings all have strange constant lines in them on certain frequencies. Some of them are beneath 20 kHz and should therefore be audible if they are really existent right ? However the phenomenon is not audible and I cant see anything static at these frequencies when i view the live "spectrum" in foobar while listening to a silence part of the audio file.

Anyway here is the phenomenon I'm talking about:





I get the same phenomenon when I'm viewing the spectrogram in Audacity:





I use a Dual 1237 A turntable which is a part of a Saba Ultra HiFi Center 9903.
The direct line signal goes from a 5 pin DIN cable adapted onto two 6,35 mm stereo phone connectors into my Lexicon I-O | 22.
I'm recording with Cubase LE 8 at 24bit 96kHz. Yes I'm aware that 24bit 96kHz is an overkill for a vinyl ripped from a standart setup.
My plan was to bounce them out of Cubase in 16bit 44.1kHz FLAC. I tried to record the vinyl on different settings, sampelrates and bits per sample.
No matter what i do i always get these strange lines in the spectogram. I even switched the 5 pin DIN cable, so at this point I'm pretty sure the reason is the Dual 1237 A turntable.

As I said the phenomenon is not audible but I'm still curious what could cause this ?
Can anyone explain me what these lines are and why they are not audible ?

Thanks for reading. Every kind of help is appreciated!




Re: Getting strange spectrograms from my vinyl rips

Reply #1
In think that the answer to the second question (i.e. why are they inaudible) is quite easy: They are too quiet, and most of them are out of the human hearing range.
If we focus on the one we see at 12Khz (the only candidate to be audible) , we see that it is between -80dbFS and -90dbFS. Frequency peaks, from what I can see, might be around -30dbFS, although most content might be between -40dBFs and -60dbFS.

So, the frequencies are between 30 to 40dB below the content. Just as a reference, the "mute" button on many devices used to be a -20dB attenuator.

So, I believe you would need to have a volume similar to that on a Music club in order to hear that line when there's silence.


As for the origin of the lines themselves, it is a bit strange.. 12Khz is not an usual frequency. 50/60Hz or 17Khz are.
I understand that the I-O 22 is the soundcard, and that you record with Cubase. Are you using ASIO for recording? Basically, to discard completely that the OS could be resampling.

Another thing to try is to discard that it is capturing emissions from other equipment nearby. Try to move the devices further from any other electronic component to see if those lines change or not.

At last, we cannot discard completely that they don't come from the disc (although I guess that's as easy to try as to record when the turntable is on, but stopped, or even rotating without a disc).

Re: Getting strange spectrograms from my vinyl rips

Reply #2
They may/may not be audible but they are definitely not normal.  Interesting that they appear to be exactly 2000 Hz apart.  Put your software in record mode with your turntable playing nothing but turning and record a minute's worth of silence.  Do you see the lines in this sample?  Now turn the turntable off/unplug it and record again.  Still lines?

Are any of your components including the computer plugged into a power strip?  Temporarily get rid of it and record a sample to check.  Also check the routing of your cables; possibly a power cable is running along side a signal cable and generating noise/hum in the signal.

Re: Getting strange spectrograms from my vinyl rips

Reply #3
It might be coming from the Lexicon.    You're going to get hum & hiss from the preamp but those higher more-specific frequencies look like  some kind of a clock.

Re: Getting strange spectrograms from my vinyl rips

Reply #4
Yesterday I spent a lot of time testing all your advices. In Cubase I'm recording with my ASIO driver. Few days ago I recorded a tape with a Sony WM-D6C Walkman in Cubase through my Lexicon.
I cant see any lines in the spectogram, so I guess we could rule out a resampling issue. I moved the turntable quite far away from any other electronic component but I still get the lines.
When I record just silence without the  the turntable spinning, I still get the lines. If I turn off the turntable, the lines will disappear but I still get strange noise around 200 Hz.
I checked the routing of all my cables and rearranged them. The turntable is not plugged into a power strip but my computer and my Lexicon are.

After all the testing I'm quite sure that I have a chassis ground/earthing problem in my setup. The preamplifier in my Lexicon seems to struggle with it. When I turn the  MIC/Line gain knob from my Lexicon all the way up to the max i get really ugly noise.
Since the turntable is pretty quiet and therefore the knobs wide open, I added a Ultra DI DI20 Box between the turntable and my Lexicon. The results are much cleaner since i can record without having the MIC/Line gain knob wide open.

I'm still not satisfied with the result. Do you guys have any tips on how to fight chassis ground/earthing problems in a setup ?
Thanks for your fast replies and your advices!

Re: Getting strange spectrograms from my vinyl rips

Reply #5
Interference from ground loops usually leads to noise frequencies which are related to the mains frequency. The noise around 200 Hz could be mains-related. The horizontal lines of your original spectrograms seem to be spaced apart at 1kHz, with every other line being stronger. This indicates a noise source with a base frequency of 1 kHz, and with a high content of harmonics. This could be impulse-like noise with one narrow spike each millisecond.

There is a 1 kHz frequency ingrained in the USB protocol. Hence the noise might originate with the USB communication. Quite how it leaks into the analog frontend isn't clear, but ground wiring is one of the prime candidates.

Combatting ground loop interference usually attempts to separate noise currents from signal currents, so that they don't share the same path anywhere. To this end, it is usually helpful to think of the current path, rather than the voltages associated with the signal. The current path is always in a loop (hence the term "circuit"), and ground loops impair a signal when there is overlap between the signal loop and the ground loop. The noise couples into the signal loop through the impedance of the shared path, which is often formed by the ground wire in a signal cable.

A systematic strategy therefore tries to identify the loops involved, and to either interrupt the offending ground loop to cut off the noise current, or to reduce the impedance of the common path, ideally down to zero. Of course, if there isn't a common path anymore, the coupling impedance is also zero.

Interrupting a ground loop isn't always feasible without serious drawbacks. Safety rules, for example, may forbid it.

In your case, the offending ground loop is very likely including the ground of the signal cable between turntable and the Lexicon, the ground of the USB cable to the computer, and probably the mains wiring of computer and turntable. Other connections may be involved depending on the overall setup. Since the USB cable is involved, this may be how the 1 kHz spikes intrude.

The DI box is a good idea in general, particularly since it is battery powered, but its effectiveness depends on how its ground wiring is done internally. It may also help to have solid extra ground wiring between the Lexicon and the turntable. The Lexicon unfortunately has no ground binding post, but the turntable may have one. You may have luck finding a suitable grounding point somewhere on the Lexicon.

Another route is to exploit the fact that the Lexicon's inputs are balanced. This means that they can work without a ground connection. You may have to build your own special cable between Lexicon and turntable to make this work. You probably used some form of adapter that you bought, and those usually connect the grounds. You wrote the adapter uses two stereo phone plugs, so they will have tip, ring and sleeve contact. The sleeve and ring contacts probably are joined together. You would need to have the sleeves disconnected. You can try to change this if you know how to use a multimeter and a soldering iron.

Re: Getting strange spectrograms from my vinyl rips

Reply #6
I had similar (but less frequent) lines when I digitized tapes from SONY TC-WR545 tape deck this summer. They were even less prominent when Dolby B was switched on. They were not audible and unfortunately I did not figure out the cause. I also could not rule out ground loop problems. I got rid of some of them by lowpass filter.

Re: Getting strange spectrograms from my vinyl rips

Reply #7
Quote
However the phenomenon is not audible...

...I'm still not satisfied with the result.
Think about that...  You can't hear it, but you're not satisfied because you have the ability to measure the noise.   You're never going to get perfection.    It's analog* so there is always some noise.  

Imagine that you can get the noise below the level where you can measure it...    Are you going to become dissatisfied again if in the future you gain the ability to measure it?

There is audible noise from the record, and perhaps some hum & hiss from the preamp.   If you want to worry about something, worry about the audible limitations of analog vinyl...   ;)  




* With digital, there is quantization noise and most digital audio originates as analog so digital isn't mathematically or measurably perfect either, although it can easily be better than human hearing and therefore it can be audibly perfect.

Re: Getting strange spectrograms from my vinyl rips

Reply #8
Think about that...  You can't hear it, but you're not satisfied because you have the ability to measure the noise.  You're never going to get perfection.    It's analog* so there is always some noise. 
If I were in his place, I would want to digitise the records with the best quality I can achieve, no matter whether I can hear the noise or not. I would only do it once, and if I would later find out that the noise becomes audible in some situations, I would be annoyed.

The kind of noise he was showing in the spectrograms was clearly not unavoidable analog noise; it pointed to a flaw in the setup somewhere. I think it is a testament to his diligence that he wants to get to the bottom of it.

Quote
Imagine that you can get the noise below the level where you can measure it...    Are you going to become dissatisfied again if in the future you gain the ability to measure it?
I would perhaps be dissatisfied in this situation, if I found that I had goofed somehow, and could have done better with more effort.

Quote
There is audible noise from the record, and perhaps some hum & hiss from the preamp.  If you want to worry about something, worry about the audible limitations of analog vinyl...  ;) 
Sure, but I consider it useful and sensible to separate the avoidable noise from the unavoidable noise, and try to get rid of the latter.

Re: Getting strange spectrograms from my vinyl rips

Reply #9
One more thought - Since it's analog vinyl, I assume you are going to apply some noise reduction?   The noise reduction just-might remove the inaudible noise.

Quote
If I were in his place, I would want to digitise the records with the best quality I can achieve, no matter whether I can hear the noise or not.
Human nature, perhaps.  But not really logical.  It's the same thinking that makes people want "high resolution" audio when there is no audible difference and it's the same thing that drives people to digitize vinyl at 24/96 or 24/192.    The whole point (hopefully) is listening to and enjoying the music.

Quote
Sure, but I consider it useful and sensible to separate the avoidable noise from the unavoidable noise, and try to get rid of the latter.
That reminds me of a joke that my statistics professor used to tell over-and-over...  I won't repeat the dumb joke, but the point was about studying what's important rather than what's easy to measure.

And so far it hasn't proven to be "avoidable".    I guess it's a matter of how much money and time (and how much time-delay) its worth to him to make an "improvement" that he (or anybody else) can't hear.

    

Re: Getting strange spectrograms from my vinyl rips

Reply #10

Sure, but I consider it useful and sensible to separate the avoidable noise from the unavoidable noise, and try to get rid of the latter.

That only makes sense if the avoidable noise is not already far less than the unavoidable noise.  In this case the avoidable noise (due to the use of 44/16 ADCs) is typically far less than the unavoidab le noise in the vinyl media.

Re: Getting strange spectrograms from my vinyl rips

Reply #11
That only makes sense if the avoidable noise is not already far less than the unavoidable noise.  In this case the avoidable noise (due to the use of 44/16 ADCs) is typically far less than the unavoidab le noise in the vinyl media.
You may have misunderstood my point. The noise I was referring to wasn't the quantisation noise of a 44/16 ADC, but the low-level tones which were visible in the spectrogram as horizontal lines. I certainly don't want to imply that anything beyond 44/16 is beneficial for ripping vinyl. I do recommend, however, to try to get rid of interference even when it looks like it is inaudible, because I wouldn't want to be in a position later on where I find that the interference becomes audible under certain circumstances. Better safe than sorry.

Re: Getting strange spectrograms from my vinyl rips

Reply #12
Quote
Better safe than sorry.
OK, you might want to get a better preamp and maybe try a different interface.  Can you bypass the preamp in the Saba Ultra HiFi Center?      And if the noise happens to be "radiating" from the  Saba Ultra HiFi Center, you might need a different turntable or maybe you can pull the turntable out of that system to isolate it and run it separately.   

Or...  You didn't say why you are digitizing your records...  Are the recordings not available digitally?

Above, I mentioned that it's human nature to want the best...    But, it's also human nature to be economical.   So, it's a question of what it's worth to you.   And of course there is some economic risk, since you don't know which components to replace.