CD versus LP - a comparison of the frequency spectra

This remark seems illogical to me. As far as I know, no brick-wall filters are used for recording analoguely mastered vinyl records, so it would be surprising if there were no content above 20 kHz.

You don't think that the cutter that makes the master disc has an upper frequency limit??? It's a mechanical filter.

Anyway, from what I understand, there is often a high "cut" at 16kHz applied for many recordings... above that, what is obtained from playback is not the original signal. Most likely noise.
 
Recordings can go very high to catch harmonics and distortions of original sound, but level is typically very low compared to pure tones (few %).
Cutting them on vinyl and playing back is difficult and finally loudspeakers will make mess of them... and anyway only bats could hear that!

https://www.channld.com/vinylanalysis1.html

ps. Linear frequency scale used earlier in this thread and in my link overemphasizes high frequencies. Logarithmic scale is the right one to use.

Cats and dogs can hear it as well, and they often are subjected to the sounds coming from audio equipment.

Besides, according to a Japanese article from the 1990's, Japanese gamelan players can subliminaly hear the difference between audio bandlimited to 26 kHz and not bandlimited to 26 kHz. Their electroencephalograms changed. It doesn't become clear to me from the article whether the test was double blind or only single blind (a.k.a. unreliable).

In any case, people often pay extra for digital recordings with content above 20 kHz, the so-called high-resolution recordings.
 
You don't think that the cutter that makes the master disc has an upper frequency limit??? It's a mechanical filter.

Anyway, from what I understand, there is often a high "cut" at 16kHz applied for many recordings... above that, what is obtained from playback is not the original signal. Most likely noise.

Sure, but how many dB/octave? Probably nowhere near the -700 dB/octave you need to prevent aliasing in a CD recording.
 
  • Like
Reactions: TNT
I found that at the gatherings of a Dutch audio forum, the people playing vinyl records usually have them playing in minutes, while those using streaming need hours to solve all computer networking issues. 😉


Sorry about your Dutch friends and their problems!
I believe it was all the Germans attending the Dutch audio forum that were having all those hours of computer problems... 🙂
This is pure speculation on my part of course...
 
  • Like
Reactions: Juhazi
Sure, but how many dB/octave? Probably nowhere near the -700 dB/octave you need to prevent aliasing in a CD recording.

What does that have to do with content in the spectrum above 20kHz in a vinyl playback chain??? Absolutely nothing! You are just trying to change the subject. Pure what-about-ism. Pathetic.

This is like responding "but my dad could beat up your dad"...

You (and I mean YOU) cannot hear anything above 20kHz, or 15kHz if you are older than about 40. So who cares about the anti-aliasing filter?

The vinyl source is not "superior" because it has content above 20kHz, but rather that is a fault of the medium - it's just noise and distortion.
 
Isn't it obvious? When the cutter rolls off at -12 dB/octave and it is preceded by a -12 dB/octave filter, both at 18 kHz, a 36 kHz signal will only be attenuated by 24 dB and can therefore still be quite visible in the spectrum. A 26 kHz signal will only be attenuated by about 13 dB then.

With a digital anti-aliasing filter that goes from 0 dB to -100 dB or less between 20 kHz and 22.05 kHz, you will have to measure very accurately to see anything around 22.05 kHz. If the sample rate is 44.1 kHz, you may see something above 22.05 kHz when you play it back over a DAC with limited image suppression, but that is completely distortion then: intermodulation distortion between the signal and the sample rate.

I'm quite sure I don't hear any difference, but what has that got to do with anything?

At the very least, ultrasonic content and slow roll-off are marketing advantages, whether they are subliminally audible to Japanese gamelan players or not. Any half-decent audio DAC nowadays supports high sample rate playback with a smooth roll-off filter with a short impulse response, so there must be a market for that.
 
  • Like
Reactions: hbtaudio and Markw4
Cutting heads don't have the same high frequency resonance as a playback stylus, which comes from the effective (meaning referenced to contact point) mass of stylus, cantilever and generator times vinyl compliance, and is near the upper end of (young peoples') hearing range. Cutting heads are resonant nearer the center of the audio range and are equalized more like a dynamic microphone.

Separately, as others have pointed out, there is no reason to assume that a CD and vinyl record should be mastered the same. Bitter experience has shown too many exceptions - just explore the Van Morrison catalog and compare the earliest (1980s) versions with the 1997 and 2008 remasterings. Only the earliest sound like the vinyl records; later versions are just awful.

An alternative "CD" source for comparison would be a homemade rip from the vinyl, so comparison would be with the same mastering. I wonder if any of us could hear the difference between a real-time vinyl playback and an A/D/A Red Book rip and reconstruction, if properly blinded. I couldn't with equipment of >15 years ago and it must be better now.

All good fortune,
Chris
 
You have to keep in mind that music is usually mastered differently for vinyl and CD. This is to compensate for shortcomings in the vinyl format which dictates things Iike summing bass to mono and compressing the top end more to prevent excessive slewing of the cutting head amp (to avoid overheating, saturation distortion). This is due to the HF needing to be cut hotter on a record to compensate for treble heavy RIAA EQ. It lowers surface noise and improves top end dynamics.

Often times, the mastering engineer will simply pay closer attention to cutting the record because he has to watch over all the other potential issues. The compression, EQ and limiting have to be optimized in order to make the analog vinyl pressing sound and play as good as it can based on the multiple Iimitations of the format itself.

With digital / CD audio there's really nothing to worry about aside from DC offset, infrasonic noise and cutting off treble sharply above 20 kHz to avoid DAC aliasing during playback. What you essentially put into it is exactly what you get out of it.
 
@CharlieLaub Half speed mastering gets around some of the top end limitations of vinyl caused by the cutting head. However, as some have already pointed out, the typical signal chain in a recording studio will cut off most of what's there above 20 kHz or less.

Noise goes up exponentially as frequency response is pushed higher up. Very few recordings have intentional music information above 20 kHz, simply because microphones don't have much response above that.

Direct recorded instruments like some keyboards, electric basses and other non-miced sources don't have much of any significant output above 20 kHz (often much less). These istruments don't generate much of any overtones or harmonics as it is. Only very high end small diaphragm condensor mics can pick up sound significantly past 20 kHz.

If you listen to classical music, most HF harmonics are generated by percussion and brass instruments ie. triangle, piccolo, flute, cymbals, trumpets, coronets, etc. The violin and piano would be the possible exceptions as would be 12 string guitar. Fender Rhodes could have a few harmonics above the highest keys.

By the time you pass the recorded audio through most of the signal chain, it already has a significant portion of the top end cut off well below 20 kHz. There are a (very) few classical and acoustic / solo vocal recordings which have content past 20 kHz, but these are extremely rare.

Sure, if you look at breaking glass, jiggling keys and other novelty sounds (unimportant to music), those are rarely recorded for music purposes. 99.9% of music doesn't have useful energy above our hearing range which is limited lower than 20 khz. Most people older than 40 yrs can't hear past 15 kHz anyways and that's being generous.
 
  • Like
Reactions: hbtaudio and Juhazi
Ehm... At least in theory, those high frequencies might have been harmonic distortion or hiss. There's no way of telling from looking at the diagram. Yes, there's more energy in the higher part of the spectrum on an LP, but it could be all garbage. I'm not saying that it is, Just that this diagram doesn't prove anything.
 
About HiRes versions, I bought some downloads from HDtracks and analyzed them with Audacity 15 years ago. My system was capable of 96kHz. Almost every HD version of old records had no content above 18kHz, which proves that they were copies of originals. Dynamics was good an that is fine. The few new HD-chain recordings had something up to 30kHz at very low level at some parts, so it was not noise. I don't have those files and captures on this laptop so I can't show them, they are in my home museum... Naturally I stopped paying for nothing and started to use streaming about five years ago.

As sidenote, I couldn't detect difference between mp3 256 or better and 96/24 even with headphones (hires original converted with Audacity)

Below my vinyl Thorens/AKG clips, the preamp has no HP filter! The silence clip shows noise from the record+deck.
Biggest differences between vinyl decks is rumble and resonance, and if the preamp has filter for it. And this is very easy to hear!

td145mk2 end groove spectr terratec.jpgtd145mk2 last track spectr.jpg
 
Last edited:
I am recording my LPs to FLACs with Audacity and yes, it's very time consuming🙄
This only was intended as a quick and funny comparison between the recording process. Actually I did not expect such a good result.
And yes, there a better and more sophisticated ways to compare these media formats.

And when it's done my turntable will become a show piece and dust collector. Because I don't have the heart to give away the LPs and the turntable.

I did an "ear-comparison" with a HD650 and couldn't hear major differences. When listening a little louder of course you can hear a very faint rumble of the record player and sometimes a crackle of the vinyl - especially during quiet passages.

Yes, the LP-rip has a wider upper frequency range because i did a 48KHz rip. CD is limited to 20KHz because of 16Bit/44,1KHz.

Yes, it's this receiver: https://www.hifiengine.com/manual_library/pioneer/sx-950.shtml I have replaced all E-caps and some transistors in PSU and phono stage. I also replaced the electrolytic capacitors, the motor, the cartridge-tone arm cables and belts of the turntable.
 
Ehm... At least in theory, those high frequencies might have been harmonic distortion or hiss. There's no way of telling from looking at the diagram. Yes, there's more energy in the higher part of the spectrum on an LP, but it could be all garbage. I'm not saying that it is, Just that this diagram doesn't prove anything.
Yes its mainly preamp noise typically. If you want to do spectral analysis to gauge performance its much better to use test tones, not music. Single tone, two-tone, chirps, multitone, white/pink noise etc, as you can then separate out the various effects and measure them. Not sure I've ever seen a really good test disc with all the test tones you'd like for vinyl measurements, they are usually oriented towards setting up geometry and tracking force...
 
You'd be shocked at how much THD vinyl playback has. Even with a high end MC cart that has a decent stylus profile ie. microline, shibata, etc.

Both my AT OC9/II or VDH SP3 carts tracking a decent piece of vinyl in the sweet spot of the arm's azimuth is capable of under 1% THD. In most cases its much higher than that, plus the surface noise and rumble.

CD playback is capable of THD under .01% with a dynamic range of over 90 dB and unmeasureable W/F. That was available in the early 1980s. Now we have 192k/24bit conversion and in some circles even better. In the real world that's not perceiveably better than 44.1k/16bit. Even 320mbps is almost impossible to discern from the uncompressed PCM file.
 
Last edited:
I couldn't detect difference between mp3 256 or better and 96/24 even with headphones
In the real world that's not perceiveably better than 44.1k/16bit. Even 320mbps is almost impossible to discern from the uncompressed PCM file.
If you guys are interested in "hearing" differences, may I suggest listening with the centre material (common-mode information) removed ? The result would be some kind of audible flanging in case of MP3 (or any MPEG) and even Dolby AC3 (448kbps) !!

One format that's good in this respect would be DTS (768kbps or above) that uses correlation based compression (vs. subband coding for MPEG & AC3), which (somehow) gives side information that "sounds" very similar to uncompressed PCM. Nevertheless, with the centre material included, the flanging becomes difficult to hear due to the masking effect that the "mid" information causes.
 
  • Like
Reactions: hbtaudio
Important to mention that vinyl quality/recording used to vary a lot according to the label and track position - inner being horrible and outter a bit better.

Back in late 70's and 80's, I use to have my favorite artists but some of them tied to a certain horrible label.
Sometimes, I prefered to listen to a certain label, that had good audio quality than listen to my favorite songs due to bad recording quality.

Distortion on inner tracks of some lables were extremely high. And here not just 0.x% of 2nd or 3rd harmonics - all kinds or high distortion.
We could minimize distortion by adding more mass to the arm, but this procedure wore out both the vinyl and the needle.

Human society is strange.

In those days, 70's and 80's, the media (vinyl) was horrible (distorted, limited in headroom etc) but the studio recording was awsome at the source on the original 24-track 2" tape recording.
Today the media is as perfect and cheap as we want but the recording is horrible (compressed, lo-fi, even clipped by design etc).
I'm talking about pop music - I know that for classical and jazz, things are better recorded.
 
There are many additional factors such as, where and when we provide higher bandwidth, this appears to influence our perception, and appreciation of what we actually can hear. https://www.earthworksaudio.com/wp-content/uploads/2012/07/The-world-beyond-20kHz.pdf

Also that many recordings (more so about capturing available dynamic range) after 1965 used companding which has not necessarily restored the same two part capture and delivery, by afterwards expanding at the same ratio.
 
Last edited:
  • Like
Reactions: hbtaudio