Bandwith limited when ripping vinyl

Special Interest Groups Recording Equipment
1

Hello,

I come here because I have a big problem when ripping vinyles and I don’t know what to do. Visibly I miss some bandwith when I try to record more than to 48 kHz.

My equiment

Turntable: Audio Technica AT-LP120-USBC
Cartridge: Audio Technica AT-95E (MM)
Phonoamp: I use the one built-in in the turntable with the Line output directly on my audio interface
Audio Interface: Steinberg UR22 (24 bit 192 kHz USB)
Audio Capturing: Steinberg Cubase AI Elements 7 (ASIO-Driver)

Some pictures

  • To connect RCA output to Jack input on the audio interface (in the MIC/LINE inputs):

  • Audio interface:

What I obtain with these settings (look from the audio interface picture):

Recorded at 96 kHz, only 44.1 or 48 kHz are present in the bandwith I think, because when I convert the file to 48 kHz, the full bandwith is here.

So here’s my question: what do I need to do? Try to adjust the input gains on the audio interface or buy a new pre-amp like a Cambridge Audio 651P?
Thank you :slight_smile:

2

Where are you looking to see that?
Does it matter what it “looks” like? Surely the “sound” is the important thing :confused:
If you are having problems with 96 kHz recording, then why not use 44.1 kHz or 48 kHz? I doubt that you are able to hear sound above 20 kHz.

What is this? It means nothing without explanation.

3

Yeah you’re right I think the 24 bits it’s the more important to have more details.

It’s the graph in GoldWave when you play a file. I see some graphs using the full width in 96 kHz, sometimes not, from some rips on Internet. When I encode from 96 to 48 kHz, the full width of the graph is used.

Here is a rip from last Daft Punk vinyl in 24/96 :

You can see the graph have the full width, not as me with 96 kHz (seems like to have 48 kHz of the 96 in fact). But maybe it’s not really a problem.

4

Audacity has a “Plot Spectrum” feature: http://manual.audacityteam.org/o/man/plot_spectrum.html
It may not be as pretty as the Goldwave image, but it is possibly more informative.

5

Don’t worry I use both softwares, Audacity for the normalisation I find it better for example. I don’t really know what should I analyze on these graphs. :blush:

6

Vinyl hasn’t such a large frequency spectrum–it is rather a myth.
I’ve learned this while analysing quadrophonic recordings. Most LP’s produce frequencies up to 15 kHz, max 24 kHz.
CD-4 (quadro sound) needs a spectrum up to 45 kHz. This is achieved by letting cut the master at half speed.
For all other recordings, 48000 Hz sample rate should be fine.

7

Audible sound is within the range 20 Hz to 20,000 Hz. This assumes “perfect” hearing for a person under 16. The ability to hear very high frequencies declines with age. For adults the upper frequency limit is usually around 16,000 Hz or less, and usually declines further as the person gets older.

Professional recording equipment is designed to handle frequencies in the range 20 Hz to 20,000 Hz as accurately as possible. This usually means that the extreme ends of their frequency range start from under 20 Hz and extend beyond 20,000 Hz, though they are usually limited so that the range does not extend too far beyond these limits.

The reason that the frequency range of professional gear is deliberately limited is because excessively high or low frequencies will damage fidelity within the audible range. Excessively low frequencies can cause unwanted mechanical vibration (for example in the tone arm of a record player) and produce unwanted heating effects (for example in speaker voice coils). Excessively high frequencies can cause intermodulation distortion that spills over into the audible range.

Ideally, and depending on the musical content, the spectrum of a high quality audio recording should show that the signal is present across the audible frequency range, without much below 20 Hz or above 20,000 Hz.

This is the spectrum of part of a rather good recording of Brahms Symphony No 4, 4th movement (Allegro energico e passionato). Note that there is little of interest below 20 Hz apart from a bit of rumble, and above 20,000 Hz there’s virtually nothing. As is typical of symphony orchestra music, the signal is strong in the mid range (300 to 3000 Hz) with significant bass down to around 40 Hz and smoothly tailing off at the high end before it disappears at around 20,000 Hz.
brahms.png

8

Yep I read some things about that here: http://people.xiph.org/~xiphmont/demo/neil-young.html

So, this a for example the graph from a vinyl I recorded in 24/48:

I guess the rip is okay in this case.

Here the graph from the Daft Punk rip I posted below in the thread (24/96):

Yup it’s really ridiculous. Just one thing I don’t understand: why a 24/48 file show the range from 0 to 23 000 Hz and a 24/96 file show the range from 0 to 48 000 Hz? :smiley:

9

You may notice that the spectrum plot in my image has the horizontal scale set to “logarithmic”. This is more like the way that ears hear. On a logarithmic scale, octaves are evenly spaced horizontally,

The maximum frequency that can be represented in PCM digital audio is half of the sample rate (known as the Nyquist frequency). So for 44100 Hz sample rate the maximum frequency range is 0 to 22050 Hz, for 48,000 Hz sample rate it is 0 to 24,000 Hz and for 96,000 Hz sample rate it is 0 to 48,000 Hz. In practice the upper limit is a little lower due to anti-alias filtering (http://en.wikipedia.org/wiki/Anti-aliasing_filter). At the low end the frequency is limited by analogue components (if it were possible for a loudspeaker to produce 0 Hz, you would feel a constant breeze flowing from it :wink: )

10

So, if I understand well, a CD quality is like the best we can find (16/44.1)? Because on the link I pasted below, they say 24 bits don’t help in playback. Seriously? I’m shocked. Because I listened to some vinyles rips and even worst… I bought some Studio Masters (on a french website which sell them - because I’m french, the equivalent for english people is HDTracks I presume), and the majority of people find the sound better in 24 bits. So, what the point about this, its it just psychological? :astonished:

2nd question: Why the vinyl sound better? Or is it just psychological too?

Just for the fun, here’s a comparison of a track I have both on vinyl and FLAC (bought on Internet). They have bot the same mastering because they have both the same Dynamic Range score of 13.

FLAC 16/44.1

WAV 24/48 vinyl rip

3rd question: What is your analyze from these two graph for a comparison, which sounds the better or there is no difference?

Thank you both to take some time to answer to my questions, I’m a newbie in this domain. :neutral_face:

11

My understanding is the 16/44.1 is sufficient for high quality playback. There have been several extensive studies using double blind testing and they all agree that people cannot tell the difference between a good quality recording played back as 16/44.1 or higher format, when played at moderate to loud volume level. If very quiet music is played back at an extremely high playback level, then some listeners are able to distinguish improved quality at 24 bits, but watch out for the loud bits - you could damage your hearing.

Extra bit depth does no harm (except for the additional file size). 32 bit float is an excellent format when processing audio because of its extreme accuracy and it is virtually impossible to clip (except in the playback system).

Back in the days when CD audio was invented, 44.1 kHz was barely enough. The old anti-aliasing filters were not that great, so the high frequency response would typically start rolling off at around 16 kHz. Modern digital filters are much better, and even cheap CD players are capable of accurately producing frequencies very close to 20 kHz (which sadly I am no longer able to hear, though I can still measure them :wink:)

I have come across some 24/96 systems that sound better at 96 kHz, and have come across other 24/96 systems that sound better at 44.1 or 48 kHz, so to some extent there seems to be some dependence on the implementation. Some sound cards seem to have a “sweet spot” - settings at which they work best. There are some 24/96 sound cards available that sound much worse than equivalently priced 16/44.1 sound cards.

The psychological effect should never be underestimated - it is a very real effect. It has been clinically proven that the placebo effect has medical benefits :astonished:

Because vinyl IS better. :smiley:
Have you ever met anyone that loves their CDs?
People may love “the music” on their CDs, but no-one loves CDs in the way that people love their records. There’s something about the ritual of playing a record that makes it all the more worthwhile - you don’t just grab it off the coffee table and stick it in the player. You handle it with care, place it on the turntable and lower the arm, then sit back and appreciate it. It’s like the difference between a painting and a photocopy.

You can only tell so much from a frequency plot - the real test is in the listening, but I’d guess they both sound pretty similar.

12

Thank you for all these answers and take the time to explain. I feel somewhat scammed with the 24 bits music we can find on the net, and now with my audio interface. Hopefully it have some nice D/A decoder so it’s still ok, and it’s useful to listen to 24/192 ridiculously encoded rips from internet :stuck_out_tongue:

I said I feel scammed because the majority of actual HDTracks/Studio Masters sold on Internet have the same clipping than the quality CD ones, so it’s exactly the same Dynamic Range between both, but not the price… :open_mouth:

Not really, the vinyl is less clear than the original master converted to 16/44.1 (so which is theorically perfect) but it is much warm and light between the different sounds (instruments or things like percussions and bass, etc). I just bought a decent phono ampli (Rega Mini Fono A2D) which can directly convert to 16/44.1 by USB so it’s nice because more is not useful if I understood all of the stuff.

In fact the #1 reason I wanted to play vinyles it’s because of the Loudness War you must know I presume. I can show an example of two tracks I have both on vinyl and FLAC CD quality bought on Internet.

FLAC 16/44.1, Dynamic Range score of 6

WAV 24/48, Dynamic Range score of 10

You can see the clipping now on CD and 16/44.1 FLAC sold on Internet, it’s awful. I can hear some CRRcRR when I listen to them, on clipped moments… Only for a volume war. They forgot that people don’t listen to CD on smartphones and others stuff. So why not make a decent mastering for CD’s like in '90, and a worst one for mp3?

The good fact with vinyls is the source is low and I can normalise the volume to whatever I want without clipping. More, the Dynamic Range score is always better on vinyls than CD, at least, if the master for the vinyl have no clipping, there is no reason the sound will be bad on the vinyl (same for the CD anyway), but the vinyl is still save from this for the moment, although they begin to reduce the mastering quality on them too. :frowning:

Examples with Katy Perry, you can see the difference from vinyls rips. (ok Katy is not the best reference but I needed an actual pop celebrity which we can guess her music will be awfully clipped). Same for Lady GaGa for example.

So yeah, the vinyl is not yet dead :smiley:

13

If true in a blind test, (so not placebo), possibly an unintentional exciter effect which could occur when converting from a 32 bit master to 24 bit, (32/24 = 1.3333333333333333333333333333333 ) , with no such exciter effect converting from a 32 bit to 16 bit because 32 is a multiple of 16 (32/16=2).

14

That’s not logical. 32 bit divided by 16 bit gives 16 bit and 32 bit divided by 24 bit equals 8 bit because the bit represents the exponent for numbers with base 2.
This means that 16 bit to 32 bit introduces 65536 new possible values inbetween the already existing 65536 levels. From 24 to 32 bits, 256 levels are introduced for each bit. In other words, the grid for 24 bit is 256 times finer than the one for 16 bit. the rounding can go in each direction.
It highly depends on the audio content, how the conversion will influence the dynamic domain.
One could possibly test the effect with lower bit formats, such as 4/12/20 or 8/16/24.

15

This recent thread contained a discussion related to this query: https://forum.audacityteam.org/t/record-vinyl-to-laptop/31206/1

As I stated there, for my money the fabled “warmth” of Vinyl versus CDs comes from distortion inherent in the cartridge/arm/phono pre-amp in use. Some folks also thing the valve (tube) amplifiers provide more “warmth” than a solid state amp - I don’t hold with that either.

WC

16

Reducing the bit-depth can add harmonics, (which could be perceived as an improvement on a dull original) …

17

That’s a fact. In the extreme case (bit depth 3), You’ll have a square wave, even if the original was a sine. But I doubt that a conversion to 24 bit produces more exciter effect than a conversion to 16 bit. That’s my point.

18

Now there’s something I can (partly) agree with you on.

It’s not all bad - I have a re-mastered Donovan CD which is breathtakingly excellent.

On the other hand when I bought Emmylou Harris’ Luxury Liner remaster on CD it was dreadful - they had pulled all the instruments forward and pushed Emmylou back into the mush. It made me rush off and convert the vinyl copy I had - after a bit of post-capture processing I got a much better result.

And don’t get me started on the re-mastered Beatles CDs grrrrrr :frowning: :unamused:

WC