I've got an "audiophile" turntable, belt driven and i have to take the platter off to change the speed from 33 to 44.. 99% of my records are 33rpm so thats where it stays..

I just did a vinyl rip on a new 7" EP i had.. only by side B did i notice it was probably a 44rpm record.. having ripped it to test it i did a pitch shift up by 1/3 and listened.. it sounded much better...

I'm just wondering if this is a legitimate way of ripping 44rpm discs... I imagine the RIAA equilisation will be applied to parts of the music at 33rpm that it wouldn't normally be done at 44rpm, such that when i then pitch shift it its not going to be quite right... not sure tho..

I'll change the player over and rip properly at 44rpm and see if i can ABX it, or at least hear a difference..

If this wasn't the case, surely 33rpm would be better as it would allow you to rip more carefully at the slower speeds, or is it the opposite that 44rpm actually sounds better in terms of what can be stored in it..

Anyone got any thoughts?! Cheers

Jim

I can not help you with the ripping question exactly. However, I do know that some audiophile quality remasters (Mobile Fidelity Sound Labs specifically) are released on vinyl that runs at 45rpms. So I assume there is some increase in quality if they are doing this. But they also use half-speed mastering.

Yes, the eq will be wrong. It should be correctable witha software filter if you can find it.
The error may be less than what people routinely do with equalizers anyway..

The other thing.. 44 rpm? The norm would be 45 rpm, and you were playing at 33 1/3.

I am in a similar situation. I use a flat preamplifier. Audacity has specific tools to correct among 33-45-78 speeds. It also has the ability to add RIAA eq after the file has been speed corrected. This works well for me. Might work for you too.

Remember that he already has the eq applied (in the preamp), but at the wrong frequencies.. so the tool would have to undo that then reapply at the right frequencies.

Hmm. Would you actually be able to get better quality by playing it slowly and then converting/applying software EQ, than if you ripped it at the right speed? Just curious.

Probably not. The main advantage of 45rpm over 33rpm is that it allows for better bass response (this applies only to 12" singles; 7" vinyl is typically pants). Most 12" 45rpm records take advantage of this, so there tends to be a fair amount of bass energy. If you play them at 33, that bass energy reduces in frequency, and there's a danger that it might get close to the arm/cartridge resonant frequency, which you definitely want to avoid.

As for the problem of inaccurate EQ... Let's face it, vinyl frequency response isn't exactly precision stuff. It's likely that after incorrect EQ has been applied, the only thing that you might notice would be a slightly altered frequency balance, and a little after-the-event EQ applied by ear will probably sort that out satisfactorily.

But the bottom line is that it's just easier to play vinyl at the speed intended.

Thanks everyone..

I did find it quite interesting messing with it.. and the end result aint terrible.. i will re rip at the correct speed..



you're right of course... it was late and i rushed the post.. the turntable doesn't have anything with it written on it like it would if it had a slider like some turntables... i took a guess.. and got it wrong

THanks for some more details on the whole ripping slowly being better/worse... it struck me as quite interesting...

Cheers

See the attached gif for an estimate of the frequency response error you've introduced by recording a 45rpm disc at 33rpm, and then resampling to achieve the correct speed in software.

The cut after 18kHz is due to the way I simulated this - you won't see it in practice. However, you will see the reduction in bass, and the gentle ripple in the mid-range. That's before you take any arm and cartridge lower frequency limits into account, which will make the bass loss worse.

As someone suggested, the cartridge response is probably worse than this, though it may already have a slight V-shaped EQ in there anyway, to which you're adding. Whether the bass loss is audible or not depends on the record and your Hi-Fi.


I know it would have been quicker for you to change the speed on your turntable than it was to calculate this error, but I was interested!

One day I should calculate the correction needed to play a Westrex pre-emphasis curve 78rpm disc via an RIAA EQ at 45rpm and still get the correct result on playback!

Cheers,
David.

David,
your frequency analysis doesn't show the reference, so we actually can't compare.
How did you calculate it ? If I shall guess: You generated some white noise, applied an EQ curve according to the RIAA curve and resampled the result to make it faster and higher at the same time ?

Pretty much - I stored the RIAA pre-emphasis and de-emphasis (approximately) as CEP FFT presets. I generated a linear tone sweep 20Hz-24kHz noise at 48kHz 32-bit sampling and applied the pre-emphasis. I set the sample rate to 35556Hz then resampled to 48kHz, applied the de-emphasis, and plotted the frequency analysis.

The reference is flat (the original material, if you like, before cutting the 45rpm disc) so what you see is the frequency response error. A -6dB tone sweep over 10 seconds analysed using those parameters in Adobe Audition gives a flat line at -36.1dB, the figures (including the 30dB offset I entered by hand - see the capture) are relative to this. However, the overall volume change isn't that important unless you're worrying about the SNR issues involved, which could be the least of your worries!

The RIAA curve I entered into CEP was only approximate, so don't take these results as being too dependable.

Cheers,
David.

At this point you have 'what's on the LP'...

Why isn't it "I set the sample rate to 35556Hz, applied the de-emphasis, then resampled to 48kHz" ?
I'd like to reproduce this to fully understand it, but don't know where to get the RIAA equalizer data...

google shows this data up, and it can be calculated as well.

i made mine using spline curves, so it may or may not be accurate, then made an impulse that i can use in foobar's convolver to correct 45-33 error, so i could play back and correct at the same time.

if i can find the impulse, i'll post it.

btw, in my experiments i didn't find any particular advantages in undercranking a recording, except that my crappy cartridge didn't drop-out as much on the slower speed.

I tried to retrieve the RIAA curve in an obscure way:
Remember that EAC can detect if a CD has been pressed/mastered with pre-emphasis (Obviously there's some special bit for each track on the CD (like the copy bits), carrying this information.) I haven't yet encountered any factory-pressed preemphasis CD. EAC can not only detect the preemphasis bit, it can also write it. I had once found out that the DACs of my CD player and MD recorders react to this bit and apply then a de-emphasis. (Obviously the preemphasis bit is also tranferred through SPDIF). So by recording the analog output to harddisc I can retrieve the deemphasis curve (After burning an audio CD with no preemphasis but with preemphasis bit.)
Unfortunately, this method doesn't seem to be that precise...Anyway, I got no sensible results.

I'm still wondering about the question, why this preemphasis for LPs is used at all. Is it similar to dolby noise reduction for cassetes ?

But the preemphasis on CDs is not the same as the RIAA eq found on vinyl records. That explains why the results you got were imprecise (or should we just say "wrong" :-)


It's to overcome the mechanical limitations of cutting vinyl records:

If bass frequencies were cut "flat", the grooves would wiggle around so much it would be impossible for any cartridge to track them (and the playing time would be seriously restricted). So the bass is reduced. (The bass is also mono-ed, to prevent large out-of-phase bass signals causing the cutter head to completely leave the acetate - which would result in a gap in the groove and consequent chaos for the playback stylus, but this isn't part of the RIAA eq per se).

If the treble frequencies were cut "flat", the vinyl surface noise would overwhelm them, so they are boosted in order to get them above the noise floor. In this respect, the boost applied to high frequencies is a little bit like Dolby for tapes, in that its purpose is to get them above the noise.

Ah, well spotted - I resampled to 48kHz because I'd stored my RIAA FFT EQ at that sample frequency (CEP doesn't always transfer FFT settings saved at one sample rate to another correctly, so I didn't apply it at 35556Hz just in case). At the end of the process, I have what you'd get by recording the 45rpm disc at 33.3rpm.

However, I should then have adjusted the sample rate to 64.8kHz and resampled back to 48kHz (or whatever) to get back to the correct speed (45rpm!) - in other words, I didn't finish the job, so the frequency axis labels on my plot are incorrect - multiply each one by 1.35 and it'll be right.

I used splines from a few fixed points to do the job approximately, since the straight line response often shown in articles and textbooks is nothing like what happens in practice.


The other interesting part about the RIAA curve is the wiggle in the middle. Cliveb's explanation in the last post is spot on - because the movement of the stylus gives an electrical output proportional to the speed of the stylus/cutter movement, rather than its displacement, you need to do something to stop the bass dwarfing the treble when the record is cut, and undo this when it is played.

However, just solving this problem would give a "constant velocity" curve of 6dB per octave right across the band - and that's not what the RIAA curve is. It puts a little less treble onto the record, and boosts the treble a little more, than true constant velocity emphasis - which gives the treble a little more headroom.

There must have been some great discussion about the ideal figures at the time - I don't suppose anyone is reading this who can provide some insight? I think it’s interesting, because many early or cheap cartridges cannot cope will with the treble levels created by the RIAA curve (how often do you hear distorted Sss on records?), but putting even less treble onto disc (and boosting more on playback) would have reduced the SNR at the expense of increased dynamic range.

Cheers,
David.

this page here helped me make my cool edit FFT curve (in a similar manner to 2bdecided it seems).

not sure how good the data is, but the output sounds pretty good, so i guess it works

http://sound.westhost.com/project25.htm