[Amanda Ghassaei] has created an awesome hack for making your own vinyl records using a laser cutter from an MP3 file. Her excellent hack uses a Processing sketch that converts a digital audio file into a vector graphics file, which is then burned onto vinyl using a laser cutter. We saw a demo of this at the FabLab11 conference, and it’s an impressive hack.
One of the best parts of her write up are the details of how she arrived at the appropriate processing settings to get the record sounding as good as possible, but still be cuttable. It’s an object lesson in how you iterate on a project, trying different approaches and settings until you find the one that works. She also decided to take it a few steps further, cutting records on paper and wood for the ultimate eco-friendly record collection.
Audiophiles should avoid this technique though. Due to limitations in the resolution of the laser cutter, [Amanda] ended up having to reduce the bandwidth of the audio signal to 4.5Khz and use a 5-bit sampling depth. That translates to a rather tinny-sounding record. Vinyl record snobs can breathe easy: this isn’t going to replace their beloved white-hot stampers. For the rest of us, there are always records etched into tortillas.
From MP3 file????
from M P 3 file???
seriously?
Do you feel like using different sound source would improve the sound much?
(“I buried paul”)
not no. another paul ;)
it did say that on the record, didn’t it? or did I dream it?
To cut a 5 bits 4.5 KHz record (these days, 24 bits 192 KHz is not uncommon), MP3 is more than enough ;-)
To expand, if you do a double-blind test to see whether you can tell audio formats and bitrates apart, you’ll find that as the bitrate gets higher, it’s indistinguishable from lossless. I don’t mean “I can’t tell the difference and therefore assume no one else can” either: people have done this test on multi-hundred-thousand-dollar club setups, pro studio equipment, and so on.
In fact, you can tell even 320kbps MP from lossless audio easily. MP3 seriously fails at reproducing wideband noise (thousands of people clapping at end of song, for example).
When I listen to a song I’m not trying to determine if it was played at Glastonbury or Madison Square Garden. Who cares about the audience, does the group sound good.
Thanks for pointing that out, MS-BOSS.
Leithoa– For some of us, the distortion described here is a lot like someone pulling their nails across a chalkboard.
The thing that gave MP3 a bad name was the early 128kbps “standard” that was used. It was a good trade off back when it was first established, storage was super expensive at the time and processing power was also barely up to the task. However, by the time “trading MP3s” got popular, it was already time to move up to 256kbps or more, but inertia of the 128kpbs standard was high. The quality of 128kbps is bad enough that you can perceive the distortion in music transmitted on ordinary FM broadcasts, that’s “too bad” in my book. By the time you take it up to ~256kpbs VBR and the more modern encoding schemes, it’s near indistinguishable when imprinted on the finest vinyl.
@Reynald
Serious, non-troll question: the range of human hearing tops out at around 20kHz. It seems to me that anything about 40kHz sample rate will keep anything we can hear safely below the nyquist frequency, so what’s the point of sampling at 192kHz?
Nyquist frequency, nyquist rate. Whatever, you know what I mean. :V
192kHz is considered bad by some, scroll down to the “192kHz considered harmful” section, although the whole thing is well worth a read.
https://xiph.org/~xiphmont/demo/neil-young.html
@Truth
Thanks for sharing that article!
There is nothing worth a read from the xiph.org guy. What a crock of bull.
@Rob Can you explain exactly which part you find wrong ? I read it all and it makes total sense to me, I can’t see any flaw in the logic.
Rob is still trying to offload his PonoPlayer. Don’t ruin it for him.
@Tony If only it when up to 300kHz, then at least then oceanographers could use it for recording Beluga whales and Bottlenose dolphins. But still it is high enough that you could record cats and rodents ( https://en.wikipedia.org/wiki/Hearing_range ).
Reconstructing an exact <=20 kHz signal from a 40 kHz sample rate is complicated. It would require very sharp filters. Doing the same thing from a 192 kHz sample rate is much easier.
one reason is to shift the ‘work’ needed up way past audio cut-off freq, mostly because the filtering is harsh if its too close to 20khz, but if its way above the human hearing cutoff point, the filtering (and dac hardware) can do a better job. that’s mostly it.
that said, there are instruments that do go way above 20khz and some people have claimed to be able to detect some presence (not sure if I would call it sound) in the upper ranges.
finally, the reason STUDIOS should sample that high is the same reason you shoot photos at raw mode, process in raw mode (or at least 16bit /pixel color) and once you’ve done all your edits with full math precision, you THEN dither down (so to speak) to 8big jpg for the user to see. all the edits at high precision cause less rounding error (etc) as you process. you would not process photos in jpg (you shouldn’t!); if you get a jpg file, immediately convert it to 24bit color, then do your edits and convert down before you save at 8bit jpg.
for me, when I see an 88.2k file or 96k file, I have -some- confidence that it was redone with better a/d systems and given more care than when it was done 20 yrs ago. but I could take those 24/96 files and convert to 44.1 and have just the same sound (very close) and not waste so much disk space.
btw, dsd is a new thing (somewhat new) and that makes the dac’s life a WHOLE LOT EASIER since its pulse width modulation instead of pcm and you almost don’t even NEED a dac to convert dsd (ie, sacd) to analog. a simple lowpass filter is mostly all you need, so its quite different from pcm and its ‘bitrate’ does not even map to the same concept.
could you email me and explain how you do it? williamhallstrom@gmail.com
Stick a mirror to a speaker, stick it to the laser head, have the laser draw a spiral at the right speed and play your audio through the speaker. No software processing required for the audio, only focusing and power for the laser… If only it were that easy.
Why isn’t it that simple? I think it would work brilliantly if tweaked well.
It’s a friggin’ wooden record, do you really think lossless would have made it sound better!?
“burned onto vinyl using a laser cutter”
Which is toxic (vinyl) and corrodes the laser cutter?
If you read the original posting you’ll nitice that “vinyl” isn’t even mentioned, just wood, acrylic, and paper. The vinyl part was dreamt up by the HAD author.
Silly people expecting HAD ‘editors’ to actually write a decent summary. Most of them can barely read, maybe one day they’ll understand some of the words.
Still, they did manage it better the last time they mentioned this (now two years old) thing: http://hackaday.com/2014/11/19/laser-cut-album-released/
As an aside, no-one worries much about the fumes being toxic. What really worries us is it makes our lasers go rusty!
Interesting proof of concept but it has a long way to go before becoming usable.
I do believe you’ve missed the point slightly.
Better sound quality than http://hackaday.com/2012/12/21/3d-printing-records/?
Almost think the 3D printed one sounds better.
It’s the same woman. Dunno why she’s spent 3 years trying to make unplayable records though.
Why not?
Because she can?
Practice at making high-resolution objects?
Sell them to hipsters?
Sell them to audiophools?
Send a few to Neil Young for the lulz?
Everyone needs a hobby?
You could invert the hack–use magnets on/under the turntable, and a magnet on the head to float the head, and use a laser (or two) to read the record. It would help the record last much longer.
Then replace the record with a thumb drive and it’ll last even longer!
You can get laser turntables, for people who just can’t throw their money away quick enough. Or for archiving delicate and irreplacable sound sources.
If I had a good MP3 file why would I want to make it into a record? I spent years making my records into MP3s.
I’ll admit even I’ve asked myself a couple times why people do certain projects but this site deals with people who are more interested in can it be done instead of should it be done and I cant argue with that.
And then much bigger hard drives came out, and Internet bandwidth multiplied by a few thousand, and suddenly you realised how crap all those 128K MP3s sound.
RIAA takedown in 3.2.. home laser cutting is destroying the music industry!
…and the rainforest!
Do you know if anyone’s able to do commercial standard stereo vinyl properly?
hey smartguys when we say 192khz recording, is sampling rate not audible sound
I’m no laser-cutting expert, but i bet she could have gotten better resolution (and sound) by letting the substrate spin beneath a stationery (or possibly single-axis) laser.
There’s much better songs on that album.
Would love some answers to the below questions if anyone has some info for me:
1. Why can laser cutting not produce industry standard stereo width vinyl?
2. Is it down to inability to produce accurate enough cutting patterns (that this somehow will not render accurate enough cutting templates/designs) or is it down to the limitations of laser cutting (that the machines somehow cannot cut with enough precision/perform the task/does it have to do with vertical+lateral cutting)?
3. Practically, is it POSSIBLE to laser cut stereo vinyl, with two channels of audio in one groove?
4. What is the challenge in doing so? Is it down to hardware limitations, or is it simply not possible to do for other reasons?
5. In the case of hardware limitations, which feature do current laser cutters lack?
6. PVC is not safe for laser cutting – is there an alternative, laser safe material which could serve as a substitute for producing industry standard stereo vinyl?