See https://pspatialaudio.com/stylus_grooves.htm for physical parameters of a phonograph record. For a 12” modern disc, a 10 kHz tone at the outer groove has a wavelength of 51 microns (micron = 1x10^-6 meter), and at the inner grove a wavelength of 20 micron. To resolve this frequency you need to sample at least twice per wave (Nyquist), and surely there will be harmonics of lower frequencies that are needed as well. So let’s say you have to sample at the inner grove at better than 5 micron. That’s crudely a resolving power of 200 lines per mm.
The best film for grain is Fuji Velvia, roughly 100 lines/mm (see https://www.kenrockwell.com/fuji/velvia100.htm). It’s possible with a digital camera with the newest sensors, say a Nikon D850 (35 megapixels IIRC, and likely twice this resolution is now available) to approximately equal the performance of Velvia. In fact a D850 can be used as a negative scanner, a decade ago a mechanical scanner was necessary. Of course such a camera only images a 35mm frame size, so it would have to be translated across the entire phonograph record. For film scanning a macro lens is used at 1:1 resolution (i.e., the object size on film equals the actual object size) - so no cheating by imaging the entire disc, and your resolution goes down by a large factor (how many 35mm frames fit on a 12” LP?). Depth of field will be an issue as well. Perhaps with bright enough illumination a very small aperture could be used, but I still doubt that given how non-flat LPs can be on the millimeter scale that adaptive focusing would not be needed. Also, to resolve 3D features (for stereo and other) depth of field actually hurts.
Ordinary desktop scanners are 600 dpi, or crudely a resolving power of 24 lines/mm. So way off what’s needed, although more expensive scanners up to “interpolated” 4800 dpi or even 9600 dpi are available. But watch out for the meaning of “interpolated”. There are drum scanners for photography that go all of the way up to 11000 dpi, just getting into the ballpark. So optically it meets the need, but not mechanically (the photograph to be scanned is wrapped on the drum). A flat bed scanner is in very close proximity to the scanned paper, so flatness of the LP hurts there too.
A DVD uses pits of .85 microns to 3.5 microns, so not that much more difficult in terms of feature size than an LP. But a DVD (or CD, or BluRay) has huge advantages in that the data is digital - so only a 1 or a 0 must be resolved, not a continuous wave - and includes error correction. Plus the surface that is read is very reflective, so saturated photodetection is all you need. Optical drives are actually quite remarkable, since they have a cheap plastic lens with little depth of field; they use active servos for both focus and tracking, and the specs on flatness of a CD or DVD are pretty loose. If you take covers off an optical drive and watch the drive gears and optical head you’ll be amazed at the amount of movement. Now think about amplifying that movement with a 12” diameter rather than 5.25”.
I think it would actually be a terrific engineering project for a student to modify an optical drive’s mechanism to see if tracking and focus could be maintained with an LP, then progress onto improving the optics and electronics to resolve the analog signal from a record. That approach likely has a greater possibility of success than a 2D scan.
Last edited:
