Printing phonograph records on a 3D printer
Amanda Gassey, an employee of instructables.com, has done an impressive job of creating 3D printing technology for gramophone records. There is still little practical benefit from this, but the detailed history of her research and experiments, laid out on instructables, reads like a fascinating detective story. The resolution of the best modern 3D printers is barely enough to achieve sound quality comparable to the first audio recordings of the 19th century, but the task is even more interesting.
Amanda had a pretty good quarter-million-dollar Objet Connex500 printer at her disposal , capable of printing at 42 microns in the X and Y axes and 16 microns in the Z axis (600x600x1600 DPI). After numerous experiments, it was possible to achieve sound reproduction with a resolution of 5-6 bits and a sampling frequency of the order of 10 kilohertz. The sound is accompanied by a characteristic periodic noise created by the raster of the printer. The gramophone surface model has such a huge number of small details that while the recording time is limited to one minute, it no longer fits in the printer's memory.
The model itself is generated as follows: a Python script processes the sound file, producing a textual sequence of sample values. This sequence is fed to the Processing library.whose main purpose is the construction of interactive graphics and diagrams. Using the ModelBuilder library , the model created in Processing is converted to the STL format that the printer understands. Before processing, the sound is compressed and equalized to fit it into the tiny dynamic and frequency range of a homemade record.
A detailed description of the process, with all the technical details and nuances, with the source code of all scripts and links to ready-made models of several songs by Nirvana, Joy Division, Daft Punk, Radiohead and others - on instructables .
Amanda had a pretty good quarter-million-dollar Objet Connex500 printer at her disposal , capable of printing at 42 microns in the X and Y axes and 16 microns in the Z axis (600x600x1600 DPI). After numerous experiments, it was possible to achieve sound reproduction with a resolution of 5-6 bits and a sampling frequency of the order of 10 kilohertz. The sound is accompanied by a characteristic periodic noise created by the raster of the printer. The gramophone surface model has such a huge number of small details that while the recording time is limited to one minute, it no longer fits in the printer's memory.
The model itself is generated as follows: a Python script processes the sound file, producing a textual sequence of sample values. This sequence is fed to the Processing library.whose main purpose is the construction of interactive graphics and diagrams. Using the ModelBuilder library , the model created in Processing is converted to the STL format that the printer understands. Before processing, the sound is compressed and equalized to fit it into the tiny dynamic and frequency range of a homemade record.
A detailed description of the process, with all the technical details and nuances, with the source code of all scripts and links to ready-made models of several songs by Nirvana, Joy Division, Daft Punk, Radiohead and others - on instructables .