A vinyl master gets bass summed to mono below roughly 150Hz, sibilance controlled harder than a stream would ever need, and a final level set by the cutting engineer rather than a limiter, because the lathe is a mechanical system with real physical limits a DAW does not have. None of that applies to a file bound for Spotify or Apple Music. The two masters are not the same file at two loudness settings. They are two different jobs.
Why does bass need to be mono below roughly 150Hz?
Because a cutting lathe steers a physical stylus left and right to encode a stereo signal, and out-of-phase low frequency content sends that stylus on a vertical excursion it was not built to sustain. Wide stereo bass on vinyl reads on the cutting head as the two channels fighting each other, and the groove can swing deep enough to cut through into the next groove or make the needle jump on playback. Keeping bass mono below the 100 to 200Hz range, with 150Hz a common working point, removes that fight while leaving the stereo image untouched everywhere it is audible. Most engineers apply this with a simple low-frequency mono filter on the master bus rather than trying to catch every bass element individually during the mix.
What does the RIAA curve actually do to the signal?
It reshapes the frequency balance specifically so the groove can hold a usable amount of music. The cutting engineer applies a pre-emphasis curve that reduces low frequencies and boosts high frequencies before the signal reaches the stylus, cut into the vinyl with less bass energy and more treble energy than the source. A turntable's phono preamp applies the inverse curve on playback, restoring a flat response. The practical reason for this trade is groove geometry: bass modulates the groove wide, and if it were cut at full level, records would hold only a few minutes of music per side. The RIAA curve, standardized industry-wide, is what makes a normal-length album side physically possible.
Why can't a brickwall-limited streaming master go straight to the cutting lathe?
Because the lathe has no equivalent of a loudness-normalized playback path to hide behind. A streaming master pushed hard into a limiter loses the transient headroom and micro-dynamics that give a cutting engineer room to work, and a mastering house asked to cut that exact file usually has to turn the whole record down further just to keep the groove from over-modulating and distorting. Furnace Record Pressing, a working pressing plant, advises sending a separate, less-limited master for vinyl rather than the same file used for digital release, specifically because a lathe calibrated for clean playback cannot cut as loud as a limiter can print. The two-master approach costs one extra bounce at the mix stage and avoids a cut that sounds thin and squashed next to the same song on a DSP.
How does side length change how loud a cut can be?
Directly, because loudness and playing time compete for the same physical groove space. A short side gives the cutting engineer wider spacing between grooves, which allows deeper, louder modulation and more low end. As a side gets longer, the same amount of vinyl has to hold more music, grooves pack closer together, and the cut gets quieter to avoid the walls of adjacent grooves touching. A 12-inch LP at 33⅓rpm generally holds 16 to 20 minutes per side before quality drops off, and cutting studios such as Diz Lathe Cuts size a release's tracklist against that ceiling before a single sound gets cut. A 7-inch single at 45rpm trades playing time for exactly the opposite: less music, more physical distance per rotation, and a louder, more detailed cut as a result.
Why does sibilance cause more trouble on vinyl than on a stream?
Because high-frequency energy modulates the groove wall at the shortest, most delicate wavelengths a stylus has to track, and a harsh S or T consonant can cut a groove excursion sharp enough to make the playback needle jump or distort. A de-esser that would be optional on a streaming master becomes closer to mandatory before a vinyl cut, tuned specifically to the sibilant range rather than a general high shelf, so the rest of the top end stays intact. Cutting engineers routinely send a test cut back with exactly this note: a vocal that sat fine in every digital release needs its sibilance tamed before the lathe will accept it cleanly.
What should actually get delivered to a cutting engineer?
An unlimited or lightly limited 24-bit WAV at the mix session's native sample rate, a mono-summed low end below the agreed crossover point, sibilance already under control, and a short note on intended side length and RPM so the cutting engineer knows the loudness ceiling before starting. Track order matters too: the loudest, most low-end-heavy song on a side is usually placed first, closest to the outer edge of the record, where the groove has the most room to move before the geometry tightens toward the center. Enormous Door Mastering sequences a tracklist against exactly this logic before a lathe session is booked, not after a test pressing comes back thin.
A reference mix bounced flat, without the streaming limiter chain, saves a round trip later. Sending only the finished digital master forces the cutting engineer to undo processing that was never meant to survive contact with a groove, and that undo work shows up as extra cost or a cut that falls short of what the songs could sound like on the format.
If a release is heading to vinyl
Budget a separate print from the streaming master and send it to the cutting engineer with the side length and RPM already decided, not left for them to guess. Full credit list at /credits, background and contact at /about.