3-08 Magazine Article "Vinyl Mastering"

[saxonrecording]

For whoever is interested...just read an article in the new issue, March 2008, of Electronic Musician "Mastering Vinyl" by Gino Robair. Some neat tips, facts and pictures.

[bancho]

Could you scan the article and send it by e-mail? :-}
I'd be very thankful!

[saxonrecording]

Here's the link for the magazines online issue.
emusician.com/
The March issue is not yet online but keep a lookout for it.

[Steve E.]

http://emusician.com/tutorials/mastering_vinyl/index1.html

[cd4cutter]

I'd like to correct what is apparently a GROSS misconception regarding the RIAA equalization curve that is applied to all modern analog disc recordings. The sited article in EM magazine just perpetuates this myth. This quotation from the article heading titled "The RIAA EQ Curve" is just plain WRONG!

". . . the EQ curve gradually rolls off the bass by 6 dB per octave starting at 1 kHz so that by 20 Hz, the level has been reduced 20 dB. Without the RIAA bass cut, only about 5 minutes of low-frequency information could be stored per 12-inch side."

There is NOT a significant bass cut applied during cutting. There is actually a BOOST. See details below. Nor is there a significant frequency change ANYWHERE in the response curve. The "typical" RIAA curve that is often published (and which is grossly misunderstood by the author of the above magazine article) is usually the EQ required to compensate a MAGNETIC phono cartridge for correct RIAA playback EQ. Consequently, this published data includes the 6dB per octave (dB/oct) treble rolloff (downward slope of 6dB/oct throughout the entire frequency range) that is required to convert the constant velocity response of a magnetic pickup to be constant amplitude which is what a crystal or ceramic cartridge is. Constant amplitude response means that the peak to peak size of the recorded wave is the same at all frequencies, i.e. it has constant amplitude at all frequencies. It is this 6dB/oct slope which accounts for most of the apparent change in response from low to high frequencies.

The actual RIAA contribution to deviation from constant amplitude response in the RECORDING is a 6db/oct ROLLUP (yes, it's a BOOST) below 50Hz (also known as the 3180 microsecond [usec] turnover). This amounts to only about 6dB of BOOST at around 25 Hz and 3dB at 50Hz deviation from flat, constant amplitude response. The purpose of this boost in the recording was apparently to reduce rumble when the boost was canceled during playback. Then the RIAA recording EQ is flat up to 500Hz where it is turned downward by about 3dB (the 318 usec turnover point) and then turned back to flat again at around 2150Hz (the 75 usec turnover). These two inflection points amount to introducing a shelf in the response between 500Hz and 2.15kHz which lowers the level to about 5dB less at 10kHz than it is at about 200Hz. The purpose of the HF cut was to reduce strain on the cutterhead as well as reducing "slope overload" or tracing distortion of conical playback styli at high frequencies. So cutting a master actually BOOSTS the bass by the RIAA curve from what it would be if it were cut with no EQ and, thus, with a constant amplitude response. Also, the treble response is slightly lowered from flat, constant amplitude response. So the "excuse" given for the treble "boost" during cutting to improve signal to noise ratio on playback is also completely wrong-headed. The treble actually requires a boost during playback to return the RIAA EQ to constant amplitude which aggravates the surface noise of the disc. The reason this was not considered objectionable was that vinyl was so much quieter than shellac that a little extra noise caused by the playback EQ was more than justified by the reduced tracing distortion in the playback. The RIAA playback curve is, of course, the reverse of the recording EQ to provide constant amplitude response at the output of the phono playback equalizer.

[Steve E.]

[EDIT BY Moderator in August 2009:

This claim by CD4cutter was challenged recently in a thread started by esteemed mastering engineer Kevin Gray. Please read it for a debate/point-counterpoint. I do not have the technical know-how to take a stand one way or the other....I'm the messenger. Judge for yourself!

https://lathetrolls.com/viewtopic.php?t=1417 ]

I'm intrigued at the RIAA curve's reduction of high frequencies during cutting, which can increase surface noise.

It sounds like the curve was put in place to account for the limitations of cutting and playback as they existed in the early 33 1/3 era.

What do you think of the RIAA curve? Are there other EQ curves that would make for better sounding cuts on modern turntable equipment, with state of the art lathes? (if such curves became standard, of course.)

[Steve E.]

cd4cutter,

Does that mean that this version of the Wikipedia article on the RIAA curve is wrong?

http://en.wikipedia.org/wiki/RIAA_equalization

RIAA equalization is a specification for the correct playback of gramophone records, established by the Recording Industry Association of America (RIAA). The purpose of the equalization is to permit greater playback times, improve sound quality, and to limit the physical extremes that would otherwise arise from recording analog records without such equalization.

The RIAA equalization curve has operated as a de facto global industry standard for the recording and playback of vinyl records since 1954. Before then, especially from 1940, each record company applied its own equalization; there were over 100 combinations of turnover and rolloff frequencies in use, the main ones being Columbia-78, Decca-U.S., European (various), Victor-78 (various), Associated, BBC, NAB, Orthacoustic, World, Columbia LP, FFRR-78 and microgroove, and AES.

Equalization practice for electrical recordings dates to the beginning of the art. In 1926 it was disclosed by Joseph P. Maxwell and Henry C. Harrison from Bell Telephone Laboratories that the recording pattern of the Western Electric "rubber line" magnetic disc cutter had a constant velocity characteristic. This meant that as frequency increased in the treble, recording amplitude decreased. Conversely, in the bass as frequency decreased, recording amplitude increased. Therefore, it was necessary to attenuate the bass frequencies below about 250 Hz, the bass turnover point, in the amplified microphone signal fed to the recording head. Otherwise, bass modulation became excessive and overcutting took place into the next record groove. When played back electrically with a magnetic pickup having a smooth response in the bass region, a complementary boost in amplitude at the bass turnover point was necessary. G. H. Miller in 1934 reported that when complementary boost at the turnover point was used in radio broadcasts of records, the reproduction was more realistic and many of the musical instruments stood out in their true form.

West in 1930 and later P. G. H. Voight (1940) showed that the early Wente-style condenser microphones contributed to a 4 to 6 dB midrange brilliance or pre-emphasis in the recording chain. This meant that the electrical recording characteristics of Western Electric licensees such as Columbia Records and Victor Talking Machine Company had a higher amplitude in the midrange region. Brilliance such as this compensated for dullness in many early magnetic pickups having drooping midrange and treble response. As a result, this practice was the empirical beginning of using pre-emphasis above 1,000 Hz in 78 and 33 1/3 rpm records, some 29 years before the RIAA curve.

Over the years a variety of record equalization practices emerged and there was no industry standard. For example, in Europe,for many years recordings required playback with a bass turnover setting of 250 to 300 Hz and a treble rolloff at 10,000 Hz ranging from 0 to -5 dB, or more. In the United States there were more varied practices and a tendency to use higher bass turnover frequencies, such as 500 Hz, as well as a greater treble rolloff like -8.5 dB, and more. The purpose was to record higher modulation levels on the record.

Evidence from the early technical literature concerning electrical recording suggests that it wasn't until the 1942-1949 period that there were serious efforts to standardize recording characteristics within an industry. Heretofore, electrical recording technology from company to company was considered a proprietary art all the way back to the 1925 Western Electric licensed method first used by Columbia and Victor. For example, what Brunswick-Balke-Collender (Brunswick Corporation) did was different from the practices of Victor.

Broadcasters were faced with having to adapt daily to the varied recording characteristics of many sources: various makers of "home recordings" readily available to the public, European recordings, lateral cut transcriptions, and vertical cut transcriptions. Efforts were started in 1942 to standardize within the National Association of Broadcasters (NAB), later known as the National Association of Radio and Television Broadcasters (NARTB). The NAB, among other items, issued recording standards in 1949 for laterally and vertically cut records, principally transcriptions. A number of 78 rpm record producers as well as early LP makers also cut their records to the NAB/NARTB lateral standard.

The lateral cut NAB curve was remarkably similar to the NBC Orthacoustic curve which evolved from practices within the National Broadcasting Company since the mid-1930s. Empirically, and not by any formula, it was learned that the bass end of the audio spectrum below 100 Hz could be boosted somewhat to override system hum and turntable rumble noises. Likewise at the treble end beginning at 1,000 Hz, if audio frequencies were boosted by 16 dB at 10,000 Hz the delicate sibilant sounds of speech and high overtones of musical instruments could survive the noise level of cellulose acetate, lacquer/aluminum, and vinyl disc media. When the record was played back using a complementary inverse curve (de-emphasis), signal to noise ratio was improved and the programming sounded more lifelike.

In a related area, around 1940 treble pre-emphasis similar to that used in the NBC Orthacoustic recording curve was first employed by Edwin Howard Armstrong in his system of Frequency Modulation (FM) radio broadcasting. FM radio receivers using Armstrong circuits and treble de-emphasis would render high quality wide-range audio output with low noise levels.

When the Columbia LP was released in June 1948, the developers subsequently published technical information about the 33 1/3 rpm, microgroove, long playing record. Columbia disclosed a recording characteristic showing that it was like the NAB curve in the treble, but had more bass boost or pre-emphasis below 150 Hz. The authors disclosed electrical network characteristics for the Columbia LP curve. This was the first such curve based on formulae.

In 1951 at the beginning of the post-World War II high fidelity (hi-fi) popularity, the Audio Engineering Society (AES) developed a standard playback curve. This was intended for use by hi-fi amplifier manufacturers. If records were engineered to sound good on hi-fi amplifiers using the AES curve, this would be a worthy goal towards standardization. This curve was defined by the time constants of audio filters and had a bass turnover of 400 Hz and a 10,000 Hz rolloff of -12 dB.

RCA Victor and Columbia were in a "market war" concerning which recorded format was going to win: the Columbia LP versus the RCA Victor 45 rpm disc (released in February 1949). Besides also being a battle of disc size and record speed, there was a technical difference in the recording characteristics. RCA Victor was using "New Orthophonic" whereas Columbia was using the LP curve.

Ultimately the New Orthophonic curve was disclosed in a publication by R. C. Moyer of RCA Victor in 1953. He traced RCA Victor characteristics back to the Western Electric "rubber line" recorder in 1925 up to the early 1950s laying claim to long-held recording practices and reasons for major changes in the intervening years. The RCA Victor New Orthophonic curve was within the tolerances for the NAB/NARTB, Columbia LP, and AES curves. It eventually became the technical predecessor to the RIAA curve and superseded all other curves. By the time of the stereo LP in 1958, the RIAA curve, identical to the RCA Victor New Orthophonic curve, became standard throughout the national and international record markets.

RIAA equalization is therefore a form of preemphasis on recording, and deemphasis on playback. A record is cut with the low frequencies reduced and the high frequencies boosted, and on playback the opposite occurs. The result is a flat frequency response, but with noise such as hiss and clicks arising from the surface of the medium itself much attenuated. The other main benefit of the system is that low frequencies, which would otherwise cause the cutter to make large excursions when cutting a groove, are much reduced, so grooves are smaller and more can be fitted in a given surface area, yielding longer playback times. This also has the benefit of eliminating physical stresses on the playback stylus which might otherwise be hard to cope with, or cause unpleasant distortion.

A potential drawback of the system is that rumble from the playback turntable's drive mechanism is greatly amplified, which means that players have to be carefully designed to avoid this.

RIAA equalization is not a simple low-pass filter. It carefully defines transition points in three places - 75 µs, 318 µs and 3180 µs, which correspond to 2122 Hz, 500 Hz and 50 Hz. Implementing this characteristic is not especially difficult, but more involved than a simple linear amplifier. The phono input of most hi-fi amplifiers have this characteristic built in, though it is omitted in many modern designs, due to the gradual obsolescence of vinyl records. A solution in this case is to buy a special preamplifier which will adapt a magnetic cartridge to a standard line-level input, and implement the RIAA equalization curve separately. Some modern turntables feature built-in preamplification to the RIAA standard. Special preamplifiers are also available for the various equalization curves used on pre-1954 records.

Digital audio editors often feature the ability to equalize audio samples using standard and custom equalization curves, removing the need for a dedicated hardware preamplifier when capturing audio with a computer. However, this can add an extra step in processing a sample, and may amplify audio quality issues of the sound card being used to capture the signal.

[Steve E.]

More discussion on this topic here:

https://lathetrolls.com/viewtopic.php?p=3254&highlight=&mforum=lathetrolls#3254

[cd4cutter]

Steve, I'm just now discovering your post and quotation from the wikipedia description of the RIAA compensation. Yes, there are errors in this description. The history of the various attempts to equalize recording and playback thru the early years of electrical recording are probably fairly accurate (I'm not going to put the effort into looking up all this stuff to verify it), but the author PERPETUATES the myth in his description that the RIAA curve cuts the bass response ("to increase recording time") and boosts the treble ("to improve signal to noise ratio") during recording. That is just not correct with regard to the RIAA curve as compared with constant amplitude recording. Refer to the graph that I have included in my other responses on this subject in other threads and you will see that the bass is actually BOOSTED below 50Hz and the treble is LOWERED to a shelved response above about 2kHz, exactly the opposite of that claimed in the perpetual myth. As mentioned elsewhere, the bass was boosted in the RIAA characteristic to allow the complementary cut during playback to reduce hum and rumble. The cut in the treble was done to reduce tracing distortion with the then-common conical playback styli and to reduce power demands on the cutter head at high frequencies. The fact that the corresponding playback EQ actually INCREASES the surface noise from the disc is totally misunderstood and incorrectly argued. People who make these arguments and descriptions are misinterpreting the 6dB/octave additional slope that is required to convert constant amplitude pickups to constant velocity as having something to do with RIAA compensation. These are totally separate phenomena as you can see from my descriptions of the constant amplitude RIAA curve in the other thread on this subject.