Better frequency response

This is where record cutters raise questions about cutting, and trade wisdom and experiment results. We love Scully, Neumann, Presto, & Rek-O-Kut lathes and Wilcox-Gay Recordios (among others). We are excited by the various modern pro and semi-pro systems, too, in production and development. We use strange, extinct disc-based dictation machines. And other stuff, too.

Moderators: piaptk, tragwag, Steve E., Aussie0zborn

Post Reply
User avatar
tape
Posts: 272
Joined: Thu Jan 25, 2007 3:52 pm

Better frequency response

Post: # 4279Unread post tape
Thu Jan 29, 2009 6:55 pm

Im using an old lathe (BELL, magnetic) from maybe the late 40'ties.
These machines has a frequency respsonse up to 8-10K at most.

If I were to build my own new cutting head, how could I improve the response? or in other words: what determines the response?

I know that it must be some combination of turntable speed while cutting, and the electronics of the cutting head.

In an earlier post Cuttercollector pointed out that, because my lathe is amercican but runs on european power frequency 50hz, it actually has a better HF response, but poorer bass response.

What combination of speed and cutting head would be needed to achive a mordern standard frequency response approx. 20-20.000 khz?

I have seen an example of a person making a cutting head, simply by gluing a needle into a speaker...maybe thats a simple way...

thanks

User avatar
cuttercollector
Posts: 431
Joined: Sun Jun 11, 2006 4:49 pm
Location: San Jose, CA

Post: # 4281Unread post cuttercollector
Fri Jan 30, 2009 3:28 am

Hmm, not sure why I would have said line frequency had anything to do with frequency response of record/playback. It would just make the machine run a little slow to put it on 50Hz without compensating by changing either the drive ratio or making the unit run on converted 60Hz.

Frequency response during cutting is obviously governed by the response of the head (major limiting factor) and the amp - more minor.
It is hard to build an amp in the last 50 years that does not respond to around at least 15KHz.
It is also governed by the media speed like analog tape.
All other things being equal, any unit should exhibit better frequency response at a higher speed.
The head in your machine probably does not go flat beyond 5Khz before it starts to roll off. You can't EQ your way back to flat because the power (even if you had enough) just does not translate through the cutter.
It will burn up rather than cut the higher frequencies at any kind of level you need.
The limits are set by internal mass of the moving parts and damping of resonances.
To get around these issues professional cutters use a servo feedback system and MUCH more power to drive the coils, which are built to handle the power. In this way you can "force" the system to flat response.
To do this in a non-feedback design very low mass must be used, but that does not lend itself to high cutting levels or the ability to handle much power. The only bright spot is we have much more efficient permanant magnets today so you get more movement for less power input AND much lower mass materials to make the moving cutter pieces out of.
A couple of the modern DJ cutters, the Vestax and one other one were non-feedback designs and got some response out above 10KHz I think.

User avatar
cuttercollector
Posts: 431
Joined: Sun Jun 11, 2006 4:49 pm
Location: San Jose, CA

Post: # 4282Unread post cuttercollector
Fri Jan 30, 2009 3:49 am

OK, I think I found what you were talking about.

I said "BTW some side effects of your 'lower speed mastering' are, you get better high frequency response from your cutter because the actual top frequency it has to cut is lower so if it were true half speed (22.5 RPM) your cutter's 8Khz frequency limit would be extended to 16Khz."

I was suggesting that if you cut at a lower speed AND drop the pitch of the audio input material by an amount such that the finished disc plays at a proper speed on a normal playback turntable running at standard speed, you get a net increase in high frequency response because the actual frequency the cutter is cutting is lower.
For instance, if you cutter was only flat to 10KHz and you cut 20KHz material at half speed, the cutter would only have to cut 10KHz and the finished product would be flat to 20KHz at normal playback speed.
But it would have to cut 10Hz ! on the bottom to have response to 20Hz when played back.
Not a normal situation anyway. The best analog recording devices are probably only reasonably flat from 40-15K with perhaps a bit of response beyond that, but not flat. Tape being better thn disc.

User avatar
markrob
Posts: 1623
Joined: Wed Nov 28, 2007 1:14 am
Location: Philadelphia Area

Post: # 4283Unread post markrob
Fri Jan 30, 2009 10:51 am

One other advantage to half speed mastering is that you can use less power for the same playback level since velocity is proportional to frequency. So, if you can cut at 5 cm/s at 8Khz. It will play back as 10 cm/s when you speed back up.

User avatar
tape
Posts: 272
Joined: Thu Jan 25, 2007 3:52 pm

Post: # 4416Unread post tape
Mon Feb 09, 2009 11:48 am

cuttercollector wrote:OK, I think I found what you were talking about.

I said "BTW some side effects of your 'lower speed mastering' are, you get better high frequency response from your cutter because the actual top frequency it has to cut is lower so if it were true half speed (22.5 RPM) your cutter's 8Khz frequency limit would be extended to 16Khz."

I was suggesting that if you cut at a lower speed AND drop the pitch of the audio input material by an amount such that the finished disc plays at a proper speed on a normal playback turntable running at standard speed, you get a net increase in high frequency response because the actual frequency the cutter is cutting is lower.
.
thank for the answer...

Im still a bit puzzled about this.

As you say, the faster the disc is rotating, the better response (like higher samplerate in digital audio) but how is that conherent with an increased HF response when cutting at half speed?

there seems to be a contradiction here (or, more likely: maybe Im getting something wrong)

User avatar
cuttercollector
Posts: 431
Joined: Sun Jun 11, 2006 4:49 pm
Location: San Jose, CA

Post: # 4419Unread post cuttercollector
Mon Feb 09, 2009 12:18 pm

The 2 facts are in opposition to each other.
Say your cutter will cut to 8kHz at 78 RPM but only 5kHz at 33RPM.
That would be kind of typical.
Now, if you cut at half speed for playback and drop the source material to half speed, the cuttter's 5kHz response becomes 10kHz on playback.
I know you would never cut a record at 33 and play back at 66 but you get the idea.
The effect is probably not linear though. I don't know if the response would fall off a cliff if you tried to cut at half 33 (16 2/3) and play back at 33, perhaps the offset would not help you as much as if you cut at 39 and played back at 78.
In the first case, say the cutter only could cut to 3kHz at 16 2/3, you would only get 6kHz on playback.
In the 2nd case if it could cut to 7kHz at 39 RPM, you could get 14kHz on 78 playback.
There are other factors too, like how stable the turntable is cutting at low speeds. If there is more speed variation,(wow and flutter) it won't go away, it will just be doubled in frequency also playing back at double (normal) speed. And the bass response gets cut off higher as well. If the cutter only will go down to 50Hz, the the bottom cutoff through half speed mastering becomes 100Hz and there goes all your bass. It sometimes is also harder to cut a quiet groove at lower speeds, but as was pointed out the peak velocity of the cutter remains the same but the effective velocity goes up because you end up playing it back faster than it was cut.

Just keep these 2 factors in mind.

1. the higher the cutting and playback speed, generally the better the frequency response, signal to noise ratio etc.

2. when you cut at a lower speed (say half) the disadvantages of cutting at that slower speed are there, but because the actual playback speed is doubled it gives you a "free" net increase in performance because the cutter does not have to cut the actual frequency or level, but only half as much. It is working "half as hard" for the same result or as hard as possible but the results are almost (see #1) twice as good at when you play it back at normal intended speed because all frequencies are doubled and effective level goes up too.

User avatar
markrob
Posts: 1623
Joined: Wed Nov 28, 2007 1:14 am
Location: Philadelphia Area

Post: # 4422Unread post markrob
Mon Feb 09, 2009 1:16 pm

One thing to keep in mind is that the cutter perfomance at lower platter speeds is not reduced. So if your head has response out to 8Khz. You will be able to cut 8 Khz at half speed. This yeilds a high frequency response to 16 Khz when played back at full speed. The playback stylus shape limits the high frequency response at lower linear velocities. In the case of cutting, the stylus is a very sharp knife edged tool and can cut very small details. In the days of 78 RPM, the playback stylus was 3 mil. conical shape. At the inside diameter of a disc, the high freq's would fall off significantly on playback. The highs were still cut into the record, but since the linear speed is reduced (due to small diameter), the high frequency details become too small for pickup stylus to trace. In the case of microgroove LP's and 45's the stylus tip is 1 mil and you can show that the high freq limit is much better at the inner diameters compared with 78's under the same conditions even though the linear velocity is higher.

Half speed mastering allowed for the production of CD-4 LP's in the 70's. They needed to have reposonse beyond 40 Khz for the FM carriers. I don't think any cutter available can work at those frequencies. I've read a paper that indicates the early CD-4's were cut with a Neuman SX-68 head using its response to only 15 Khz and mastering at 1:2.7 (12.2 RPM !). This allowed them to cut to 40 Khz. Also note that to play these disks back requires a better stylus shape with a small contact area. Hope this makes some sense.

Mark

User avatar
cuttercollector
Posts: 431
Joined: Sun Jun 11, 2006 4:49 pm
Location: San Jose, CA

Post: # 4425Unread post cuttercollector
Mon Feb 09, 2009 1:58 pm

Hmm, interesting point regarding cutter response at slower speeds.
I guess I stand corrected.
Not to hijack this thread, but then are you saying that given the same cutter with a response to say, 15kHz and modern microgroove stylus shapes for playback there is no frequency response or dynamic range (signal to noise) advantage to cutting and playback at 78 RPM vs. 33 1/3? The linear velocity advantage would still be there if you could play it back, which brings us to levels cut.
Again, leaving half speed mastering out of the picture for a moment, cutting and playing at actual normal speed, what are the net advantages and disadvantages of higher media speed? If as you say the cutter cuts to 15kHz no matter what, and the maximum level it cuts is the same in terms of actual groove mechanical deviation, is the only advantage that the high frequency details are longer and thus easier to trace on playback? I know the actual quiet groove noise goes up with speed but does actual effective level go up more?
The interrelationships of all these factors can be confusing. (just started a new topic to discuss all this)

Back to the half speed mastering with a home cutter from the 40s, another issue might be that any turntable rumble issues that get cut into the disc will also be doubled in pitch when played back at normal speed.

Whether there is any theoretical advantage to cutting at higher speeds in general or not, think of it this way, your cutter response say 40-8,000 on a good day with the wind at it's back, is the "window" superimposed on the program material - a "filter" if you will. If your program material is 20-20,000 and you half speed master that means that all frequecies are halved. This would mean you are cutting program material that now is shifted in pitch to 10-10,000 with a cutter still cutting the range of 40-8,000 into the disc turning at half speed. Now when you play it back at normal speed the window has shifted and the playback frequency range is 80-16,000 because all pitches are returned to normal and the record is spinning twice as fast as when it was cut.

User avatar
markrob
Posts: 1623
Joined: Wed Nov 28, 2007 1:14 am
Location: Philadelphia Area

Post: # 4427Unread post markrob
Mon Feb 09, 2009 3:37 pm

Hi cc,

I don't believe there will be any downside to dynamic range, but I'm certainly no expert on this. I scanned a couple of pages from F.V. Hunt's 1962 AES paper, "The Rational Design of Phonograph Pickups" that deals with scanning loss.

http://home.comcast.net/~markrob1066/pwpimages/ScanLoss1.jpg

http://home.comcast.net/~markrob1066/pwpimages/ScanLoss2.jpg

Hope this makes some sense.

The area of concern, as you point out are the low frequencies. You need to be sure that the signal chain driving the head has a low freq response able to handle the slowed down audio. I believe the head will have response down to DC. I tried a test with my Presto 1C. I used a dial indicator and a DC source to the coil to see if the stylus would deflect statically. It did. I think a voice coil style head should as well. In this region, the head is in constant amplitude mode. In the case of a Presto head, you would have to change the series resistor to set the L/R time constant for 250 Hz. instead of 500 Hz. The high freq turnover would need to be 150 us or 1061 Hz. as well.

The rumble issue you bring up is also interesting. I'm going to make a wild guess that the rumble spectral components are related to the platter rotation. If so, then you would expect the components to shift downward in relation to the platter speed. In this case there would be no difference. If they are not related to the platter speed, then you would be correct. It would be interesting to test this.

User avatar
cd4cutter
Posts: 67
Joined: Tue Jul 08, 2008 3:01 pm

Post: # 4430Unread post cd4cutter
Mon Feb 09, 2009 6:57 pm

Mark, your links to the technical paper on scanning loss are interesting. This is indicative of the thinking going on in the late 1950s and early '60s when mono records were still popular. People were more concerned with high frequency LEVEL loss than distortion at that time. The term "scanning loss" was taken to mean a loss of HF frequency response, typically occuring at the inner diameters because of the reduced linear groove speed and the smaller wavelengths as compared with those at the outer diameters. Scanning loss was a result of playback stylus shape AND tracking force which caused the deformation of the small geometries of the groove wall. You will note that much of the discussion in this paper addresses the groove wall deformation issue and there is little mention of distortion. Notice also that tracking forces of 5 to 10 grams was still considered normal at this time. Note that there is some mention in this paper of the practice of compensating for scanning loss during the recording process. This technique was called using "diameter compensation", and it entailed the dialing in of increasing amounts of HF boost as the recorded diameter became smaller. Some recording lathes of the period had pots or switches included in them to be used to switch in this diameter compensation as the cutter traversed to smaller diameters.

All that emphasis changed just about when this paper was published with the introduction of stereo records in 1958. As stereo records became the accepted norm, the problem of TRACING DISTORTION became the dominant issue as it was of far more importance than mere scanning loss. Mono lateral records did not suffer from many of the problems of tracing distortion. Yes, the tracing distortion mechanism was in play - it was called "the pinch effect" in those days. Meaning that the improper fit of the stylus in the groove caused the stylus to be "pinched" upward when the groove wavelengths were very small and/or the recorded level was very high. But this had little effect in the generation of distortion products in the lateral motion of the stylus. The net effect of the pinching was to cause scanning loss - merely a loss of HF response. But the pinching caused some vertical motion of the stylus which was 100 PER CENT DISTORTION! None of this vertical motion was intentional or desirable in the case of mono lateral cut records. This motion still existed in the playback of stereo records. But now, this vertical motion was actually appearing in BOTH channels of the cartridge output signals. Suddenly, listeners were hearing gobs of distortion when playing stereo records (or when playing mono records with a stereo pickup) that they had not been accustomed to hearing when playing mono laterals with a lateral pickup. There was A LOT of theorizing, discussion, and general hand-wringing in the hifi world that this terrible sound would be the death of stereo before it ever got off the ground.

All of which led to the rapid development of better playback cartridges. The first thing they did was to reduce the playback stylus from a 1.0 mil conical to a 0.7 mil conical shape. Some hifi cartridges also offered 0.6 and even as small as 0.4 mil conical styli (ADC was the predominant supplier of these). This considerably reduced the tracing distortion. Then the re-introduction of elliptical or "bi-radial" playback styli AND the availability of cartridge designs that allowed MUCH lower tracking forces - down to the 1 to 2 gram range was crucial to the furtherance of the stereo record. I say "re-introduction" because the elliptical stylus was originally invented by none other than Thomas Edison around 1904. Then as now, Edison recognized the importance of reducing the scanning radius to improve high frequency reproduction of his wax cylinder records. But these more expensive styli were not adopted by most of the makers of lateral-cut records and their phonographs and they fell out of general use for decades. The adoption of Edison's elliptical stylus shape and even further improvements into line contact styli was the key to making stereo records listenable for the more critical listeners.
Collecting moss, phonos, and radios in the mountains of WNC

User avatar
markrob
Posts: 1623
Joined: Wed Nov 28, 2007 1:14 am
Location: Philadelphia Area

Post: # 4432Unread post markrob
Mon Feb 09, 2009 7:38 pm

cd4cutter wrote: All of which led to the rapid development of better playback cartridges. The first thing they did was to reduce the playback stylus from a 1.0 mil conical to a 0.7 mil conical shape. Some hifi cartridges also offered 0.6 and even as small as 0.4 mil conical styli (ADC was the predominant supplier of these). This considerably reduced the tracing distortion. Then the re-introduction of elliptical or "bi-radial" playback styli AND the availability of cartridge designs that allowed MUCH lower tracking forces - down to the 1 to 2 gram range was crucial to the furtherance of the stereo record. I say "re-introduction" because the elliptical stylus was originally invented by none other than Thomas Edison around 1904. Then as now, Edison recognized the importance of reducing the scanning radius to improve high frequency reproduction of his wax cylinder records. But these more expensive styli were not adopted by most of the makers of lateral-cut records and their phonographs and they fell out of general use for decades. The adoption of Edison's elliptical stylus shape and even further improvements into line contact styli was the key to making stereo records listenable for the more critical listeners.
Hi CD4,

Thanks for the comments. I'm just home experimenter, but you've obviously been around this professionally for some time.

In the same book I own with the collections of papers on disk reproduction (Edited by H.E. Roys) , I there is an 1963 AES paper by Fox and Woodward of RCA that describes a system for correcting tracing distortion of stereo recordings. It uses a taped delay line to pre-distort the cutting signal so that the playback has reduced distortion. I gather that this was a component of what was called Dynagroove by RCA. Was this still in use into the 70's? Or, did better playback stylii render it obsolete?

It was also interesing that the same results can be acheived by recording onto a master disk and re-recording with inverted polarity to a 2nd generation disc. In essence cancelling out the non-linearity in the playback process.

BTW, if you can find a copy of "Disc Recording and Reproduction" by H.E. Roys, its a great resource. It is a collection of most of the major landmark technical papers on disc recording spanning 1926 to 1971. I'm just amazed at the effort that went into engineering recording systems.

Mark

User avatar
cd4cutter
Posts: 67
Joined: Tue Jul 08, 2008 3:01 pm

Post: # 4435Unread post cd4cutter
Tue Feb 10, 2009 4:10 pm

Ed Roys was an engineering manager at the Indianapolis record plant of RCA Records. He retired a little before I went to work there, so I never got to meet him. Yes, the "pre-distortion" technique written about in the Woodward paper was employed in the original RCA Dynagroove records. In fact, it was the most significant aspect of the process. The other main change was in the physical profile of the records which came in later production. The records were made thinner in the music area. This profile reduced the weight of the records from about 116 grams to about 98 grams. The new profile was called "Dynaflex" and was cursed by most record buyers, but it did actually allow the records to lie flatter on a full 12 inch turntable and exhibit better damping because of the closer contact with the turntable platter over a wider area of the record.

Anyway, the pre-distortion device was called the "Dynamic Recording Correlator" (DRC), and it was really an analog computer capable of dynamically (yes) adjusting the distortion of the waveform with increasing signal frequency and level. And, yes, it did this with the use of a tapped analog delay line and very rapid signal switching via critically biased transistors set up at multiple taps along the delay line. You youngsters don't remember or can't imagine that digital computers were almost nonexistent in the early 1960s, and few of them even then had the computing power to create these waveforms operating in real time. The DRC system actually worked pretty well, but it was designed to accommodate the tracing distortion inherent in the use of 0.7 mil conical styli which were the most common type used in the early 1960s for stereo records. It was difficult to recalibrate the thing for other stylus sizes, so nobody ever did. The system was difficult to keep in alignment, so the mastering studios gradually quit using it. We had four RCA company mastering studios in operation at this time: New York, Chicago, Los Angeles, and Nashville. All of them had the DRC available, but it was phased out of use in all of them eventually. The independent mastering studio was not yet a significant player in the record business, and all the major labels including Columbia and Capitol had their own company-owned mastering studios. None of these other studios adopted the DRC. Don't know why, but it was probably because they didn't want to pay for it.

Yes, the introduction of the elliptical and later stylus shapes rendered the effectiveness of the DRC moot. But it's interesting to note (almost nobody knows this) that all three generations of the JVC CD-4 cutting systems employed their own version of the DRC which was optimized for the Shibata stylus shape! They called this circuit "Neutrex I". They also included a second pre-distorter called "Neutrex II" which addressed the effect of intermodulation distortion (IM) caused by the tracing distortion of the baseband signal on the carrier signal. This circuit used a portion of the baseband signal, properly filtered and phased, to add to the modulation of the carrier which pretty effectively cancelled out the IM from the tracing distortion. CD-4 was a very sophisticated system and more complicated in detail than casual observers realize. Most people don't know the extent to which analog engineering was stretched to make this system work. CD-4 was the ultimate analog stylus-in-groove technology until the CED videodisc system was developed.

As a consequence of the use of Neutrex I, you probably will hear the BEST, lowest distortion analog disc reproduction that it is possible to hear when you play a CD-4 record with a Shibata stylus and listen to only the baseband (normal stereo) reproduction. It's so clean, it almost sounds like digital. :D
Collecting moss, phonos, and radios in the mountains of WNC

User avatar
doneth
Posts: 1
Joined: Sun Aug 27, 2017 9:17 am

Re: Better frequency response

Post: # 47659Unread post doneth
Sun Aug 27, 2017 3:22 pm

I was working in a record shop at the time and I can tell you that the reason the super-skinny Dynaflex records failed was that too many of them were so badly warped that they could not be played. In the store we theorized that they were being ripped out of the waffle iron before they had cooled completely -- don't know if that makes any sense or not, but we got more exchanges with those than any other product and the new set was usually just as bad.

User avatar
fredbissnette
Posts: 383
Joined: Tue Oct 11, 2016 5:38 pm

Re: Better frequency response

Post: # 47721Unread post fredbissnette
Sun Sep 03, 2017 12:14 am

https://www.youtube.com/watch?v=wYssxBqeutI&feature=youtu.be


i cut this with a tungsten needle and a homebuilt lathe

it is possible to get good results
Instagram @styluspressurerecords

Post Reply