Transfer Function of Tape?

[giro1991]

Hey all,
I'm sure these prove interesting questions.
Providing there isn't a filter in the signal path when recording to ferric oxide is there a general low pass transfer function associated with tape or is any high-cut purely the tape machine?

Judging by some graphs on google, these indicate the response to continue beyond 20khz but others there is obvious lowpass, so unsure what the actual response is. Anyone care to enlighten?

Another good question but regarding emphasis and noise reduction this time.
If you were to record an analog musical signal to tape, and enable the Dolby NR C, play it back but with de-emphasis disabled and provided the levels are slightly boosted can this be used in the sense of harmonic excitation?

I propose this idea now because I learned about psycho-acoustical masking;
When you consider lows mask highs (according to bark scale) I'm starting to think emphasis was primarily conceived for this reason and it's application for noise reduction secondary or accidental.
To re-inforces this is the fact uS time constant is often associated with high shelf or 'Air' EQ, maybe also related to temporal masking.

Most of available literature on this subject (psycho-acoustics) is way over my head and not exactly free but I know that when I tried this out, there appears to be added "presence".

 

[cliffforrest]

Dredging my memory .... Basic tape upper frequency limit is a function of the tape speed past the head gap and the wavelength of the frequency of interest.

Crudely, if the wavelength is short enough so the flux reverses before that bit of tape has gone past the head gap, it will exhibit flux cancellation.. Increasing tape speed and reducing head gap will have first order effects.

Someone will have a more precise technical reply, doubtless!

 

[sreten]

Hi,

Tape HF level and bandwidth is determined by head gap width and tape speed.

Dolby C pre-emphasis is a very poor excuse for an effective harmonic exciter.

rgds, sreten.

 

[benb]

It's complicated - many things affect tape frequency response, including the level of the bias (ultrasonic signal used to "linearize" the tape) used in recording. This search should give a lot of info:
https://www.google.com/search?q=tape+bias+and+eq+alignment

 

[giro1991]

Thanks for the replies.

I have done some comprehensive reading and discovered that harmonic distortion occurs mostly when signal is under-biased which is interesting (see attachment).

Good source: Lecture7-Magnetic_recording.pdf

Harder to find still; what about phase/group delay of tape, is the rolloff linear phase?

Dolby C pre-emphasis is a very poor excuse for an effective harmonic exciter.

Well it sure is an ok substitute if on a budget .
Best

 

[multisync]

I used to have a reel to reel, still do, its somewhere in the basement. If I remember correctly there was a switch on the R to R for different types of tape. The switch controlled the tape bias levels. Tape bias is used to overcome the hysteresis effect of magnetic tape. There was another switch for different NAB equalization settings as well. NAB is like RIAA but for tape. Ideally tape bias should be adjusted for every brand/type of tape to get the best performance out of the machine. On the data tape drives I worked on in the 70's and 80's there was no bias adjustment. Distortion was not important. There was no Dolby on the R to R. I added an outboard DBX 224 to help with the noise.

 

[giro1991]

Eh, certainly an interesting post but it didn't answer my question about group delay

 

[benb]

Thanks for the replies.

I have done some comprehensive reading and discovered that harmonic distortion occurs mostly when signal is under-biased which is interesting (see attachment).

Good source: Lecture7-Magnetic_recording.pdf

https://www.google.co.uk/url?sa=t&r...kPh3bSr9znXYZ1bXV4j3nEA&bvm=bv.96952980,d.ZGU
That's got some good stuff, except for this statement:

This is referenced to the input level that produces 3% harmonic distortion, a figure that would not be acceptable in a digital system.

Firstly, that distortion is NOT an overall figure, but only applies to the highest "acceptable" signal level - the distortion goes down substantially with a lower signal.

Also, since the tape signal response has a smooth transition to saturation (like many tube amps, and unlike the abrupt switch to flat as in high-feedback solid-state amplifiers), the distortion consists almost exclusively of lower harmonics, which are not as audible as higher harmonics at the same level, and thus not nearly as "damaging" to the sound. This is once again the fallacy of using an overall THD number to judge the expected quality of an audio signal path.

Harder to find still; what about phase/group delay of tape, is the rolloff linear phase?

Why do you want to know this? What will it tell you?

 

[jcx]

...only applies to the highest "acceptable" signal level

quite wrong - "0 dB" in analog recording practice is not "the highest" - it was common to run much more heavily into mag tape compression

mining my own post's:

some time ago (while Quantegy was still in the analog master tape business) I looked master tape performance # for a audioasylum post:

"
http://www.quantegy.com/ProductGroups/Audio/gp9.htm (I think this is “the good stuff” and it’s only 75 dB s/n re the 3% (mostly 3rd harmonic) distortion level)
to have usable S/N analog recording relies on compression/expansion (a nonlinear, ie a lossy, distorting process) like Dolby SR to “add” 10dB/24dB “S/N” low/hi frequency getting to ~ 85-100 dB ~= 16 bit digital resolution

media_2: analog tape characteristics (that’s 20 Hz on the left edge of the graph, i.e. where the analog tape fr graphs don’t get to)

the upper end of the analog tape fr curve can extend considerably beyond CD audio’s 22 KHz Nyquist limit but the bias frequency imposes a necessary low pass filtering requirement on the analog tape as well, from the 7 ½ ips curves’ 20KHz roll off we can guess that 30 ips recordings could reach 80KHz bandwidth

so the message is mixed, if you value low distortion with your S/N then 16 bit digital is competitive or ahead (noise shaped dithering can give 16 bit digital the few extra dB needed to get 100+ dB weighted s/n)

on 20 – 20 KHz “audio” frequency response flatness CD 16/44 clearly wins

if you require ultrasonic frequency extension fast analog tape can compete with digital up to 192K sample rates

if you want to even up the comparison standard by allowing the digital domain equivalents of the analog compression/expansion that is absolutely required to make analog tape usable, then look at HDCD which claims 120 dB “S/N” and is available on consumer playback devices clearly giving the specification win to consumer CD audio in the resolution category

of course no professional would call 16/44 an acceptable substitute for analog tape in the mastering process, today’s better 24/96 digital systems can actually deliver nearly 120 dB s/n with ~110 dB thd and are a practical replacement for analog tape.."

 

[giro1991]

Thanks for the pointers.

Why do you want to know this? What will it tell you?

Well I wondered if one can use tape as a substitute to linear phase filtration (a LPF with Q of 0.58 aka Bessel response) when digitising analog sound!

If it's the tape head which affects the response, how does it affect the phase response?
i.e. It's either time-continuous or not.