Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Analog tape noob: explanation needed (Read 13145 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Analog tape noob: explanation needed

I am an audio restoration technician who already digitized over 1000 open reel tapes, the studio has Otari MTR-15 and Studer A810, and some other Studer and Revox machines. However, the maintenance of machines are done by (old) technicians from another company and I never know their "secrets". I searched terms like bias, MRL tapes, ips but I am not sure that what I read are correct or not. My questions are listed below:

1. Bias
Most websites/wikipedia etc says bias is a high frequency signal (over 100khz?) added in recording to improve sound quality. Then I immediately thought about noise-shaping in digital audio. However, noise-shaping is spreading noise in high frequency range (15khz-22khz in 44.1khz sampling rate, for example) but bias (from what I understand) is like a single high amplitude sine wave. Will the bias signal cause audible intermodulation distortion? Will the tape machine filter high frequency in playback mode? If I turn the reel very very slowly, can I hear the bias signal? (actually I've tried but I can hear nothing). What I really like to see is the FFT analysis of tape signal, however the recording interface in the studio only supports up to 96khz so it may not be able to record the bias signal.

2. Tape speed vs playback amplitude
We are required to set the playback level to -4dBVU when the tape is playing the test tone before the audio starts. Some tapes are recorded in 1.875 ips but the machines only support as low as 3.75 ips. I suggested to record at 96khz (we normally use 48khz) in 3.75 ips then change the playback rate to 48khz, but my supervisor said If I want to do this, then I need to set the playback level to -8dBVU because the slower the speed, the higher the amplitude. Is this correct? At least in digital audio, playback rate will not affect amplitude, it is quite strange to me.

3. CCIR vs NAB
CCIR sounds brighter than NAB. Their differences are similar to playback EQ in cassette players. However, only type 2 (CrO2) and type 4 (metal) tapes should use the darker playback EQ while type 1 tapes should use the brighter EQ. Is this true for open reel tapes as well? I mean, some (cheap) tapes should record and play using CCIR, some (expensive) tapes should use NAB? Or all tapes can use either CCIR or NAB?

Thank you very much!!


Analog tape noob: explanation needed

Reply #1
1. Bias

Necessary to overcame the hysteresis problem inherent in recording to magnetic media. Essentially the magnetizing process is non-linear at the "zero crossing" of the baseband signal, so it is modulated onto a bias current which keeps the it in the linear part of the response curve.

Sheesh, reading my own explanation does my head in. Try and find a (physical) copy of an audio engineering textbook from the 60's to the 80's to get a much better explanation than mine.


2. Tape speed vs playback amplitude

Yes, this is inherent in the response of magnetic tape playback. Open Reel machines have to be set to the correct response curve; usually CCIR or NAB. If you play back, say,  a CCIR recording at half (or twice) the speed that it was recorded at, then of course you are in a different of the response curve. Make sense? Once again, an old textbook will explain this better than I can.

3. CCIR vs NAB

Basically CCIR = Euro. NAB = US.

Match your machine's setting to whatever the tape was recorded on.

 

Analog tape noob: explanation needed

Reply #3
Thank you very much!

Analog tape noob: explanation needed

Reply #4
...(old) technicians from another company and I never know their "secrets".
Luckily some of the (old) experts don't mind sharing their knowledge, like Richard Hess.
The MRL (calibration tape) website also has lots of interesting free documents.

Quote
Bias
Most bias frequencies are around 150 kHz. There's a list of several recorders and their bias freq. on the Hess website. The bias tone will be filtered out in the recorder's playback amp. AFAIK you should be able to hear the signal when you play (reel-rock) the tape very slowly.
BTW bias is only relevant for recording, not for playback.

Quote
2. Tape speed vs playback amplitude
1.875 ips is very slow, so it's probably a recording where audio quality wasn't very important. If you can more or less correct the PB-EQ differences your solution might work pretty well.
At low tape speeds the record level is usually set lower to avoid saturation of the high frequencies. "the slower the speed, the higher the amplitude" doesn't make sense. When you push the tape to the PB head during fast forward or rewind, the output usually overloads, so I'd say faster is louder.

Quote
3. CCIR vs NAB
All tapes can use either IEC (CCIR) or NAB, but like bbl wrote it's important to use the same for REC and PB and write on the tape box which one was used. Having both 1kHz and 10kHz tones on the tape can help a lot.


Analog tape noob: explanation needed

Reply #5
Luckily some of the (old) experts don't mind sharing their knowledge, like Richard Hess.
The MRL (calibration tape) website also has lots of interesting free documents.

Thank you for pointing me to these websites.

Most bias frequencies are around 150 kHz. There's a list of several recorders and their bias freq. on the Hess website. The bias tone will be filtered out in the recorder's playback amp. AFAIK you should be able to hear the signal when you play (reel-rock) the tape very slowly.
BTW bias is only relevant for recording, not for playback.

Maybe I was not slow enough, I will try again next time.

1.875 ips is very slow, so it's probably a recording where audio quality wasn't very important. If you can more or less correct the PB-EQ differences your solution might work pretty well.
At low tape speeds the record level is usually set lower to avoid saturation of the high frequencies. "the slower the speed, the higher the amplitude" doesn't make sense. When you push the tape to the PB head during fast forward or rewind, the output usually overloads, so I'd say faster is louder.

I will try to digitize a 7.5 ips tape using different speeds and examine their loudness and EQ next time.

All tapes can use either IEC (CCIR) or NAB, but like bbl wrote it's important to use the same for REC and PB and write on the tape box which one was used. Having both 1kHz and 10kHz tones on the tape can help a lot.

One of my supervisors said some tapes like Scotch 806 should use NAB and others like Ampex 631 should use CCIR. I don't think it is reasonable because sometimes sound quality is strange if I follow this rule. There is no instruction written on the tape boxes as well. As a result we can only use our ears to determine what type to use.

Analog tape noob: explanation needed

Reply #6
I am an audio restoration technician who already digitized over 1000 open reel tapes, the studio has Otari MTR-15 and Studer A810, and some other Studer and Revox machines. However, the maintenance of machines are done by (old) technicians from another company and I never know their "secrets". I searched terms like bias, MRL tapes, ips but I am not sure that what I read are correct or not. My questions are listed below:

1. Bias
Most websites/wikipedia etc says bias is a high frequency signal (over 100khz?) added in recording to improve sound quality. Then I immediately thought about noise-shaping in digital audio. However, noise-shaping is spreading noise in high frequency range (15khz-22khz in 44.1khz sampling rate, for example) but bias (from what I understand) is like a single high amplitude sine wave.


Bias frequencies used with analog tape range from a few multiples of the highest frequency recorded to several MHz.

Quote
Will the bias signal cause audible intermodulation distortion?


No, bias greatly reduces nonlinear distortion.

A short course about analog audio recording

Quote
Will the tape machine filter high frequency in playback mode?


Yes.

Quote
If I turn the reel very very slowly, can I hear the bias signal?


Maybe, maybe not. High frequency response is about physical wavelengths and most bias frequencies are above the highest frequency (shortest physical wave) that a tape head will respond to.

These days people have a process for compensating for flutter and wow in analog tape that is based on recovering the bias frequency, but special playback heads are needed.

Quote
(actually I've tried but I can hear nothing). What I really like to see is the FFT analysis of tape signal, however the recording interface in the studio only supports up to 96khz so it may not be able to record the bias signal.


Many older analog tape machines used bias frequencies below 96 Khz.  But most analog tape heads do not respond well that high. Not at all!

Quote
2. Tape speed vs playback amplitude
We are required to set the playback level to -4dBVU when the tape is playing the test tone before the audio starts. Some tapes are recorded in 1.875 ips but the machines only support as low as 3.75 ips. I suggested to record at 96khz (we normally use 48khz) in 3.75 ips then change the playback rate to 48khz, but my supervisor said If I want to do this, then I need to set the playback level to -8dBVU because the slower the speed, the higher the amplitude. Is this correct? At least in digital audio, playback rate will not affect amplitude, it is quite strange to me.


Check the reference I provided for this equation:

The equation approximating the output voltage [e(x)] from a magnetic recorder reproduce head is:

e(x) = -µ0VwM0 (Hgg/i) k? [e-kd] [(1-e-k?)/k?] [sin(kg/2)/(kg/2)] cos(kx)

V is the physical speed of the tape, so a faster moving tape generates a higher output voltage all other things being equal.

Quote
3. CCIR vs NAB
CCIR sounds brighter than NAB. Their differences are similar to playback EQ in cassette players. However, only type 2 (CrO2) and type 4 (metal) tapes should use the darker playback EQ while type 1 tapes should use the brighter EQ. Is this true for open reel tapes as well? I mean, some (cheap) tapes should record and play using CCIR, some (expensive) tapes should use NAB? Or all tapes can use either CCIR or NAB?


That issue is described here:

NAB versus CCIR equalization



Analog tape noob: explanation needed

Reply #7
One of my supervisors said some tapes like Scotch 806 should use NAB and others like Ampex 631 should use CCIR.



Too bad.  The most important thing about comparing and using recording tape is to bias them properly. That alone goes a long way towards making them perform well.

The purpose of CCIR and NAB is to maintain compatibility with other recorders, not match different recording tape.

You should try to find the manufacturer's instructions for biasing and doing other adjustements on one of the recorders you have on hand.

Just reading it would probably be a revelation to not only you, but your supervisor it would seem!

Quote
I don't think it is reasonable because sometimes sound quality is strange if I follow this rule. There is no instruction written on the tape boxes as well. As a result we can only use our ears to determine what type to use.


Not at all. Find the manuals for setting up your tape machines - they generally give generalized instructions for setting up your analog recorders that should give good results with just about any kind of tape at any speed.

Analog tape noob: explanation needed

Reply #8
Thanks Arnold! Especially the CCIR and NAB response graph. With this I can approximate a better compensation EQ when digitizing in different tape speeds. Also thanks for providing other documentations as well.

Analog tape noob: explanation needed

Reply #9
Thanks Arnold! Especially the CCIR and NAB response graph. With this I can approximate a better compensation EQ when digitizing in different tape speeds. Also thanks for providing other documentations as well.
The Studer A810 in your studio should be able to switch between CCIR and NAB. Wouldn't that be the easiest solution ?

Analog tape noob: explanation needed

Reply #10
Thanks Arnold! Especially the CCIR and NAB response graph. With this I can approximate a better compensation EQ when digitizing in different tape speeds. Also thanks for providing other documentations as well.
The Studer A810 in your studio should be able to switch between CCIR and NAB. Wouldn't that be the easiest solution ?


I mean when I need to digitize a 1.875 ips tape in 3.75 ips speed. However after I read the pdf I found that the EQ curves are for 15 ips.

I don't have any 1.875 ips tapes in the studio today so I used a 7.5 ips tape and digitized it at 3.75 ips, 7.5 ips and 15 ips at 96khz, then resampled and changed their playback rate and examined their differences. Assuming the resampler is good enough (Adobe Audition 1.5), the EQ differences between 7.5 and 15 ips are not so big, though audible (13/15 after ABX) but their EQ differences with the 3.75 ips recording are very obvious. For loudness of a specific part (around 1 minute), I measured their average RMS power. The 3.75 ips recording is around -27.1dB, the 7.5 ips one is around -26.7dB, the 15 ips one is around -26.2dB. I did not make a RMAA tape to check the response since I can't do anything I like in the studio 

The tape machine is Otari MTR-15 in CCIR mode. Other machines are occupied by others so I could not check them.

Analog tape noob: explanation needed

Reply #11
Thanks Arnold! Especially the CCIR and NAB response graph. With this I can approximate a better compensation EQ when digitizing in different tape speeds. Also thanks for providing other documentations as well.
The Studer A810 in your studio should be able to switch between CCIR and NAB. Wouldn't that be the easiest solution ?


I mean when I need to digitize a 1.875 ips tape in 3.75 ips speed. However after I read the pdf I found that the EQ curves are for 15 ips.

I don't have any 1.875 ips tapes in the studio today so I used a 7.5 ips tape and digitized it at 3.75 ips, 7.5 ips and 15 ips at 96khz, then resampled and changed their playback rate and examined their differences. Assuming the resampler is good enough (Adobe Audition 1.5), the EQ differences between 7.5 and 15 ips are not so big, though audible (13/15 after ABX) but their EQ differences with the 3.75 ips recording are very obvious. For loudness of a specific part (around 1 minute), I measured their average RMS power. The 3.75 ips recording is around -27.1dB, the 7.5 ips one is around -26.7dB, the 15 ips one is around -26.2dB. I did not make a RMAA tape to check the response since I can't do anything I like in the studio 

The tape machine is Otari MTR-15 in CCIR mode. Other machines are occupied by others so I could not check them.


The best way to handle situations where you have to play a tape under nonstandard conditions is to digitize a test tape that you either have or make, and then figure out what it takes in the way of equalization to make the transcription match the test tape.

Analog tape noob: explanation needed

Reply #12
Thanks again Arnold, I am now using your method.

Analog tape noob: explanation needed

Reply #13
Quote
Will the bias signal cause audible intermodulation distortion?


No, bias greatly reduces nonlinear distortion.
You can very easily get audible intermodulation distortion by trying to record a high frequency tone with some mathematical relationship to the bias frequency that places the distortion product within the audible band.

Have you never tried recording a frequency sweep to analogue tape? A typical 20Hz-20kHz sweep will expose this problem on many tape recorders. The high teens of kilohertz are enough to upset most cassette decks.


No reason to worry about this when playing back existing tapes - you have what you have, and it's unlikely to be a recording of high frequency sine waves.

Cheers,
David.

Analog tape noob: explanation needed

Reply #14
Quote
Will the bias signal cause audible intermodulation distortion?


No, bias greatly reduces nonlinear distortion.
You can very easily get audible intermodulation distortion by trying to record a high frequency tone with some mathematical relationship to the bias frequency that places the distortion product within the audible band.


If you stick to answering the OPs question instead of making up your own up on the fly, the only reasonable comparison is to performance with no bias.

When you are talking about a system with from a few percent to over 10% nonlinear distortion, all sorts of crazy thing can happen. However, the question asked (as opposed to the question newly imagined) was whether or not bias increases or decreases nonlinear distortion.  Got a relevant answer to the question that was actually asked? ;-)

Quote
Have you never tried recording a frequency sweep to analogue tape?


What do you think? All things considered, that question could be taken in a very negative way. The first tape machine I ever measured was a Wollensak mono unit with tubes. If memory serves, the tape used was SOTA for the day - Scotch 111.

Quote
A typical 20Hz-20kHz sweep will expose this problem on many tape recorders. The high teens of kilohertz are enough to upset most cassette decks.


Feel free to post your results of a reasonable test showing this defect. ;-)

I don't remember too much about the last time I did any tests of a cassette machine, but I do remember obtaining results that looked pretty consistent with reasonable expectations for cassette, but ghastly by modern standards. It was recent enough that I did the tests using the Audio Rghtmark program.

In some ways the audio cassette made the LP format look pretty good in comparison.



Analog tape noob: explanation needed

Reply #15
The first tape machine I ever measured was a Wollensak mono unit with tubes. If memory serves, the tape used was SOTA for the day - Scotch 111.

I too first started recording with a Wollensak with tubes, probably the same model, and Scotch 111. That was actually a very well built machine.

Analog tape noob: explanation needed

Reply #16
If you stick to answering the OPs question
I was making no pretence to usefully answer the OP's question - I was just pointing out an interesting but largely useless fact.

Quote
Feel free to post your results of a reasonable test showing this defect. ;-)
If only I hadn't dumped my fascinating collection of tone sweeps recorded to tape and then re-digitised. (sadly, I'm not joking!). I can see a tape deck right now, but wouldn't know where to find a blank tape. The teenage me would never have believed this to be possible

Cheers,
David.
(not nearly as old as you, but contemplating middle age / pipe-and-slippers anyway)

Analog tape noob: explanation needed

Reply #17
Too bad all of my cassette recorders were broken and discarded several years ago, otherwise I can easily make a sweep tape and check it at home. I still have some new TDK tapes at home (the cheapest D60 model)

Analog tape noob: explanation needed

Reply #18
People who are hearing audible beats with the bias oscillator are probably doing sweeps at 0dB.
As far as I can remember, it is not the hf audio that causes the beat but the distortion products that result from the pre-emphasis boosting the hf into overload.
Pre-emphasis ('equalization') relies on the fact that the hf component of real signals is at a low level allowing it to be boosted during record and then reduced on playback to improve the sn ratio.

Analog tape noob: explanation needed

Reply #19
Cassette decks often had a "beat switch" that changed the bias oscillator frequency if you had problems trying to record FM radio, where the 19 KHz pilot was often present at significant levels in the tuner output. You can bet that the manufacturers wouldn't go to the expense of adding the switch if it weren't a common problem. It was apparently a cheaper solution than trying to filter out the tone in the low end systems (radio/cassette boomboxes), though 19 KHz LC filters were quite common in more expensive decks - switchable on some models capable of HF response to over 20 KHz.
Regards,
   Don Hills
"People hear what they see." - Doris Day

Analog tape noob: explanation needed

Reply #20
I was only five (I'm now in my thirties), but I'm fairly sure the "beat" switch on my music centre (!) affected AM recording (only?).


You might be right about pre-emphasis. I've never seen the pre-emphasis curve for a cassette deck - though, thinking about it, maybe some Sony tapes came with it printed on the wrapper? Especially for Chrome tapes? Can't really remember - certainly can't remember what it looked like.


Analog tape noob: explanation needed

Reply #21
Cassette decks often had a "beat switch" that changed the bias oscillator frequency if you had problems trying to record FM radio, where the 19 KHz pilot was often present at significant levels in the tuner output. You can bet that the manufacturers wouldn't go to the expense of adding the switch if it weren't a common problem. It was apparently a cheaper solution than trying to filter out the tone in the low end systems (radio/cassette boomboxes), though 19 KHz LC filters were quite common in more expensive decks - switchable on some models capable of HF response to over 20 KHz.



I've never ever seen this in my life. There's plenty of tech information about legacy cassette decks online, so please provide evidence that even 10% (a very weak example of "often", but I'll be generous) of them had this alleged feature. I couldn't even find a relevant reference with Google...

Analog tape noob: explanation needed

Reply #22
People who are hearing audible beats with the bias oscillator are probably doing sweeps at 0dB.


That makes sense. The preponderance of cassette machines with 19 KHz filters had 19 KHz filters to keep the HF tone from affecting Dolby tracking during recording.

I suspect that a typical music signal was bouncing around to much for the ear to actually lock onto some spurious response except under some pretty rare circumstances. 

Since the filters were simple LC networks, there might be an audible effect as they were switched in and out. I never did a DBT but I was under the impression that I could detect them on most of the equipment of the day that had  them.

As I said before, when you have a medium with from 1 to 10% nonlinear distortion, all sorts of crazy things can happen. 

Cassette tape was like the MP3 format of the day in the sense that when you measured it you wondered how so many bad things could be happening and still sound reasonably acceptable. For example the standard was to test FR at -20 dB  because over most of the life of the medium at -10 dB very little actually worked above 5 KHz. 

FM Stereo Tech Details On Wikipedia  puts the amplitude of the 19 KHz  tone at 10% of FS.

In combination with the high nonlinearity of the medium (on the order of 3% at 0 dB  the FM stereo pilot tone could easily create spurious tones that were only 40-50 dB down, which would be above the noise floor, as ridiculously high as it was.  Consider that recording above 0 dB often happened because of the poor transient response of the electrical meters that were used.


Analog tape noob: explanation needed

Reply #23
Cassette decks often had a "beat switch" that changed the bias oscillator frequency if you had problems trying to record FM radio, where the 19 KHz pilot was often present at significant levels in the tuner output. You can bet that the manufacturers wouldn't go to the expense of adding the switch if it weren't a common problem. It was apparently a cheaper solution than trying to filter out the tone in the low end systems (radio/cassette boomboxes), though 19 KHz LC filters were quite common in more expensive decks - switchable on some models capable of HF response to over 20 KHz.



I've never ever seen this in my life. There's plenty of tech information about legacy cassette decks online, so please provide evidence that even 10% (a very weak example of "often", but I'll be generous) of them had this alleged feature. I couldn't even find a relevant reference with Google...



Google "beat cancel" rather than "beat switch".

Analog tape noob: explanation needed

Reply #24
I've remembered: the FM pilot tone is removed by the "MPX filter" (multiplex filter) in a couple of tape decks I still have..

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


Cassette decks often had a "beat switch" that changed the bias oscillator frequency if you had problems trying to record FM radio, where the 19 KHz pilot was often present at significant levels in the tuner output. You can bet that the manufacturers wouldn't go to the expense of adding the switch if it weren't a common problem. It was apparently a cheaper solution than trying to filter out the tone in the low end systems (radio/cassette boomboxes), though 19 KHz LC filters were quite common in more expensive decks - switchable on some models capable of HF response to over 20 KHz.


I've never ever seen this in my life. There's plenty of tech information about legacy cassette decks online, so please provide evidence that even 10% (a very weak example of "often", but I'll be generous) of them had this alleged feature. I couldn't even find a relevant reference with Google...

The beat switch was commonly found on things that contained both a radio and a cassette deck. I haven't personally seen one on a stand-alone cassette deck. If you google radio cassette beat cut switch you can find plenty of archived manuals that mention it.

e.g. see PDF page 6, section "recording from the radio" here:
http://www.robertsradio.co.uk/assets/User_...PDFs/cd9909.pdf
"When recording from the MW or LW bands a whistle may occur. To remove the whistle slide the beat cut switch (25) on the rear of your CD radio cassette to the position giving best reception."

Cheers,
David.