Nagra 4.2 track layout

25 replies [Last post]
galleard
Offline
Joined: Apr 18 2007

Playing back some film soundtrack recorded on a Nagra 4.2 (circa 1983), using my Teac A3340S 1/4 track machine. Audio is mainly FX so absolute synchronisation is not vital. I'll sync up with picture in the timeline.

I'm getting clean signal from tracks 1 & 4, but 2 & 3 have a 'hum' signal on top of the audio. Internet research suggests that the Nagra was a full track mono recorder, but could also record a synchronising signal. Out of interest - where is this sync signal on tape? Are the tape heads actually twin track with sync track in the centre?

I can get by with audio 1 & 4, but a full understanding of the issue would be useful.

Colin

Rob James
Offline
Joined: Jun 26 2001

The sync track (assuming this was recorded on a Neo Pilot machine) is indeed in the middle. If it's a mono recording then the sync head consists of two out of phase gaps and the audio is intended to be read by a full-track head which cancels out the sync pulse. For a fuller explanation see here:

http://chambinator.free.fr/english/kudelsus.htm#pilote

If it was recorded on a stereo Nagra then the Pilotone sync system is a little different in that it uses a 13.5kHz FM signal. (still in the middle of the tape.) Later models had the FM board replaced with a TimeCode board, but the recording principle was the same. Worth noting that unless you tranfer using a full track head (or possibly tracks 1&4 if the azimuth is up to it) you may get thumps on edits. Just filter the recordings steeply at 50 Hz and they should go away.

Hope this helps.

Rob The picture is only there to keep the sound in sync

John Willett
John Willett's picture
Offline
Joined: Jun 1 2001

The Nagra 4.2

The NAGRA 4.2 is a portable mono full track 6.35-mm (¼ inch) analogue audio tape recorder designed for radio, cinema and television applications.

This recorder has three speeds: 38, 19 and 9.5 cm/s (15, 7 ½, 3 ¾ ips), NAB or CCIR standard. Two mic. inputs, switchable to dynamic, T or P (48V) condenser. One voltage or current line input. One direct mixer input, 4.4V or 1.55V balanced line output, depending on the transformer installed.

The NAGRA 4.2 contains separate recording and playback heads for confidence playback, built-in monitoring speaker switchable to source or tape, a reference generator, modulometer, alarm indicator for power supply and tape transport. It also has high-pass filters for recording or playback and a switchable automatic level control and limiter.

The machine can be fitted with an optional NEOPILOT 50/60 Hz sync system with a built-in quartz generator and may also be powered from either an external supply (ATN-4) or internal batteries.

Your post implies you are playing a full-track (ie 1 track) tape back on a 4-track machine :confused:

John
 
A picture tells a thousand words, but sound tells a thousand pictures.

MAGLINK
Offline
Joined: Mar 8 2007

I spent the first five years of my career in the Tyne Tees Tv transfer bay with a pilotone nagra going to a sondor 16mm mag machine.
We also had an SLO which had a scope on it so you could set the 50hz pilot tone to the correct level, the whole lot was then referenced to 50 mains for transfer.

15 years ago a YTV we still used mag on touch of frost but then ran a video picture synced up to timecode with a colin broad pilot to timecode box.

As the guys say the nagra recorded a 50 hz pulse that was locked to a crystal, the camera also had such a crystal so that it ran at 25fps.

galleard
Offline
Joined: Apr 18 2007
Rob James wrote:
If it's a mono recording then the sync head consists of two out of phase gaps

The audio is mono, and tape box labelled as such. Thanks for the link. If the sync head is half track width as shown in the diagram then I should hear it on all four quarter tracks - but (luckily) I don't. So it must be as you say a narrow centre track.

Quote:
Worth noting that unless you transfer using a full track head (or possibly tracks 1&4 if the azimuth is up to it) you may get thumps on edits. Just filter the recordings steeply at 50 Hz and they should go away.

OK. Currently I'm recording the two clean tracks to twin channel WAVs. If there is a problem then I suppose I can ditch one of the tracks. Sounds OK on headphones. Soundbooth's bar meters do show some slight difference in response and levels from the two channels, so you may be right.

Thanks for the explanation.

galleard
Offline
Joined: Apr 18 2007
John Willett wrote:
Your post implies you are playing a full-track (i.e. 1 track) tape back on a 4-track machine

Yes - needs must.

But in this case, since I don't have access to a Nagra or full track machine for playback, using the two outer quarter track channels would seem a lucky solution. I'd have had a problem with a 1/2 track machine.

Rob James
Offline
Joined: Jun 26 2001

A 1/4 track machine where you can play both the outer tracks is a good solution provided you adjust the head azimuth to suit.

That diagram is not strictly accurate. The audio record head bias wipes the sync pulse except in the narrow middle area. However, true half track heads have a very narrow guard band and will likely pick up the Pilotone. Proper 2-track heads have a wider guard band that misses the sync track.

When I started in sound transfer at Ealing Studios at the BBC we had at least ten different head blocks for the various weird and wonderful sync systems. I cannot remember what they all were now but there were a hell of a lot of them.

Rob The picture is only there to keep the sound in sync

galleard
Offline
Joined: Apr 18 2007
Rob James wrote:
A 1/4 track machine where you can play both the outer tracks is a good solution provided you adjust the head azimuth to suit.

Doesn't poor azimuth tended to knock out high frequencies? I don't have a set up tape or equipment to measure it, so if I had a go I'd probably make it worse rather than better.

Fortunately all the audio is fx. I guess dialogue would pose a much greater problem.

Quote:
When I started in sound transfer at Ealing Studios at the BBC we had at least ten different head blocks for the various weird and wonderful sync systems. I cannot remember what they all were now but there were a hell of a lot of them.

Ah, standardisation. So it's just luck then that I found a solution!

Many thanks to all for comments.

Rob James
Offline
Joined: Jun 26 2001

In the absence of a scope and experience you would be well advised to leave the mechanical azimuth alone. However, there is something you can do once you've done the transfers. If there is any line up tone on the tapes, split the stereo file into two monos and look at the waveform. Make sure that the peaks line up. If they don't, you can pull one track into alignment. Then compare the results on programme material. Put simply you are looking for maximum treble output and the minimum of 'phasey' effects. You could just use one track instead of two but I seem to remember that, in practice, you may get lower noise if you sum the two. I'm sure one of the more technical folks will be along to give a definitive answer.

As a BBC engineer once said to me, "Standards? I love standards, they keep me in business"

Rob The picture is only there to keep the sound in sync

tom hardwick
Offline
Joined: Apr 8 1999

Ah, the Teac A3340S. How I loved that handsome beast, but it sure needed add-on noise reduction units to compete with upper-class cassette.

galleard
Offline
Joined: Apr 18 2007
Rob James wrote:
However, there is something you can do once you've done the transfers. If there is any line up tone on the tapes, split the stereo file into two monos and look at the waveform. Make sure that the peaks line up. If they don't, you can pull one track into alignment.

Interesting. Yes, there is tone at the start of each tape. The sound recordist stated the tone was -8dB and from the waveform (and using a calculator) it seemed to be at about 1063 Hz. I've opened the two tracks in Soundbooth, and zoomed right into the waveform of the tone. Channel 1 is 0.00002 seconds behind channel 4. I'm not sure how significant this is.

I've spent the afternoon trying to find a way of splitting the two tracks in Soundbooth, to manually re-sync. But I don't think it's possible. So instead I've imported the twin tracks into Audacity, used "split stereo tracks", then when zoomed right in, manually dragged one track to match the other by eye.

Quote:
As a BBC engineer once said to me, "Standards? I love standards, they keep me in business"

:)

Rob James
Offline
Joined: Jun 26 2001

To be honest, with that small an error you are unlikely hear a difference. I can't be bothered to do the maths but I think that equates to 0.2 mS at 10kHz (which is the point at which azimuth is usually checked.) From that you can work out the phase error if your brain works that way.
All you are doing with this adjustment is removing the difference in azimuth between your machine's replay head and the Nagra record head.

The nominal line up tone frequency is usually 1kHz so, if your calculations are correct, either the Nagra was running a little slow, your machine is running a little fast or the tone oscillator was a fraction off. (None of this is anything to worry about unless you are dealing with music and people with perfect pitch. :) That said, my recollection is that Nagra oscillators usually worked at 900Hz so, if anything sounds speeded up, there's the answer. EDIT I just checked on Google and the User Manual http://www.csuohio.edu/class/com/dvcomm/Nagra%204_2_Op_Manual%20Short.pdf
says 1.1kHz, so even less of an error if your calculations are correct.

More to the point, in broadcast terms, the -8dB tone should be lined up to read -18dBFs on the workstation input meter. (Assuming this is calibrated in dBFs) Although in theory the tape recording should not exceed 100% i.e. 0dB, in practice there are likely to be transients that exceed this, even if the Nagra limiter was used, which will cause unpleasant digital distortion and aliasing if you line up to -8dBFs. This gives a 10dB margin for analogue overshoots and is the EBU standard line up level. (AES is -20dBFs)

Rob The picture is only there to keep the sound in sync

galleard
Offline
Joined: Apr 18 2007
Rob James wrote:
To be honest, with that small an error you are unlikely hear a difference. I can't be bothered to do the maths but I think that equates to 0.2 mS at 10kHz (which is the point at which azimuth is usually checked.) From that you can work out the phase error if your brain works that way.
All you are doing with this adjustment is removing the difference in azimuth between your machine's replay head and the Nagra record head.

Fair enough. Worthwhile to get an idea of the error involved. I don't want to make work for myself if it's unnecessary.

Quote:
More to the point, in broadcast terms, the -8dB tone should be lined up to read -18dBFs on the workstation input meter. (Assuming this is calibrated in dBFs) Although in theory the tape recording should not exceed 100% i.e. 0dB, in practice there are likely to be transients that exceed this, even if the Nagra limiter was used, which will cause unpleasant digital distortion and aliasing if you line up to -8dBFs. This gives a 10dB margin for analogue overshoots and is the EBU standard line up level. (AES is -20dBFs)

An issue I was confronting. Soundbooth record bar graph meters go up to 0dB full scale. Initially I set the meters to read -8dB on this tone, but found a lot of signal peaks got clipped. I've tried reducing this in steps to -10dB, then -12db, finally -15db. If -18dBFs is the norm then that gives me confidence to set that and proceed with this task. Thanks.

John Willett
John Willett's picture
Offline
Joined: Jun 1 2001
galleard wrote:
Yes - needs must.

But in this case, since I don't have access to a Nagra or full track machine for playback, using the two outer quarter track channels would seem a lucky solution. I'd have had a problem with a 1/2 track machine.

OK - if the tape has pilot tone on it you need to listen to the mono signal from two tracks - preferably the inner two - so the pilot tone totally cancels on the mono playback.

If you are using a quarter-track stereo machine, rather than a true 4-track you may have problem as one head will read an inner track and the other an outer and they may not balance.

It would be better to use a half-track steero machine rather than a quarter track.

John
 
A picture tells a thousand words, but sound tells a thousand pictures.

MAGLINK
Offline
Joined: Mar 8 2007

Yup as Rob says nagra was 900hz lower than the std 1khz. It was to allow you to hear the bleep tone as you spooled thru the tape in fast forward or rewind. Studio machines used 400hz for slate marking but they ran a lot quicker than the nagra spool times. The recordist always hit the tone button at the end of every take or at the end of a reel with three bleeps.

I also recall that the line up was -8db on the nagra scale so that you averaged at around 0 on their meter which equals -10db fullscale.

As Rob says the line up tone should be set at -18 db full scale and you should be ok for transfer.

Rob James
Offline
Joined: Jun 26 2001

John is quite correct if you are using a conventional 4 track recorder where the track pairs are 1&3 and 2&4 but I had the impression you are using a full four track with simultaneous access to any or all tracks. EG a Tascam or Fostex. In this case the pragmatic approach is to use the outer pair, 1&4. Although there is a chance azimuth error will be greater, I've always found this gives a cleaner signal from Neopilot tapes. (Yes, I've done this as well, although not for some years.)

Rob The picture is only there to keep the sound in sync

John Willett
John Willett's picture
Offline
Joined: Jun 1 2001
Rob James wrote:
the pragmatic approach is to use the outer pair, 1&4. Although there is a chance azimuth error will be greater,

I'm curious why you say this Rob, as I would have thought that 2 & 3 (inner tracks) would be better as the original would be full track audio and the inner tracks don't suffer from edge-effect.

John
 
A picture tells a thousand words, but sound tells a thousand pictures.

galleard
Offline
Joined: Apr 18 2007
John Willett wrote:
OK - if the tape has pilot tone on it you need to listen to the mono signal from two tracks - preferably the inner two - so the pilot tone totally cancels on the mono playback.

I haven't tried mixing the audio from two inner 1/4 tacks, ie track 2 and track 3. From what's been said, this should cancel the tone. The link Rob gave from earlier stated the tone was cancelled at source in the playback head. Is this 'cancellation' as good further down the audio path? If I've got a slight azimuth issue as well then maybe the tone won't completely cancel and I'll have to knock out what's left with software.

I have tried a straight software solution. The rumble filter in Soundbooth set to about 70Hz does seem to kill this tone, but I assume it'll affect all low frequencies.

So all in all I'll keep it simple and use channel 1 & 4.

Quote:
If you are using a quarter-track stereo machine, rather than a true 4-track you may have problem as one head will read an inner track and the other an outer and they may not balance.

No, the Teac A3340S is a true four track, four channel machine.

Colin

galleard
Offline
Joined: Apr 18 2007
Rob James wrote:
John is quite correct if you are using a conventional 4 track recorder where the track pairs are 1&3 and 2&4 but I had the impression you are using a full four track with simultaneous access to any or all tracks.

Yes, the Teac has independent access to all four channels.

Quote:
(Yes, I've done this as well, although not for some years.)

Seems the way to go from advice here.

galleard
Offline
Joined: Apr 18 2007
JGNattrass wrote:
Yup as Rob says nagra was 900hz lower than the std 1khz. It was to allow you to hear the bleep tone as you spooled thru the tape in fast forward or rewind. Studio machines used 400hz for slate marking but they ran a lot quicker than the nagra spool times. The recordist always hit the tone button at the end of every take or at the end of a reel with three bleeps.

Having never being initiated into the black art of sound departments I'm unsure of normal practice. On these tapes, the recordist laid 15 seconds or so of tone at or near the start, then most of the boards were end taps on the microphone. I haven't had any of the film transferred yet, but I've no reason to think sync'ing up on the timeline will be a problem.

Colin

Rob James
Offline
Joined: Jun 26 2001

OK, John first.

I agree about edge effect. However, to minimise the Neopilot buzz, staying away from the centre of the tape has always worked better for me. If we didn't have a full track head to hand we used to use twin track in preference to half track because of the bigger guard band in the middle. (with the outputs summed)

Colin, where to start? Assuming that the film was shot at 25fps the Neopilot tone on the tape will be at 50Hz. When transfering to film one of two methods was used. Either the 50 Hz was used to set the tape playback speed referenced against mains or a crystal and the sound was transferred to 16mm magnetic film with the mag film recorder also referenced to mains. Thus you ended up with one frame of mag film for each frame of picture film. The alternative was to use the nominal 50Hz Neopilot on the tape to drive the mag machine via a socking great amplifier if an old machine (Westrex etc.) or to control the stepper drive motor in later examples (Albrecht etc.) via bi-phase pulse to achieve the same result.
In this case you will have film telecined at (hopefully) 25fps and I suggest your best bet is to analyse the Neopilot frequency as best you can then varispeed the sound in the DAW so that it is running at 25fps. E.g. if the Neopilot is running at 55Hz then you need to varispeed slower by 9.091%. Even then don't expect the sound to stay in sync for long because the Nagra will have drifted in speed when recording.

If by any chance, the film was shot at 24fps then that's another ball game.

Rob The picture is only there to keep the sound in sync

MAGLINK
Offline
Joined: Mar 8 2007

This box will lock 50 hz sync pulses to timecode, we used to use it at YTV to enable a video picture to lock with the albrecht mag machines on touch of frost.

Also if you can get a hold of a nagra 4.2 it will play back in sync locked to its own crystal, it may be worth borrowing or hiring one for a day to transfer the sound to disk:http://www.warehousesound.co.uk/soundhire/equipmentInfo.php?equipID=288

We used to use nagras for location sync playback on The Tube and other music shows.

galleard
Offline
Joined: Apr 18 2007
Rob James wrote:
Assuming that the film was shot at 25fps the Neopilot tone on the tape will be at 50Hz.

I've transferred a couple of rolls now, and heard the following from the recordist's notes on tape. Two cameras are mentioned, Aaton and Arriflex ST. Sound tape is crystal synced to 50Hz.

Quote:
In this case you will have film telecined at (hopefully) 25fps and I suggest your best bet is to analyse the Neopilot frequency as best you can then varispeed the sound in the DAW so that it is running at 25fps. E.g. if the Neopilot is running at 55Hz then you need to varispeed slower by 9.091%. Even then don't expect the sound to stay in sync for long because the Nagra will have drifted in speed when recording.

I'm not entirely sure what to expect, but the quantity of this material is not too large. If I'm forced to do some timeline speed changes then I'll have to.

JGNattrass wrote:
if you can get a hold of a nagra 4.2 it will play back in sync locked to its own crystal, it may be worth borrowing or hiring one for a day

I currently have eight 5" reels now, possibly more later. Hiring would be expensive. My major cost will be getting the picture transferred by telecine to AVIs. I'm economising by transferring the audio myself, hence the original enquiry. With no dialogue to worry about, I hope that sync drift from using reel to reel will not be excessive. In any case, as with a film cutting reel, I'll have to sync up each shot once I get video and audio into the timeline.

The main issue was to get the best from the audio with my equipment. I'll see how it goes.

Colin

Rob James
Offline
Joined: Jun 26 2001

Let us know how you get on. If we can help in any way, let us know.

Rob The picture is only there to keep the sound in sync

MAGLINK
Offline
Joined: Mar 8 2007

A 4.2 hire would be £30 a day is that too expensive to get them transfered in sync???

galleard
Offline
Joined: Apr 18 2007
JGNattrass wrote:
A 4.2 hire would be £30 a day is that too expensive to get them transfered in sync???

I agree this doesn't sound expensive. But costs are rarely this low - carriage & insurance on top, and possibly a second session needed. I'm aware of a possible drift out of sync over time, which a better machine would minimise. If sync was vital as for dialogue then there would be a case for this, but for my needs this 'best solution' is an expense I can save.

Colin