Goon tapes question for Ted Kendall

[TIMTAPE]

Hello Ted.
I'm hoping you're the same fellow who has done conversion of old Goon Shows to digital.
I have on CD some of what are probably some of the Goon shows you worked on. I must admit I was amazed at the lack of tape noise in most of them. Did you employ any sort of noise gating/noise reduction/ expansion in converting these tapes?

Cheers Tim

[jamesperrett]

I'm not Ted, but I also restore old recordings and I find that clients are often surprised when I tell them I've not applied any noise reduction to their old recordings. Decent reel to reel recordings are much quieter than many people expect. I guess they are used to cassette standards of noise.

Cheers

James.

[ppppenguin]

I'm not Ted either but he has told me of the trials and tribulations of piecing the Goon Shows together for re-issue. This includes working with different versions used for domestic, overseas and transcription services. I believe the last were on acetate discs. I know he is very expert at getting the best out of any groove. Different stylus sizes are just the start - he has designed specialist hardware too. His experience goes back well before digital techniques were available.

[Ted Kendall]

Yes and no. The raw signal-to-noise ratio on an original Goon Show tape (TLO in the parlance) can nudge 65dB - this is at 15ips, full track. Having transferred this properly into digits, which is a story in itself, you don't need to do much to the basic sound of one of these, unless there are technical faults to correct, such as top loss or magnetisation rumbles. Inserts are generally off disc, so the clicks and crackle come out of those, unless of course the gag involves the sound coming off a scratchy disc! A careful pick-over for thumps, fader crackles and other nasties, and the job's just about done.

Unfortunately...(as Hoffnung said), most Goon Shows don't survive in that form. To take an example, this very day I have been working on "The Spectre of Tintagel" for the next Goon Show Compendium volume. The TLO no longer exists, so the available sources were the Transcription Service disc issued in the 1950s (cut to 29'30"), a second-generation dub with mild compression made in the 1960s for the Transcription reissue (cut to 27") and a domestic 3.75 ips recording made off FM at the time of the original transmission (with the musical items omitted). The procedure here is to select the version with the best overall sound as the basis and restore the cuts from the best available source. In this case, the TS2 tape was the best starting point, as the compression was not obtrusive. Surprisingly enough, it was possible to bring the domestic tape to the point where it would cut seamlessly into the main take. The closing sig was cut on the TS2 version, and had to be restored from the TS1 disc. And that, my dears, is the bones of the story.

Now, kit. You can't do this for twopence halfpenny. Professional tapes generally get played either on a Studer or a magnificent old Telefunken with the butchest dual capstan arrangement I've ever seen, which makes short work of wrinkled or cupped acetate stock. Domestic tapes are played on a Studer 810, which has adjustable replay time-constants (essential since there was no standard at 3.75 until 1966) and very low noise replay electronics. This last is essential when large amounts of treble boost sometimes have to be applied. Convertors are Prism, workstations are Sadie and noise is dealt with using Cedar Cambridge. The whole point of this, however, is that any processing should not draw attention to itself except by excellence of the results. If processing is audible, something's gone wrong or the quart wasn't going to come out of the pint pot in the first place.

[TIMTAPE]

Thanks for the reply Ted.

To be more specific, assuming you had a pristine Goon Show TLO, would you still apply dehiss or other more broadband NR?

Cheers Tim

[Ted Kendall]

Broadband NR is a dangerous thing, in my view. If you start pumping the background up and down at LF, the reverberant cues on which the ear relies to anchor the sound in the room are compromised, hence the "there's no there there!" effect.

Fortunately, the most troublesome noise is above about 2kHz - a moment's contemplation of the replay curve will illustrate why - and here it is possible to reduce hiss substantially without causing damage. Obviously, you need software which allows you apply different amounts of NR in different parts of the spectrum, and Cedar NR-5 allows this. A typical TLO would just have a skim of 3-4dB above 3kHz or so, plus a tad of extreme top lift. On the fifth series, where the top was a little gentle because that series was done on AXBTs, I used more NR to allow top lift lower down, giving more crispness without a noise penalty.

With all this goes a serious point. If you are going to do this sort of thing, PLEASE retain a pristine original or a dub thereof. I have recently had the maddening experience of remastering a Navy Lark and having to use an AM tape covered in interference because the only remaining copy of a decent tape from FM had been ruined by somebody "cleaning it up" with one of these 30 buck dehiss packages. It wasn't the cyclic hiss I couldn't live with, it was the omnipresent space monkeys over everything. Whatever you do by way of restoration, somebody somewhere sometime is going to qvetch about it, and if you have preserved the source intact, your posterior is covered.

[ppppenguin]

Originally Posted by Ted Kendall

With all this goes a serious point. If you are going to do this sort of thing, PLEASE retain a pristine original or a dub thereof.

This raises the question of what is a good dub. Clearly a decent AD converter will have a noise floor way below any likely source and plenty of dynamic range too but do you really to sample above 44.1/48kHz? Does the ultrasonic info (if any) on a disc make any difference? Is it just good practice to sample at 96kHz or 192kHz

[paolo]

192Khz 24 bit is the norm these days letting shapers do the downconversion. Better to oversample than undersample. Remember, rubbish in, rubbish out

[jamesperrett]

Originally Posted by ppppenguin

Does the ultrasonic info (if any) on a disc make any difference? Is it just good practice to sample at 96kHz or 192kHz

I'm going to keep the quote from Ted in there because it is very important for anyone copying from old sources to keep the original and the straight digital copy.

Most of my tape dubs are done at 44.1kHz 24 bit. There is a case for for material that will be de-clicked to be recorded at a higher sample rate as it makes it easier for the de-click algorithm to discriminate between clicks and wanted audio. Most expert opinion that I've seen suggests that there is no point in going higher than 96kHz as the performance of an analogue to digital convertor will be worse at 192kHz.

The exception to this rule would be if you are trying to recover the bias signal from the tape in order to remove speed variations - as in the process mentioned at

http://chace.com/services/id/13,170

Cheers

James.

[Ted Kendall]

44.1 is good enough for most purposes - extra bit depth is, to my mind, more useful - and my Prism is to all intents and purposes completely transparent at any sensible sample rate. The archive community is gravitating towards 24/96, which at least gives plenty of room for error. 192 is, as yet, for the spec-chasers, in my view - it's unlikely to do any harm if correctly engineered, but is wasteful of disc space. Cedar can clean up a signal at any sample rate, anyway, and should a particular process require a wider bandwidth, such as ultra-accurate digital filtering, then upsampling will meet the case.

The retrospective clocking of an analogue tape recording with the bias signal is one of those ideas, like synchronising multiple copies of a disc, which is at first sight so elegant that it must work. In practice, I doubt whether it can deliver useful results outside the laboratory, much as we might wish it to. There are other ways of straightening wow out, but they are as yet very processor-intensive and hands-on. One day, maybe...

[TIMTAPE]

Ted,

Thanks for your replies.

I probably framed my question wrongly. I've been a big fan of Goon Shows for many years and still listen to them. Maybe it's partly because much of my listening to them over the years has been over local and national AM radio where dynamics are limited, but I find the CD versions I have are frustrating to listen to for very long because there is too much dynamic range for comfortable listening, even in a quiet environment.

I find if I compress the audio peaks by 5 to 10 db they are much easier to listen to and I can hear much more of the quiet offside jokes such as when Spike Milligan mutters "you're not the only one" etc.

In the end I guess this is a question of personal preference, but I must say that if I had been remastering these old shows (I also remaster analog tapes to digital commercially) I would have moderated the dynamics a bit even at the risk of being accused by some of engaging in "volume wars", and also at the risk of allowing some tape noise to be more audible.

Your thoughts.

Regards, Tim Gillett

[Ted Kendall]

Noise reduction has little to do with dynamic range, save that compressing a hissy tape will make it hissier. And as someone once said, once you've made the hamburger you can't get the cow back.

The point of the Goon Show issues, particularly those in the Compendium, was to preserve these shows in the best possible quality, and that involves retaining the intended dynamics of the recordings, which were subject to fairly strict guidelines at the time. Obvious fader errors are ironed out, and, should the average level drop in a scene, there is a little manual gain riding just to bring it back on the cam, so to speak, but nothing else. Sometimes, as I have said, I must accept a compressed dub as a source because it is the best in other respects, but TS didn't bend them double the way modern AM transmission in particular does.

In any case, and this has been mooted many times, not least by Glenn Gould, perhaps the consumer should have access to a compressor so he can screw it up...I mean "optimise the dynamics" for himself. With computer technology, this is now economic. Whether it's sensible is another matter, but then, you bought the bridge - blow it up if you wish.

[TIMTAPE]

Ted,

I suppose that since the advent of the CD as a consumer medium, when remastering shows like the Goons one really can give people the full dynamics as they were originally recorded. My question is whether this is necessarily a good thing. Obviously you are very firm in your view that it always is a good thing. It seems I'm more flexible on this.

The Goon Show is not the 1812 Overture complete with fortissimo's and canon blasts. It's a comedy show. The most important thing is hearing all the voices clearly, following the story and getting all the gags. Big dynamics mitigate against this. Or is there something I have missed here?

Regards Tim

[Ted Kendall]

To put it as plainly as I can:

This is expected to be the last time these shows will be remastered.

The avowed aim of the series is to bring the listener as close as possible to the original performances, and that includes preserving the original intended dynamics.

You cannot undo compression.

You can compress the sound off the CDs if you prefer - you bought it - you do what you like with it; but if you think dynamic range has no place in comedy, think again. Light and shade is essential to a programme of any length, and the lack of it is a prime cause of listener fatigue. Furthermore, one of the glories of The Goon Show is the quality of the sound. Any dynamic compression necessary was done at the time of recording by gain-riding.

The series is a critical and commercial success on this basis.

If it ain't broke, don't fix it.

[TIMTAPE]

Ted,

To return the favour and put my point as plainly as I can:

Of course you cannot undo compression. Of course even a comedy show has to have dynamic light and shade.

But somewhere in between the often severe crushing of dynamics of an AM radio Goon Show broadcast, and your otherwise immaculately produced CD transfers, is a whole dynamic range middle ground. A sensible compromise. It was that middle ground I was referring to.

If you will not concede that even 0.5db of limiting might be an improvement for the average listener then I guess we will just have to agree to disagree.

Regards, Tim

[Kat Manton]

Hi,

I've been following this discussion with some interest, both as a Goons fan and from a technical perspective.

I don't need to express my opinions concerning the preservation (or otherwise) of dynamic range; Turn Me Up! has it pretty well covered.

Neither Cobweb nor I have experienced any discomfort or fatigue listening to material featuring a good dynamic range, we just turn the volume up so the 'quiet bits' are comfortably audible and the 'loud bits' are as loud as they're supposed to be.

On the other hand, we both find material lacking in dynamic range can be boring and ultimately tiring to listen to. Worse still are recordings where I can hear compression 'pumping' and/or background noise varying; I find myself listening to the processing and getting annoyed rather than listening to and enjoying the the content!

Sorry Tim; I'm with Ted here. If he'd compressed the recordings to suit your preference, they wouldn't suit mine - and I wouldn't be able to do anything to get the dynamic range back. Once it's gone, it's gone.

As it is, both our preferences can be satisfied if recordings are released with their dynamic range intact. You're welcome to apply compression if that's what you prefer. I'll just turn the volume up

Regards, Kat

[TIMTAPE]

Kat,
In an earlier post I specifically stated that I wasnt trying to enter into a "loudness wars" debate as I knew that I might be misinterpreted as saying that.

I wasnt making a general comment about the industry. I was referring specifically to a set of Goon Show CD's I own that Ted had remastered, and may I say, he has remastered briliantly, with the possible exception of the dynamic range issue.

You said that the dynamic range on these recordings that would suit me would not suit you and Cobweb. But have you heard the dynamic range that would suit me? Obviously you havent.

Most people who buy such CD's havent a clue even what dynamic range is except in the most general terms and would be at a loss to pick a specific example of limiting in a recording, far less to estimate just how much limiting in db's was used. In discussions of dynamic range and compression on other audio recording discussion forums I have at times referred to specific examples of compression on a contemporary release to teach newer recordists what to listen for so that they can get a handle on the subject - without my necessarily trying to impose my personal preferences either.
For many, audio processing issues such as this are largely a mystery.

You are of course free to side with Ted on this but it would be a far more enlightening discussion if we used actual audio (A and B) examples to illustrate exactly what I mean. We could also use waveform graphs as per the loudness wars website you quoted too. And if we did the latter I very much doubt that the dynamic range changes that would suit me would result in a graph with the terribly crushed waveform illustrated there. It would be far far less severe limiting than that.

Kat, do you understand what I am, saying here?

Cheers Tim

[TIMTAPE]

Thinking about this further (I realize this entire thread goes somewhat beyond the scope of the subject of vintage tape recorders) it's important to remember that back in the 1950's they didnt have the multitracking recording facilities taken for granted in later decades. With the later multitracking, each of the microphones and other inputs used in the making of the Goon Show could have been recorded to its own separate track. Later on, at leisure, post production staff could have set about rebalancing each voice/microphone/track against the others in ways that with a single track recording is just not possible.

Recording a live show such as this to a single track is by comparison perilously difficult and even with the best panel operators in the world, the balancing results can never be what they could have been with the luxury of multitrack facilities.

I salute the technical staff as well as the performers involved in The Goons for the recorded results they achieved on a single track. But it means that the only options to try and rebalance relative levels later on are severely limited to manipulating the gain and EQ of that one track, because that is all we have to work with.
Still, that is something, and I for one am a firm believer in doing so insofar as it is done unobtrusively and with careful judgement.

Kat, a person experienced in the use of compressers/limiters is often able to limit louder portions of a track, even a composite mix such as this, without introducing annoying pumping of the quieter background sounds. In fact this can be done so well that it takes a skilled ear to even detect that it has occurred at all. Skilled use of limiting doesnt draw attention to itself but merely makes the program more listenable to a wider audience. Naturally there are definite limits to how far one can use this technique without introducing annoying artifacts.

I recall remastering some interview tapes some years ago of a well known sporting identity. They were recorded in London in 1959. They were not of professional audio standard and one of their main failings was that the volume of the interviewer's questions was somewhat louder than that of the sportsman he was interviewing. You can imagine that this would be very offputting to someone listening. In the process of remastering I applied some limiting of the program so that the interviewer's voice level was now reduced somewhat so that the important voice came up closer in level to his. It did not solve the problem completely of course but it helped. Obviously if the interviewer had had available a 2 track recorder where he could have recorded himself and the sportsman individually the options would have been much wider later on.

For many years the recommended technique in recording Oral History tapes (and now to digital) has been to use the two microphone technique onto two independent tracks.

I suspect the problem can come when we fail to understand this dilemma. Also there can be a tendency to regard well loved recordings such as this almost as Holy Relics which must not be changed or manipulated in any way. While it's true due to the simple way they were recorded that the options are limited, it can still be worthwhile doing what we can. One of those options is limiting dynamic range, especially in the interests of rebalancing voice and microphone levels insofar as that's possible.

I hope this background information explains more clearly where I have been coming from in these posts.

Regards Tim

[Ted Kendall]

Firstly, thank you, Kat - nail struck firmly on head.

Now, Tim. Arthur Haddy used to say "if you can't go straight to two-track, you shouldn't be in this business". Like all Haddyisms, it wrapped a grain of truth in picturesque exaggeration, but the basic point is sound - if you can't produce something like it on the day, no amount of remixing is going to help. To suggest that The Goon Show suffers by not being on multitrack is bordering on the absurd. In fact, I did remix The Last Goon Show of All (1972) from eight-track with my esteemed colleague, Dirk Maggs, and the reason that exercise was worthwhile is that a lot of the art of sound balance used in the original shows was sacrificed to stereo and expediency. Putting back the naturalness took a lot of jiggery-pokery, paradoxically enough.

My point all along has been that The Goon Show was compressed as much as necessary in the first place - manually. The dynamic range broadcast by the BBC at the time took account of reception conditions and was restricted to 22dB, the difference between PPM"6" and PPM"1". Occasionally accidents happened, as I have described, and these have been corrected as far as possible. I recall a gunshot which came out far too loud, having been cut onto disc with a cutter which went into paroxysms at the transient. The solution here was to drop the level at the shot, using an edit in the DAW, and trim the return fade so that it was inaudible. Problem solved - without shoving the programme through a compressor. The intended dynamic range was restored.

I have no quarrel with compression as such, if it is applied where necessary. In my view, if compression is audible, you have failed. The sound does not occur in nature, and this distracts the brain from the programme material. The most usual reason for compression is to restore a natural-sounding dynamic range to a voice or instrument miked closely (and therefore exhibiting an exaggerated dynamic range) or to make an electric bass or similar instrument sit nicely in the mix. I recorded an album with a famous black singer-songwriter which has a large degree of compression on it - but I'd defy you to hear it as such.

As for your example of imbalance between interviewer and interviewee, the answer is not, in my view, to shove it through a compressor - you have two speakers, each holding the stage in turn. Simply cut at the change of voice, adjust the clip as necessary, and cut back when the voice changes again. Mistake corrected, microdynamics preserved - and if you're clever about where you cut, the change in tape noise will be masked.

The idea of compression as a panacea goes back at least to the 1940s, when Sarnoff thought he could sack all his expensive balance engineers by putting compressors on every mic - of course it didn't work, and half the things people try to do with compressors don't work to this day, because they are so frequently misapplied. Gain-riding is a skill so undervalued these days that it is in danger of disappearing altogether - but the fact that it is under the control of the operator and his ears is what makes it so valuable.

[Kat Manton]

Hi,

Tim, I suspect I'm being misunderstood. I'm not averse to the use of compressors/limiters and have used them on individual tracks, groups and the stereo mix (along with sometimes very extreme EQ, deliberate tape saturation and all manner of 'creative mangling'.) My Linux-based home studio setup (which I'm still setting up) includes JAMin and a 3-head mono valve tape recorder, among other things. But that's all concerned with the creation of an 'original recording' in the first place.

However, this thread concerns half-century old recordings made on equipment with limited dynamic range (compared to modern technology) being remastered and reissued on a now 29-year-old delivery medium, itself with limited dynamic range (compared to modern technology.)

Aiming for a 'dynamic range middle ground' would inevitably involve personal preference. Given I find practically all FM radio stations over-compressed and can barely tolerate AM radio, I suspect my preference leans towards a higher dynamic range than is typical.

As has already been stated, not compromising dynamic range means you, Tim, can apply compression and I can turn the volume up. So we're both happy and there isn't a problem, is there?

The 'preservation/archival' angle hasn't escaped me. Releasing these shows in a form that's as close as possible to the original recording effectively provides multiple 'off-site backups' and goes a long way towards ensuring their continued survival.

At least this thread means I'm not put off by "Remastered using new material and the latest technology to give the best possible sound quality" appearing in Amazon's description. Usually a phrase like that has me looking for earlier releases...

Regards, Kat.