Recording bars and TONE

[Jamie Brooklyn]

Hi guys,

I was once told when in production to record 30 seconds to a minute of bars and tone, but is tone only something you can record when working with a mixer?

I use a Sony Z1P camera, with a Rode NTG-1 plugged directly into it's XLR input, so does the tone not apply to me?

I only ask because when I put my Z1P's bars on, no tone comes through, just bars, which leads me to believe that I'd need a mixer to actually emit some tone.

What exactly would I do with tone once I had recorded it? This brings me to my next semi-relevant question of bars, which as far as I understand, are recorded to calibrate your editing monitor correctly. Is this correct?

Thanks guys, your help is much appreciated :)

[Richard Crowley]

The purpose of tone (and color-bars, for that matter) was to record a known calibrated standard signal so that you could "decode" the reproduced signal during post-production. Back in analog days, the only way to determine where a particular audio level fell was to record a reference calibrated tone of some known level. Here in the digital era, levels are self-identified because they are absolutely (in the mathematical sense of the term) recorded in numeric values. So when recording directly into a digital camera (or an external digital recorder), in effect, audio levels are "self-identifying".

HOWEVER, whenever you must interconnect two pieces of equipment using an ANALOG signal, then you need some way to coordinate levels between them. So most production-style mixers include a reference tone generator that you can use to coordinate maximum levels between the analog mixer operating points and the analog input to the camcorder and/or audio recorder.

Likewise, for video, color-bars were used to provide a known reference for analog video reproduction. Old analog video tape/cassette recorders had dozens (or scores or even hundreds) of controls and adjustments to get the reproduced signal to be most equivalent to the original video signal going INTO the recorder. But in the digital (and particularly post-tape) era, the need for color-bars is, strictly-speaking, superfluous.

In cameras (as in audio) there is still PART of the signal path that is analog. In cameras, some designs turn the analog light values into digital right in the imaging chip(s). But those cameras that still have analog circuits for the video signal arguably have some need to record an artificially-generated reference signal (i.e SMPTE color-bars, etc.) to show that all (except the optical/imaging) parts of the camera are operating properly.

Indeed, at the reproduction end (most importantly including post-production editing, etc.) there is a need for reference signals (both audio and video) to calibrate your monitor speakers and picture display. But it is completely valid, here in the digital era, to have these audio and video reference signals generated by the editing workstation. By the time you get the digital video and audio recordings, the recording "reference" is integral to the recording method and reference signals no longer serve the same purpose as they once did.

[Jamie Brooklyn]

Richard,

When you say "but it is completely valid, here in the digital era, to have these audio and video reference signals generated by the editing workstation", are you saying that the workstation produces the bars and tone, based on the numeric values that the camera is outputting?

And what exactly do you mean by "the recording ""reference"" is integral to the recording method"? What is the difference between recording reference and signal?

Thanks for what was an above and beyond answer, much appreciated.

[Rick L. Allen]

Recorded SMPTE/NTSC video bars and 1 KHz audio tone (at -18db - broadcast level) are the reference signal. Recording bars & tone even on digital media (P2, EX. etc.) are important to proper setup and establishing normal/broadcast/in spec video & audio levels in the field and throughout your edit system. I see many video and edit problems that are the direct result of inattention to this detail and a lack of quality control throughout the acquisition, production and delivery process.
SMPTE color bars - Wikipedia, the free encyclopedia

[Richard Crowley]

Originally Posted by Jamie Brooklyn

... are you saying that the workstation produces the bars and tone, based on the numeric values that the camera is outputting?

No. Digital signals are ABSOLUTE. There is ZERO ambiguity in interpretation. I am talking about test signals like bars which you can use to calibrate your computer monitor for accuracy of picture adjustment, color grading etc. And audio tones that you can use to adjust your monitor speakers for frequency response and acoustic level (SPL). These have NOTHING to do with whatever video came in from outside sources (including whatever came from your camera). You should calibrate your picture monitor and speakers for best editing results. You can do this before you have shot or ingested even a single frame of video or second of audio.

And what exactly do you mean by "the recording ""reference"" is integral to the recording method"?What is the difference between recording reference and signal?

I mean that a digital value is a digital value. Whether it represents the amount of money in your bank account, the temperature down at the airport, the audio signal level for 1/48000 of a second, or the values of red, green, and blue for a particular pixel in a particular frame of video (or image from a digital still camera, etc.)

In digital recording there IS NO DIFFERENCE between recording "reference" vs. "signal". And that is EXACTLY why tone and bars are superfluous in the modern era. You cut directly to the key issue. :-)

There are a finite number of bits available to represent the entire gamut of possible signals (audio or video) out in the real world. We typically must adjust what the microphone or image chip picks up so that it fits within our finite range of digital values. For example all decent audio is at least 16 bits "deep", and we often use 24 bit recording to accommodate a wider dynamic range. And you will find digital values of 8-bits, or maybe even 10-bits or 12-bits for professional gear. Special "digital cinematography" cameras can store video data with even wider range of values.

But in ANY case, whether you are storing bank balances, temperatures, audio samples, or video pixels, you have some FINITE number of bits. If you try to store a value that is even +1 above the maximum capacity of the value, you "clip" the data. This produces ugly distortion in audio, and "blows out" bright areas in video. Clearly "clipping" is not desirable if you are keeping bank balance figures for rich people(*), or if you are measuring airport temperatures at noon in a desert. So we design mechanisms (hardware, codecs, file formats, software, etc.) to either allow for the widest anticipated values, or else we somehow "limit" the range of values to fit within the mechanism.

There is also a computer technique called "floating point" where the numeric value reference is "adjustable", but the reference value is stored along with the signal value, so the number is still "absolute".

But in any case, audio and video values have some fixed number of bits of "precision" or "depth". And a word with all the bits turned on except the MSB (most significant bit) represents a value that is exactly 1/2 the range of the system. An audio (or video) value is a finite number within the allowable range, and the number is its own interpretation of "level".

Now Mr. Allen brings up an important point. We still have large parts of the distribution chain (including networks and local broadcasters) who either still have analog parts of their signal chain OR they are still using antiquated "setup standards" from back in the analog era. Many still want "Tone and Bars" reference somewhere, typically before the program starts. They also want quaint "countdown" devices (like the famous Academy Leader countdown we are all familiar with. Since we don't cue up film (or even tape) anymore the necessity of a countdown is a throwback to a bygone era.

But nevertheless, if you want to distribute your finished production to someone who wants bars, tone, countdown, slate, etc. etc. then you are compelled to oblige their fetishes whether it makes any sense in the digital era or not.

Note that the tone and bars you put on the head of a finished production are NOT the ones that came out of the camera. They are generated fresh in the digital domain by the editing software.. They allow analog-domain interpretation of the digital signals (which intrinsically need no such "interpretation").

I apologize for likely rambling too far afield for your simple question. But these old analog-era fetishes are rather a hot-button for me and they get more ridiculous every day. ESPECIALLY for file-based distribution (vs. digital tape). If I was a busy traffic coordinator at a local TV station ingesting commercials into the automation system, it would be particularly irksome to have to slog through all that antique slate tone/bars, countdown, 2-pop, etc, etc. for video that came in via email (or file transfer) as a video file. We have meta-data tags for video files now that can contain MUCH better information than anyone ever put into a "slate" at the head of a tape.

(*) Chances are excellent that if you are reading this on a computer, YOU are in the"1%" of the richest people on the planet.

[Trevor Dennis]

Richard, don't apologise. From a lurker’s point of view, this is pure gold.

[Brian Drysdale]

Since we're talking about a tape camera, it would be a good idea to put up some bars even if only to reduce the risk of drop out etc or at least ideally don't shoot material over the first 30 seconds or so.

Some of these things are belts and braces and if there's a means of a level being knocked out in a network or system, you can bet it's going to happen some time. I hear there are professional editors who adjust their monitors to what looks nice to them, rather than correctly calibrating them. Which may or may not work for them if all they're just doing some cutting, although not the best idea for colour correction etc.

[Rick L. Allen]

Originally Posted by Brian Drysdale

Since we're talking about a tape camera, it would be a good idea to put up some bars even if only to reduce the risk of drop out etc or at least ideally don't shoot material over the first 30 seconds or so.

Absolutely! If you're still using tape 90% of drop outs occur in the first minute of the tape (mostly wear and tear from threading and tape tensioning). Shooting bars & tone over the first minute of tape also gives your deck clean time code for preroll & ingest.

Originally Posted by Brian Drysdale

I hear there are professional editors who adjust their monitors to what looks nice to them, rather than correctly calibrating them. Which may or may not work for them if all they're just doing some cutting, although not the best idea for colour correction etc.

I can adjust a monitor to look great but if it isn't properly calibrated to a reference signal (bars) then the picture it gives me is meaningless. I need to see the real, footage warts and all. Thanks to Richard & Brian for salient points.

[Jamie Brooklyn]

Thanks for the replies guys, and especially to Richard for what were above and beyond answers!

Every time I have question now I re-read what he said and usually find the answer :)

Although a little fuzzy on some details, more and more makes sense the deeper I get into production. I'm looking to make the mistakes now rather than later.

Recording bars for the first minute if not for anything else but avoiding drop out, and achieving clean time code, are two ideas I'm definitely sticking with.

Thanks again for the support.