DAT versus dvx audio

Re: DAT versus dvx audio

[quote author=MIDIMan link=board=Links;num=1080751160;start=0#2 date=03/31/04 at 09:08:55]

Digital audio is essentially the same whether you record it on a DAT or your DVX.  Anything hotter than 0db is clipped.
[/quote]

Hmmm... you must keep in mind that we're talking about recording ANALOG audio (that is, converting an analog signal to digital, then recording it), in which case the ANALOG and A/D components of the system are often what make the quality difference. 0dB is the max level for digital, but different A/D converters have very different amounts of headroom - capability to handle signals over 0dB - and different techniques for mitigating peaks.

The preamps are also extremely important in the process; I've found preliminarily that my external preamps give far better results than the DVX's. I'm planning on performing/providing some comprehensive test results sometime soon; let me know if you're interested...

Two interesting things in this discussion are the statement that the level meters never showed peaks, and that levels coming in "too hot" from a soundboard are problematic.

1) If the level meters never showed peaks, I surmise that the AGC/ALC (I never use it, either!) was distorting the signal pretty nastily in the analog realm (admittedly, because it was pushed past its limit with a line-level signal hitting a mic-level circuit). So, technically, this may not be "clipping" - are playback levels hitting peaks, or are they at the -3dB that ALC seems to prefer? If so, then it may be just really nasty limiting... I'd like to hear a sample!

2) With PROPER line level settings, I've found that the manual level controls on the DVX allow a pretty flexible range of input levels... can someone (midiman) say more about problem's they've had with sounboard feed?
 
Re: DAT versus dvx audio

[quote author=Rush link=board=Links;num=1080751160;start=0#9 date=04/01/04 at 21:08:01]
 Also, as far as I know, DAT records at 96Khz, DV at 48Khz.  I have been told by audio guys that it doesn't step out beyond the normal hearing range of the human ear, it just makes the sampling more accurate and clear. [/quote]

Actually most DATs record at (user choice) 44.1K or 48K; some now do 24bit, but there's no standard that I've heard of for 96K DAT. So, for the common DAT, the digital representation is exactly the same as for DV. For a 24-bit DAT, resolution of sound amplitude is 256 greater, but "range" of amplitude and noise floor are unaffected, and determined by analog components in the recording chain. Recording to an HD at 96K sampling gives you twice the frequency range (up to 48K) - going far beyond the generally "accepted" range of human hearing (up to ~22K).

Most professional DATs will have better analog componentry than the DVX; that is what makes the difference.
 
Re: DAT versus dvx audio

[quote author=FatDaddy link=board=Links;num=1080751160;start=0#19 date=04/02/04 at 08:28:27]
Any thoughts?  Is 16bit = to 48K?
[/quote]

The DVX records in either 16 bit/48khz, or 12 bit/32khz.

The DV specification also provides for 16 bit/44.1 khz, but that's not supported for recording by the DVX. So if you're recording 16 bit, you're recording 48khz. And if you're recording 48khz, you're recording 16-bit.
 
Re: DAT versus dvx audio

[quote author=ullanta link=board=Links;num=1080751160;start=0#21 date=04/02/04 at 10:12:27]

Actually most DATs record at (user choice) 44.1K or 48K; some now do 24bit, but there's no standard that I've heard of for 96K DAT.  So, for the common DAT, the digital representation is exactly the same as for DV.  For a 24-bit DAT, resolution of sound amplitude is 256 greater, but "range" of amplitude and noise floor are unaffected, and determined by analog components in the recording chain.  Recording to an HD at 96K sampling gives you twice the frequency range (up to 48K) - going far beyond the generally "accepted" range of human hearing (up to ~22K).

Most professional DATs will have better analog componentry than the DVX; that is what makes the difference. [/quote]
Thanks Ullanta!
 
Re: DAT versus dvx audio

[quote author=ullanta link=board=Links;num=1080751160;start=0#20 date=04/02/04 at 10:00:45]

can someone (midiman) say more about problem's they've had with sounboard feed?[/quote]

The problems I've had with a mixing console feed have been the direct result of the sound engineer. When I've set up my own mix, I've had excellent results. The bottom line is that when you tap into a house P/A system, the engineer (regardless of how good they are) is mixing for the house and not for your camera. The multiple feeds are then assembled in post.

Any time I go into an environment to film a live event, I always set up my own audio sources (usually several wireless systems and a minidisk). Each camera operator on the shoot with me is responsible for monitoring their audio feeds and making necessary adjustments to the input levels.

I agree with you that different systems have different methods of limiting audio peaks (ALC, AGC), but any digital signal hotter than 0db is clipped. It's just the nature of digital audio.

MIDIMan
 
Re: DAT versus dvx audio

[quote author=MIDIMan link=board=Links;num=1080751160;start=0#24 date=04/02/04 at 12:13:17]

I agree with you that different systems have different methods of limiting audio peaks (ALC, AGC), but any digital signal hotter than 0db is clipped.  It's just the nature of digital audio.

MIDIMan[/quote]

Sorry, MIDIMan, but that's really not quite right. You have to keep the analog/digital distinction in mind. There is NO SUCH THING as a digital signal hotter than 0 dB - it such a value cannot be represented digitally. That's why hot ANALOG signals have to be clipped to fit the digital domain. And there's little or no A/D circuitry that doesn't have SOME headroom above 0dB before it clips; the level of headroom is one of the big selling points of lots of A/D equipment.


(Note - I'm not discussing certain internal digital representations of DAW mixing busses which have greater than final range in their representations and can have "clipping" due to bus summation and then data truncation down to final representation; e.g., the protools internal represention is 48 bits, some use 32 bit floating point... but inputs and the final product have to fit a standard in which 0dB is the highest representable value)
 
Re: DAT versus dvx audio

[quote author=ullanta link=board=Links;num=1080751160;start=0#20 date=04/02/04 at 10:00:45]

 If so, then it may be just really nasty limiting... I'd like to hear a sample!
[/quote]


You're absolutely right here, but the effect would be about the same...clipping...a truncated waveform with severely flat tops.

I guess what I should have said is that line level signals fed at mic level, in the absence of limiting, would clip. In the presence of severe limiting, the effect is still distortion.
 
Re: DAT versus dvx audio

Some boring electro-acoustic remixed theory (if anyone is interested and/or unfamiliar with the subject) :

Hz/KHz measure the frequency — the number of times a sound vibrates per second. It's a temporal measure, but it deals with pitch actually.

A human ear in perfect condition can hear sounds from 20Hz (very low 20 vibrations/sec) to 20KHz (very high 20,000 vibrations/sec). Around 40 years old, a human being is more likely to be able to hear sounds up to 16KHz. This ear decadence keeps going on, falling down to about 13KHz at the age of 70 (all of these are average figures of course). Our "biological microphone" is mostly affected in the high spectrum, not the low spectrum, as our tympanum gets stiffer with age — like a very rusted piano string.

I couldn't exactly explain why without being boring to death, but A/D converters need to double their analysis range in order to render sound without any audible harmonic distortion, so that a 48KHz converter can sample sounds up to 24KHz (4KHz above any human ear's range).

So we see that the 96KHz issue is not a really important one. Just a theoretical plus.

What's important though is that when you get 96KHz, your sound is now analyzed in 24 bits, not 16 bits anymore.

Bits deal with dynamic range. Let's imagine a straight oblique line representing the ascent of sound from silence to deafening rattle. If we want to translate it into the digital language, we'll have to draw a "staircase" mimicking the oblique line (an oblique line is a digital impossibility). The more "steps" we'll use, the more realistic will be the mimicked line — very much like the pixel issue in digital video. A sample with a 24 bit resolution will simply have many more "steps" to imitate the line than a 16 bit sample.

Real sounds look much more like curved lines than straight lines, and a curved line is even more a digital impossibility than a straight oblique. Music for instance is composed of thousands of notes and harmonics mixed together, quite a challenge for a digital converter to redraw when it is only allowed to draw verticals and horizontals — zeroes and ones.

That's the reason why 12 bit recordings (the old standard of the 80's) sound so metallic : the "steps" are simply not numerous enough to give a satisfying illusion. Unless you can't do otherwise, unless your life depends on it, avoid using this option on your DVX.
 
Re: DAT versus dvx audio

[quote author=Anartiste link=board=Links;num=1080751160;start=25#28 date=04/05/04 at 18:16:33]
So we see that the 96KHz issue is not a really important one. Just a theoretical plus.[/quote]
Far more than theoretical. Many recent studies have shown that people can definitely be "affected by" frequencies that they can't "hear." People can tell the difference (well, some people) between 96Khz and 48KHz... parts of the body other than the classical hearing organs are involved in ways that aren't completely inderstood yet.

More practically, however, once you start processing and combining signals, the extra precision in frequency specification afforded by a higher sample rate can make quite a difference, too.


What's important though is that when you get 96KHz, your sound is now analyzed in 24 bits, not 16 bits anymore.
Click to expand...

Sample rate (Hz) and sample size (bits) are unrelated... you can record 96 KHz at 16, 24, 8, or 3 bits... 44.1/24 and 48/24 are pretty common.



As a side note, for the discussion of aging ears - yes, most ears degrade with age; but when testing MP3 compression way back in the 90's, researchers discovered that in a small percentage of people, hearing continues to improve throughout life, and become more discerning of frequencies above 22 KHz or so... (they couldn't stand to listen to the MP3's!)

-Barry
 
The 2 bit post

The 2 bit post

Okay, I confess I'm a bit :D fed up of all these KHz debates, so I sometimes jump a bit (that makes 2 bits now...) hastily on certain conclusions.

I maintain though that this KHz issue is a mainly theoretical debate "in real life circumstances", at least far less concrete than the accuracy of a mix, the proper acoustic rendition of a room or the artistic choice of an echo, between other things. To hear the difference between 48KHz and 96KHz, you'll need a silent place, a very good PA system, top end monitors and, most of all, perfect ears —or any other organ capable of audition.

Resolution, to me, seems a more crucial issue, as the more smoothness we get in analysis, the less "digital" the sound will sound (see what I mean by "digital" here, don't you ?  ;))

You're absolutely right and I'm a moron : yes, you can digitize 96KHz in 3 bits if you want, but the thing is that 24 bits and 96KHz are delivered in the same package by manufacturers. Unless my memory (which, alas, is not 24 bits) betrays me one more time, I've always seen them paired like marketing siamese brothers.

I fully share your opinion when you say that there are still mysteries in the way we percieve our environment. Psychoacoustic is something I believe in. Daniel Lanois says that the obsession of sound engineers and producers for high frequencies has made music production look like a clinic, something less organic, less true to our experience of real life. I agree with this idea : some limited technologies, although less "realistic", provide a more true to life feeling, a paradox we DVXusers experience everyday in preferring 24/25 progressive images/sec to 50/60 interlaced images/sec. And I guess there are some of us who still think a vinyl record has more musical qualities than a CD, or that a Marschall amp will always sound better than any preamp plug-in. Precisely because of (thanks to) their limitations.

I'm not far from having this opinion. It's not an absolute belief, merely a psychoacoustic statement.

And I can't stop thinking that, while the available dynamic has technically doubled in the last twenty years, it has nearly been divided by two in the real world, due to the increasing dictature of compressors at all stages of the production line : recording/mixing/mastering/broadcasting.

Fear of silence ?
 
Re: DAT versus dvx audio

Smoothness... smoothness...

I'm not sure why you attribute smoothness to sample size but not sample rate. Sample rate contributes just as much...

When you talk about things sounding "clinical" or "digital", I think PART OF what you're talking about are some of the effects i was talking about. Now, in the 44.1-48K world, people started moving back towards tube equipment in the analog parts of their signal chains to counteract this. To some extent, people have felt less of a need to do so when sampling at higher frequencies. "The Warmth is Back!", some say.

I haven't worked at 192KHz myself. I've done enough recording and testing at 88.2KHz and 98KHz to hear a difference, especially on quiet, close sources such as acoustic guitar; or in nice concert halls on quieter passages.

As far as compression... all I can say is, listen to some classical music once in a while?

Anartiste, I agree in general with most of what you say, despite any technical little quibbles. "Precise", "accurate", and "good" are three different things, and I certainly prioritize "good" for (most) musical sound.
 
Re: DAT versus dvx audio

I'm sorry if this topic is getting out of its original tracks because of the two of us arguing on their common love of sound. But this is an argument driven by passion, so I hope everybody will forgive this topic hijacking — as well as my poor English that reaches its limits very quickly when talking technical subjects.

Dear Ullanta, I don't want to deny the technological progress in the frequency/resolution quest. I guess you're right concerning the frequency issue — I don't want to nibble and quibble about that and will leave you any final word you want in the technical part of this dialogue. If we could have 1MHz/128 bit recordings, this would be fine for me.

When I say that dynamic has been divided by two during the last twenty years, I am mostly pointing mainstream culture (pop music and films broadcasted on FM and TV). Of course, music genres with large scale dynamics like classical music, world music and jazz have taken a great advantage of this new fidelity in recording and as an amateur, I appreciate that every day. But I think it's a charming paradox we must talk about old genres to pop up the advantages of new technology.

Most of what we hear today is not betrayed by technology, even if it's not totally satisfying yet, but by human nature (the dark side of it). My paraphrasing Daniel Lanois was more intended to outline the fact that, while technology has been giving us more frequency range, we (musicians/sound engineers/producers) started to become more and more obsessed by the high frequencies, unfortunately too often at the detriment of general balance. Just like the excessive use of compressors, this drift is due to human lack of taste or — worse —  human greed, not to technological failures.

The best part of analog technology is, to my opinion, yet unchallenged by digital technology, and the opposite is true. They're copmplementary. Our challenge is to use both for what they can offer at their best and take away our partial points of view. Quite a tough challenge indeed !

In a time when we easily get more concerned about the potentialities of the machines that we are constantly tempted to buy than by the content they're supposed to capture, I think it's wise to remain technologically limited so that our psyche can give the best of its creativity.

Our fathers used to show off with their cars, we do the same with our computers. The pathology has just changed its shape, not its nature.
 
Re: DAT versus dvx audio

I've mixed live bands/events for a while, the moment some dude walks up with a camera in one hand and a cable in the other my pre-canned speech comes out: "You're gonna get an aux, you can't touch the feed/mix, I take no adjustment requests after sound check".

It's an inside joke for live sound engineers because you're going to get the worst output remaining (after all live components have been patched). Most of the feeds remaining usually (depending on sound reinforcement gear) are pre eq also so it's like "Man, I can only imagine what that crap is going to sound like" as you walk off and I giggle (actually, I would have noted the dvx and would have taken more care  ;)). And levels depend on many environment variables for example the comedian in your case has some skit he does with an amp on stage that's way to loud so stage volume is used and you end up with massive vocals on your video with little or no music (except for bleed). And if he's paying you to tape it, you're obviously screwed because it's gone forever.

And the bottom line is that if the house or the talent hasn't specifically instructed the sound engineer to take care of you (paid for two independent mixes) it will be bad far more times than not. I recently had a situation where I know the guy was being overpaid to video an event and he pulled an aux mix from 19 pre-eq inputs, the show was behind schedule so he "got what he got", I guarantee it was unusable when they went to edit.

My 2 cents from that angle, no matter what, always take care of you own sound source if at all possible. The house sound guy is going to be a jerk:) It's actually a noted point of professionalism when I see a video crew shooting an event that I'm mixing and I've not been approached at all (their own audio). You get far more crowd and ambience also flying you're own audio.

Hope that's helpful (I know it's off topic a bit but it's in reference to your specifically mentioned scenario).
 
Re: DAT versus dvx audio

[quote author=nullphonic link=board=Links;num=1080751160;start=25#33 date=04/08/04 at 22:35:19](I know it's off topic a bit but it's in reference to your specifically mentioned scenario).[/quote]

Actually, I think you're the only one on-topic... close to topic?
 
Re: DAT versus dvx audio

[quote author=Anartiste link=board=Links;num=1080751160;start=25#32 date=04/08/04 at 15:34:21]
When I say that dynamic has been divided by two during the last twenty years, I am mostly pointing mainstream culture (pop music and films broadcasted on FM and TV). Of course, music genres with large scale dynamics like classical music, world music and jazz have taken a great advantage of this new fidelity in recording and as an amateur, I appreciate that every day. But I think it's a charming paradox we must talk about old genres to pop up the advantages of new technology.
[/quote]

I just read an article on preparing broadcast mixes that nicely traces a lot of this process of dynamics-squashing, looking at both technology and market pressures... very interesting, and I'll post a reference as soon as I find it again!

And I don't want a last word, but if I must, I'd speak one of agreement: sound, skill, and artistry in recording must come before technology quibbles. Mic placement, mic choice, room acoustics, and to some extent ;-) the performance will always be more important than any of the differences we're discussing in sampling. So for sound improvement, look to yourself and what you've got until you're sure there's no more you can do, before turning to more/better equipment...
 
You must log in or register to reply here.
Back
Top