If I were you I'd be worried about someone discovering a weapon that can be built with common materials, portable and powerful enough to destroy a country.
Yeah, if someone beats me to it what would I do with my weekends?
10.5.8 runs on a G4 out of the box, the current version (10.6.3) is Intel only. And the G4 iMac wasn't introduced until 2002, so none of your story stacks up.
With an H910, the pitch shift is constant as set by a knob. A glissando remains a glissando, and amateur's wobbly voice remains just as wobbly, just changed in frequency.
Which is why it's so hard to detect vocals that were shifted using one. It doesn't have the tell-tale impossible snap to note; all variations are preserved. The only give away is munchkinisation from excessive shift ranges.
It'll put you off key in a different key.
Or a fraction of a key for as long as the effect isn't bypassed, which is how they were used to correct pitch: if the second note in the third bar of the second verse was flat, set the machine to shift up by the same amount and hit the bypass button at exactly the right time while recording to another track. Yes, I conceded the H910 isn't automatic, but it may surprise you to learn that there was a time when faders didn't record moves and editing was done by cutting tape on a metal block with a razor blade.
In particular, an H910 is useless for pitch correction when singing live since it does a constant shift.
AutoTune is nearly useless live too, since it tends to lock on to the lowest fundamental frequency it can detect, which usually isn't the vocal unless the vocal mic is completely isolated (which means pretty much never). In those situations the pitch correction is keyed by a sequencer track, so it isn't automatic either; you're simply replacing someone twiddling a knob live on a H910 with a recording device playing back pre-recorded knob twiddling.
And AutoTune can't fix bad timing or lack of breath control, so truly bad singers still mime (hello Britney).
Autotuners let any idiot sing in tune since they pull your voice sharp or flat as necessary.
Provided any idiot is less than half a semitone out (or tone if the key has been set), more than that and they snap to the wrong note. They don't work by magically divining the note you intended; if you can't carry a tune in a bucket, autotune won't help automatically.
Your glib "Wrong" post response with a 1975 effects device was misleading bullshit.
It's as much bullshit as saying that a third of a decade didn't exist, how obviously a technique is used is related to when it was first introduced, or that pitch correction being automatic or manual somehow makes a philosophical difference (either way, it's still corrected), so I think we're even.
I assume, though, that it couldn't have been better than early digital autotuning, or they wouldn't have switched?
Like much studio gear, price is the critical factor here. The basic Autotune software is around 5% the price of an Eventide ($5,500+), and you can't easily pirate hardware. The same can be said for LA-4A compressors and Pultec EQs...rare and expensive in meatspace, today anyone with ProTools can get the Bombfactory plugin equivalent for a fraction of the price.
Maybe you could shift the pitch a little with this thing, like compressing a few seconds off a song and preserving the pitch
Early Eventides shifted pitch by (short version) writing data to a memory buffer at one clock speed and reading it out at another in real time. This meant they could achieve up to a full octave shift using less than 16K RAM.
Modern autotune could take droning and make it sound about like nickleback, with little or no effort from the studio technician.
To do that you have to program the notes into a sequencer, which isn't difficult, but neither is it little or no effort. But that isn't "auto" tune, it's manual tune: a human is still required to enter the pitch data, which only differs from moving a dial in precision and convenience.
And there's a limit to how far you can shift pitch, particularly vocals, before it starts sounding unnatural.
This changed singing style dramatically, which was my point.
Microphones changed singing style dramatically; opera fans ridiculed the crooners who were inaudible without one. Now they're standard, even in opera.
Compressors changed singing style dramatically because they allowed singers to get the vocal qualities of both whispering and shouting without having to worry about levels or mic technique. Now they're standard.
Distortion pedals changed guitar style dramatically with power chords and hammer solos. Now they're standard.
What I find interesting is that nobody has satisfactorily explained to me why distortion is acceptable when autotune isn't without using the same arguments that were used against distortion or any of the other technological aids in their day. It seems to me to be a completely arbitrary distinction, more concerned with methods than results. And though they're not my cup of tea, I'd argue that in the context of heavily synthesised music like techno and R&B the synthetic voice quality is quite fitting...could it be possible that the kids like the autotune sound precisely because their elders don't and it's a defining feature of the music of their era, and the objections are simply history repeating?
The H910 had to be switched in and out manually to be used as a pitch corrector, so no, it wasn't autotuning. It was, however, still used to artificially manipulate vocals to the correct pitch, so the distinction is operation method, not principle.
Autotuning was a function on the 1987 Eventide H3000 (IIRC), so the statement that there was no such thing in the 80's is still incorrect.
Is it because Sci-Fi works well for a feature length movie but not for a serial production show with a much smaller per-hour budget?
That's not how it works: you can't amortise your costs across the entire length of a production run until you know how long that is, which in turn is determined by ratings. In other words, if you spent $20M on props and sets and your show only lasts six episodes, you're looking at >$3.33M an episode before running costs, but if it runs four complete seasons these fixed costs are only $384,615 per episode...the problem is there's no guarantee you'll last four seasons. Commissioning editors only look at the first number ($20M), which is why they run screaming for the hills when the syllables "sci" and "fi" are used consecutively in their presence.
A movie might be more expensive, but it has a number of advantages over TV. First, TV is advertising supported, while a movie has direct revenue streams from tickets, DVDs and merchandise, so there's greater potential to make back a large investment quickly. Second, a movie doesn't have to fit schedule demographics the way a TV show does, since it doesn't need to sell advertising slots. Third, a movie doesn't rely on drawing in the same crowd week after week, so it can afford to suck (note to Michael Bay: just because it can doesn't mean it must).
Of course, the big screen is better when it comes to special effects driven claptrap, but without the sensory overload you have to rely on plot, which is also a problem for television.
I suspect that is why he said something like a DI box.
All dogs have four legs, and my cat has four legs, but that doesn't mean my cat is a dog. An attenuator that doesn't convert from unbalanced to balanced is nothing like a DI box, it's just an attenuator.
The idea is the conversion from line level to mic level, not the unbalanced to balanced conversion.
Yes, that's what I said, and that's why a DI box is the wrong choice.
But by using a voltage 1:1 balun on the output of a DI box, you should in theory be able to get the desired unbalanced line level signal you need.
A balun is unnecessary, since the output impedance of a DI is lower than the impedance of an electret mic input; depending on whether it's active or passive DI, you could either connect pin 2 instead of pin 3 to input ground, or just leave pin 2 unconnected (shunting to ground through a 10k resistor would be better). Either way, it's still more trouble than just using the right device (an attenuator) in the first place.
Unlike with pro audio, in consumer audio you want to deal with both channels in every device, requiring either 2 DI boxes, or one box with dual channel support. The costs for that would be absurd.
A stereo Behringer DI-120 is about Au$40 (and it has switchable 0/20/40dB attenuation). But it's still not the right tool for the job.
Considering the quality of standard computer audio components, you could probably get away with using a dual op-amp chip and a could of resistors to design a simple amplifier with a gain of less than one.
It could be done with a simple two resistor voltage divider and DC blocking capacitor per channel for less than a dollar. An op-amp based circuit is unnecessarily complex, won't sound any better, and with a PSU it'd cost about as much as the basic USB interface I mentioned.
Sensitivity is to power what speed is to velocity: similar, but not the same.
Input sensitivity, or RMS input voltage for full scale output, is useful when the circuit gain is fixed (or has some calibrated nominal value). So yes, for most consumer gear with standard input types (including standard sound cards and hi-fi phonogram or tape deck inputs) sensitivity is correct enough to get the job done.
For mixing desks and associated equipment input sensitivity is an indeterminate variable, so an alternative reference is required, and that is best expressed in relation to power into a given load since it allows easy mapping of gain structures in a signal chain*. Impedance matching transformers may seem archaic, but they're still quite common in high end gear, and as the 600 Ohm balanced line is a long established standard 0dBm and 0dBu are the same to allow compatibility with solid state I/O (which for most practical purposes can be said to be immune to loading effects).
*Decibels are a measure of power rather than amplitude; it's technically incorrect to measure voltage-derived scales in dB, but it means you can add and subtract rather than use long multiplication and division, so only electronic engineers who want to share the misery give a toss.
Three characteristics matter, actually: phantom power, impedance (Z) and sensitivity.
Most PC mics are electret (condensor) types, which typically require a low power DC supply to operate. Fine for a mic, but when you try to connect a line in the DC offset is superimposed on the signal, which can cause half-wave clipping if the output of the device feeding it isn't properly isolated. Most line outputs are DC isolated with capacitors, so this generally isn't a problem, but you can never be sure.
However, the problem with using capacitors to block DC is that they create a filter. If the source impedance doesn't match the load (such as when you plug a high-Z line out into a low-Z mic in), this filter will be audible.
Mic inputs tend to have 20-40dB more gain than line level inputs, and this is indeed referred to as sensitivity (no, it isn't power; see my other comment in the thread).
Not quite correct: 0dBm is defined as 1mW into a 600 Ohm load, and this has long been a standard where transformer balanced gear is used. For unbalanced equipment with unpredictable input and output impedances, dbV (1.2V) or dBu (0.775V) are used instead.
No, a DI boxes are for driving balanced lines, and while this may be a rash generalisation, I don't think any computers have XLR inputs on the built-in sound card. Besides, not all DI boxes have an attenuator (pad), which is what's needed here.
Assuming there isn't a system setting that allows gain switching on the input (quite possible), simpler and probably cheaper would be a basic interface like the Behringer UCA200.
Slashdot Mirror
If I were you I'd be worried about someone discovering a weapon that can be built with common materials, portable and powerful enough to destroy a country.
Yeah, if someone beats me to it what would I do with my weekends?
I'm sure there's some kind of inverse sexual harassment thing going on there.
"She constantly gives graphic descriptions of her genitals and demands we feel her breasts. Oh, and she's a coprophile".
Compared to all the hokum on the net, "unlikely" is gold standard proof.
Pretty much every TSA employee I have ever met was a cock.
But not in a big way...
10.5.8 runs on a G4 out of the box, the current version (10.6.3) is Intel only. And the G4 iMac wasn't introduced until 2002, so none of your story stacks up.
No problem, I just patented the superconducting looseleaf binder.
The pair pictured aren't farmers, they're erosion enthusiasts.
Not likely.
With enough beer that's swimming training.
...couldn't that space in the CPU be used for better things than a redundant graphics circuit?
At first I read that as "retarded graphics circuit". Still made perfect sense...
That's what the roll of carpet and shovel is for...
With an H910, the pitch shift is constant as set by a knob. A glissando remains a glissando, and amateur's wobbly voice remains just as wobbly, just changed in frequency.
Which is why it's so hard to detect vocals that were shifted using one. It doesn't have the tell-tale impossible snap to note; all variations are preserved. The only give away is munchkinisation from excessive shift ranges.
It'll put you off key in a different key.
Or a fraction of a key for as long as the effect isn't bypassed, which is how they were used to correct pitch: if the second note in the third bar of the second verse was flat, set the machine to shift up by the same amount and hit the bypass button at exactly the right time while recording to another track. Yes, I conceded the H910 isn't automatic, but it may surprise you to learn that there was a time when faders didn't record moves and editing was done by cutting tape on a metal block with a razor blade.
In particular, an H910 is useless for pitch correction when singing live since it does a constant shift.
AutoTune is nearly useless live too, since it tends to lock on to the lowest fundamental frequency it can detect, which usually isn't the vocal unless the vocal mic is completely isolated (which means pretty much never). In those situations the pitch correction is keyed by a sequencer track, so it isn't automatic either; you're simply replacing someone twiddling a knob live on a H910 with a recording device playing back pre-recorded knob twiddling.
And AutoTune can't fix bad timing or lack of breath control, so truly bad singers still mime (hello Britney).
Autotuners let any idiot sing in tune since they pull your voice sharp or flat as necessary.
Provided any idiot is less than half a semitone out (or tone if the key has been set), more than that and they snap to the wrong note. They don't work by magically divining the note you intended; if you can't carry a tune in a bucket, autotune won't help automatically.
Your glib "Wrong" post response with a 1975 effects device was misleading bullshit.
It's as much bullshit as saying that a third of a decade didn't exist, how obviously a technique is used is related to when it was first introduced, or that pitch correction being automatic or manual somehow makes a philosophical difference (either way, it's still corrected), so I think we're even.
I assume, though, that it couldn't have been better than early digital autotuning, or they wouldn't have switched?
Like much studio gear, price is the critical factor here. The basic Autotune software is around 5% the price of an Eventide ($5,500+), and you can't easily pirate hardware. The same can be said for LA-4A compressors and Pultec EQs...rare and expensive in meatspace, today anyone with ProTools can get the Bombfactory plugin equivalent for a fraction of the price.
Maybe you could shift the pitch a little with this thing, like compressing a few seconds off a song and preserving the pitch
Early Eventides shifted pitch by (short version) writing data to a memory buffer at one clock speed and reading it out at another in real time. This meant they could achieve up to a full octave shift using less than 16K RAM.
Modern autotune could take droning and make it sound about like nickleback, with little or no effort from the studio technician.
To do that you have to program the notes into a sequencer, which isn't difficult, but neither is it little or no effort. But that isn't "auto" tune, it's manual tune: a human is still required to enter the pitch data, which only differs from moving a dial in precision and convenience.
And there's a limit to how far you can shift pitch, particularly vocals, before it starts sounding unnatural.
This changed singing style dramatically, which was my point.
Microphones changed singing style dramatically; opera fans ridiculed the crooners who were inaudible without one. Now they're standard, even in opera.
Compressors changed singing style dramatically because they allowed singers to get the vocal qualities of both whispering and shouting without having to worry about levels or mic technique. Now they're standard.
Distortion pedals changed guitar style dramatically with power chords and hammer solos. Now they're standard.
What I find interesting is that nobody has satisfactorily explained to me why distortion is acceptable when autotune isn't without using the same arguments that were used against distortion or any of the other technological aids in their day. It seems to me to be a completely arbitrary distinction, more concerned with methods than results. And though they're not my cup of tea, I'd argue that in the context of heavily synthesised music like techno and R&B the synthetic voice quality is quite fitting...could it be possible that the kids like the autotune sound precisely because their elders don't and it's a defining feature of the music of their era, and the objections are simply history repeating?
The H910 had to be switched in and out manually to be used as a pitch corrector, so no, it wasn't autotuning. It was, however, still used to artificially manipulate vocals to the correct pitch, so the distinction is operation method, not principle.
Autotuning was a function on the 1987 Eventide H3000 (IIRC), so the statement that there was no such thing in the 80's is still incorrect.
The eighties did not have autotune.
Wrong.
Lady Gaga as a drunken sailor
That would be Amy Winehouse.
Is it because Sci-Fi works well for a feature length movie but not for a serial production show with a much smaller per-hour budget?
That's not how it works: you can't amortise your costs across the entire length of a production run until you know how long that is, which in turn is determined by ratings. In other words, if you spent $20M on props and sets and your show only lasts six episodes, you're looking at >$3.33M an episode before running costs, but if it runs four complete seasons these fixed costs are only $384,615 per episode...the problem is there's no guarantee you'll last four seasons. Commissioning editors only look at the first number ($20M), which is why they run screaming for the hills when the syllables "sci" and "fi" are used consecutively in their presence.
A movie might be more expensive, but it has a number of advantages over TV. First, TV is advertising supported, while a movie has direct revenue streams from tickets, DVDs and merchandise, so there's greater potential to make back a large investment quickly. Second, a movie doesn't have to fit schedule demographics the way a TV show does, since it doesn't need to sell advertising slots. Third, a movie doesn't rely on drawing in the same crowd week after week, so it can afford to suck (note to Michael Bay: just because it can doesn't mean it must).
Of course, the big screen is better when it comes to special effects driven claptrap, but without the sensory overload you have to rely on plot, which is also a problem for television.
Besides, glasses make people look more intelligent.
Nobel prize material right here.
Sorry. I've been meaning to get onto that, but these peasants with pitchforks keep burning down my castle...
I suspect that is why he said something like a DI box.
All dogs have four legs, and my cat has four legs, but that doesn't mean my cat is a dog. An attenuator that doesn't convert from unbalanced to balanced is nothing like a DI box, it's just an attenuator.
The idea is the conversion from line level to mic level, not the unbalanced to balanced conversion.
Yes, that's what I said, and that's why a DI box is the wrong choice.
But by using a voltage 1:1 balun on the output of a DI box, you should in theory be able to get the desired unbalanced line level signal you need.
A balun is unnecessary, since the output impedance of a DI is lower than the impedance of an electret mic input; depending on whether it's active or passive DI, you could either connect pin 2 instead of pin 3 to input ground, or just leave pin 2 unconnected (shunting to ground through a 10k resistor would be better). Either way, it's still more trouble than just using the right device (an attenuator) in the first place.
Unlike with pro audio, in consumer audio you want to deal with both channels in every device, requiring either 2 DI boxes, or one box with dual channel support. The costs for that would be absurd.
A stereo Behringer DI-120 is about Au$40 (and it has switchable 0/20/40dB attenuation). But it's still not the right tool for the job.
Considering the quality of standard computer audio components, you could probably get away with using a dual op-amp chip and a could of resistors to design a simple amplifier with a gain of less than one.
It could be done with a simple two resistor voltage divider and DC blocking capacitor per channel for less than a dollar. An op-amp based circuit is unnecessarily complex, won't sound any better, and with a PSU it'd cost about as much as the basic USB interface I mentioned.
Sensitivity is to power what speed is to velocity: similar, but not the same.
Input sensitivity, or RMS input voltage for full scale output, is useful when the circuit gain is fixed (or has some calibrated nominal value). So yes, for most consumer gear with standard input types (including standard sound cards and hi-fi phonogram or tape deck inputs) sensitivity is correct enough to get the job done.
For mixing desks and associated equipment input sensitivity is an indeterminate variable, so an alternative reference is required, and that is best expressed in relation to power into a given load since it allows easy mapping of gain structures in a signal chain*. Impedance matching transformers may seem archaic, but they're still quite common in high end gear, and as the 600 Ohm balanced line is a long established standard 0dBm and 0dBu are the same to allow compatibility with solid state I/O (which for most practical purposes can be said to be immune to loading effects).
*Decibels are a measure of power rather than amplitude; it's technically incorrect to measure voltage-derived scales in dB, but it means you can add and subtract rather than use long multiplication and division, so only electronic engineers who want to share the misery give a toss.
Three characteristics matter, actually: phantom power, impedance (Z) and sensitivity.
Most PC mics are electret (condensor) types, which typically require a low power DC supply to operate. Fine for a mic, but when you try to connect a line in the DC offset is superimposed on the signal, which can cause half-wave clipping if the output of the device feeding it isn't properly isolated. Most line outputs are DC isolated with capacitors, so this generally isn't a problem, but you can never be sure.
However, the problem with using capacitors to block DC is that they create a filter. If the source impedance doesn't match the load (such as when you plug a high-Z line out into a low-Z mic in), this filter will be audible.
Mic inputs tend to have 20-40dB more gain than line level inputs, and this is indeed referred to as sensitivity (no, it isn't power; see my other comment in the thread).
Not quite correct: 0dBm is defined as 1mW into a 600 Ohm load, and this has long been a standard where transformer balanced gear is used. For unbalanced equipment with unpredictable input and output impedances, dbV (1.2V) or dBu (0.775V) are used instead.
Ironically, if you wanted to record the sound from an acoustic instrument, you would use a microphone and the mic-in socket.
Only if you hate your ears and believe they should be punished.
No, a DI boxes are for driving balanced lines, and while this may be a rash generalisation, I don't think any computers have XLR inputs on the built-in sound card. Besides, not all DI boxes have an attenuator (pad), which is what's needed here.
Assuming there isn't a system setting that allows gain switching on the input (quite possible), simpler and probably cheaper would be a basic interface like the Behringer UCA200.