wst3
Max Output Level: -55.5 dBFS
- Total Posts : 1979
- Joined: 2003/11/04 10:28:11
- Location: Pottstown, PA 19464
- Status: offline
RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2008/07/19 21:25:31
(permalink)
ORIGINAL: yep
Respectfully, I beg to differ from the differer. and I'm going to have to differ, at least a little, with your differ<G>! ORIGINAL: yep
Certainly a device which simply used a capacitor to achieve a nominal impedance at a certain frequency would be an example of clearly dishonest "gaming the system."
But I think we can still talk intelligently about high-impedance vs low-impedance inputs and we can even use nominal, generic numbers to do so.
And this is where I respectfully disagree. I've been a gear end-user nearly all my life, I spent several years designing and testing gear, and these days I assemble gear into systems for clients. In every case the more I knew the better. We can not really talk intelligently about impedance using generic figures. In fact generic numbers are at the root of the problem, even if they were used innocuously. There is so much we do not know about audio... why do two amplifiers with identical specifications (measured carefully no less, with the same test equipment in the same lab) sound different? And that's just the start<G>! As an end user I still use my ears first... if I like the sound of something I'm not likely to care a whit about the specifications. I'll probably still measure it when I get home, because I am curious, but that's it. As a device designer I live and die by specifications. My client expects me to meet or exceed the specifications we agree to during the initial product design. As a system designer I live and die by specifications again, I EXPECT the source to be able to drive the input, and if they get either end wrong I have problems! Then I send it back, and get something that does work! Same goes for the handful of idiots that still haven't figured out how to terminate the shield!!!!! So, qualifying device specifications can only be good for everyone. ORIGINAL: yep
Vague, imperfect, and even outright fraudulent freq charts and equipment specs are nothing new, and there is no standard by which these things are tested (especially something as variable as impedance). Moreover, the specifics of how impedance changes over time are basically the entirety of the "sound" of any preamp input stage. To demand that they all meet certain spec at all frequencies is not only unnecessary but counter-productive to the purpose of trying to encourage better quality devices. Moreover, the testing required to put out a comprehensive and accurate impedance chart is expensive, time-consuming, and meaningless unless all other manufacturers are using the same method. I dislike vague and imperfect specs, and I really dislike fraud. No, it isn't new, but that does not make it right. There are standards, and we continue to work towards better standards every year. Entire committees at AES, IEEE, and other bodies are trying to understand how to measure things and why. And while that initial purchase might be a bit dear for most of us, an AP System Two will make any measurement you can imagine is seconds... so will their Portable One for that matter, and it's a lot cheaper<G>! ORIGINAL: yep
Frankly, even among high-end audio manufacturers it is not uncommon to "predict" (or just guess) the specs based on the design of the device rather than perform actual batch testing. If actual testing is done, is the test data the result of a randomized sample off the line, or of the first one the guy made in his garage, or of the one that tested best out of 100 they tried? Or are they simply made up? And how would we know and to whom would we complain if the data was any of the above? Believe it or not, many of the above. The target specifications are usually the result of design objectives, calculations based on the final design, and measurements made on prototypes and early production runs. Really ambitious companies do test samples from every run, and some companies actually test every unit they make. If samples are used there are well known, and generally accepted ways to handle that. ORIGINAL: yep
And perhaps more importantly, who is to say that, say, some kinds of frequency-dependent reactance in the lower mids is not a good thing? I think you might be mixing topics here... no one is suggesting that there is a correct or proper specification for anything, frequency response, source or input impedance, distortion, etc. We don't yet know what values for these easy specifications make things sound really cool... so we certainly don't have a list of "right" specs. ORIGINAL: yep
The OP's original point and facts were a very good one. A lot of manufacturers are making instrument inputs that are outright wrong, and that violate standard practice and common decency flagrantly. /quote]
Ah, yes, now we are back on topic<G>...
designing an input with a relatively low, and primarily resistive input impedance and then labeling it as an instrument input is folly. It violates generally accepted practices!
I think we all agree on that<G>!
ORIGINAL: yep
But this fraud is not the same thing as imperfect specs. If we were to test the input impedance of a Sansamp, an API DI, a blackface Twin, and a Mackie mixer, and they were all somewhat different, what would that tell us? And if the Mackie mixer tested "best," (which would not surprise me in the least) would that be useful information?
Again define best! The devices you list would absolutely result in different measurements across the board. What we still need to do is learn to correlate those measurements to how people hear. Which is, of course, the real root of the problem... no two people hear the same... what a mess!
ORIGINAL: yep
I think vague spec is just fine for the most part, as long as the manufacturer is ethical about it. And if the manufacturer is not, then it doesn't really matter how detailed their lies are.
Well then you are in good shape, because vague specs are part of the landscape these days. I think manufacturers should provide complete and accurate specifications. Fortunately for me, in my day job, when something does not work as advertised we can send it back for a full credit. It isn't that easy in the MI space. And that's too bad!
You are quite right about that last bit... lies damned lies, and specifications!!!
-- Bill Audio Enterprise KB3KJF
|
yep
Max Output Level: -34.5 dBFS
- Total Posts : 4057
- Joined: 2004/01/26 15:21:41
- Location: Hub of the Universe
- Status: offline
RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2008/07/19 22:44:26
(permalink)
ORIGINAL: wst3
ORIGINAL: yep
Frankly, even among high-end audio manufacturers it is not uncommon to "predict" (or just guess) the specs based on the design of the device rather than perform actual batch testing. If actual testing is done, is the test data the result of a randomized sample off the line, or of the first one the guy made in his garage, or of the one that tested best out of 100 they tried? Or are they simply made up? And how would we know and to whom would we complain if the data was any of the above?
Believe it or not, many of the above..... Forgive me, but you're missing my point. My question was a rheotrical one, i.e. that the specs are meaningless, since we don't really know what they mean without knowing the methodology. Similarly, my comment about Mackies testing out "best" was in the same spirit. My point is not that detailed specs are bad, it is that that stuff like impedance specs are not actually useful for end users. We would need a detailed report on what the measurement methods were and where the data came from in order to compare them to any other set of data. You can hand a skilled engineer a random piece of kit and a random set of target specs and he could devise a "test" that would produce or closely approximate those specs pretty easily, without outright lying. The ways to produce the desired specs are so myriad and obvious that they're not even worth describing. It's like looking for a way to get a T-shirt dirty. Perhaps more to the point, there is nothing at all to stop them from outright lying. Where there is no accountability, and no way to compare specs for company or product X to product Y, they are meaningless. Again, I am not arguing against manufacturers putting out more detailed product data, I'm just saying that without a basis for comparison, the data is basically so much marketing spin, or at least indistinguishable from it. If Ivory Soap is 99% "pure," what is the other 1%? And if Dove is 1/4 moisturizer, and if pretty much all soap, including Dove, is mostly compounds made from caustic fat, then what do any of these stats mean? Calling a line input an instrument input is, at best, dishonest. But we don't need a moisture test to tell when we're standing on dry land. Cheers.
|
Shadow of The Wind
Max Output Level: -82 dBFS
- Total Posts : 425
- Joined: 2005/06/09 17:39:20
- Location: Mountain View, CA (German)
- Status: offline
RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2008/07/20 02:37:59
(permalink)
- Today's electronic components make it easy to realize input impedances where the imaginary part is negligible.
- For a real high-Z input, FETs or tubes should be preferred, while for low input impedances bipolar transistors are the better choice. High-Z inputs should be dedicated inputs, not switchable ones.
- Actually, a good electronic circuit is only a few dollars per channel more expensive than a bad one. But it adds up. If it is 10 dollars per channel... You get the picture. You get what you pay for.
- A transformer preserves energy, a resistor dissipates energy. A transformer with a ration of 2:1 transformer will show one quarter of the input impedance at the output and one half of the voltage (and twice the current). If you use a 750 kOhm resistor to increase your input impedance from 250 kOhm to 1MOhm, you only get one quarter of the voltage instead of one half (transformer). Thus, the transformer gives you twice the signal or 6 dB more because it is lossless (i.e. good transformers are a good thing). And it makes a difference on which end you install the transformer / resistor.
- The output impedance of a guitar is complex and it depends on how you adjust the pots. This is one of the reasons why changing the volume changes the sound.
- All high impedance microphones have phantom powered amplifiers inside the microphone and symmetric outputs. I still do not understand why guitars don't.
Wilko
|
wst3
Max Output Level: -55.5 dBFS
- Total Posts : 1979
- Joined: 2003/11/04 10:28:11
- Location: Pottstown, PA 19464
- Status: offline
RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2008/07/20 08:33:57
(permalink)
ORIGINAL: yep
Forgive me, but you're missing my point. My question was a rheotrical one, i.e. that the specs are meaningless, since we don't really know what they mean without knowing the methodology. Similarly, my comment about Mackies testing out "best" was in the same spirit. No, I understood your premise, but I disagreed<g>, and apparently did a pathetic job of explaining! Specifications are not meaningless if they are done well. I can learn a tremendous amount from a fully qualified specification. Your point is that you need to know the method with which they were made, and we agree on that, since the method is part of the "fully qualified", e.g. EIN is meaningless if we don't know how the test fixture is constructed. A good specification will tell you that. ORIGINAL: yep
My point is not that detailed specs are bad, it is that that stuff like impedance specs are not actually useful for end users. We would need a detailed report on what the measurement methods were and where the data came from in order to compare them to any other set of data. You can hand a skilled engineer a random piece of kit and a random set of target specs and he could devise a "test" that would produce or closely approximate those specs pretty easily, without outright lying. The ways to produce the desired specs are so myriad and obvious that they're not even worth describing. It's like looking for a way to get a T-shirt dirty. Perhaps more to the point, there is nothing at all to stop them from outright lying.
See, we agree on the first part, it's the second part that's causing problems. In my world, A/V system design and installation, if a specification is incomplete we probably won't use it. If a specification is wrong we'll send it back. This does prevent most of the pro audio manufacturers from lying. As far as creating a test to produce a desired outcome, this is certainly true, but if they detail the test conditions an equally skilled engineer can detect the "error in their ways", and still make an informed decision. ORIGINAL: yep
Where there is no accountability, and no way to compare specs for company or product X to product Y, they are meaningless.
This comes under the heading of "it depends"<G> In the pro audio space they are absolutely accountable. When a manufacturer provides incomplete, or worse inaccurate specifications word spreads like wildfire in the community. Fortunately there are enough manufacturers out there who care about this that there are plenty of lines to choose from. As an example, the Rane application note #110, "Audio Interconnections" was corrected years ago to reflect new understandings about shield termination and shield current induced noise. (Their designs were also corrected.) Well for some strange reason the text was corrected, but the drawings were not updated. When they caught it they sent out an announcement to all the usual suspects. We all had a good chuckle at their expense, and then it was back to business as usual. They didn't have to broadcast their error, but they are accountable to the marketplace, and so they made sure that the marketplace knew they had corrected the error. ORIGINAL: yep
Again, I am not arguing against manufacturers putting out more detailed product data, I'm just saying that without a basis for comparison, the data is basically so much marketing spin, or at least indistinguishable from it. If Ivory Soap is 99% "pure," what is the other 1%? And if Dove is 1/4 moisturizer, and if pretty much all soap, including Dove, is mostly compounds made from caustic fat, then what do any of these stats mean?
There's a huge difference between 99% pure and "0.01% THD+N/N at 1 kHz at maximum output with a 50 ohm source impedance". This tells me that they drove the system with a 1 kHz sine wave at a level that drove, the system to maximum output. The signal source had a 50 ohm source impedance, so it was higher than most of the gear I own, and thus my own use will probably have less input noise. It also tells me that the measurement was made by removing the excitation signal, so that the actual result includes noise. ORIGINAL: yep
Calling a line input an instrument input is, at best, dishonest. But we don't need a moisture test to tell when we're standing on dry land.
True, but how do we know that it is a line level input and not an instrument input? A line level input can have an input impedance of anywhere from 5K ohms (still 10x the expected source impedance) to several hundred Kohms. It should handle a nominal source level of either -10 dBV or +4 dBu without attenuation. And in this case we really don't care too much about the reactive portion of the input impedance, since the source impedance is expected to be purely resistive. If there is a large reactive component it could actually be problematic, so they should still report it. An instrument input needs to handle signal levels in the range of -20 dBV or lower, and it needs to have a much higher input impedance. The reactive nature of the impedance is an issue. However, and on this we do agree, even if I create an input circuit that accurately mimics the input stage on my favorite blackface twin it isn't going to sound the same. I might still like to know, but of course in order to make use of that information I'd have to know what that twin input stage looks like! So while I don't need a test to know I am standing on dry land, I do still appreciate accurate and complete specifications. Maybe I'm weird?
-- Bill Audio Enterprise KB3KJF
|
wst3
Max Output Level: -55.5 dBFS
- Total Posts : 1979
- Joined: 2003/11/04 10:28:11
- Location: Pottstown, PA 19464
- Status: offline
RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2008/07/20 08:47:33
(permalink)
ORIGINAL: Shadow of The Wind
- Today's electronic components make it easy to realize input impedances where the imaginary part is negligible.
True! They also make it possible to realize an input impedance that looks like whatever the designer wants it to look like. ORIGINAL: Shadow of The Wind
- For a real high-Z input, FETs or tubes should be preferred, while for low input impedances bipolar transistors are the better choice. High-Z inputs should be dedicated inputs, not switchable ones.
Could not agree more with that second part, except that when not in use the High Z input needs to be disconnected and shunted. ORIGINAL: Shadow of The Wind
- Actually, a good electronic circuit is only a few dollars per channel more expensive than a bad one. But it adds up. If it is 10 dollars per channel... You get the picture. You get what you pay for.
Having fought that battle back when I was a designer I'm going to agree, and then keep my opinions to myself<G>. Otherwise I'd put everyone to sleep<G>! ORIGINAL: Shadow of The Wind
- A transformer preserves energy, a resistor dissipates energy. A transformer with a ration of 2:1 transformer will show one quarter of the input impedance at the output and one half of the voltage (and twice the current). If you use a 750 kOhm resistor to increase your input impedance from 250 kOhm to 1MOhm, you only get one quarter of the voltage instead of one half (transformer). Thus, the transformer gives you twice the signal or 6 dB more because it is lossless (i.e. good transformers are a good thing). And it makes a difference on which end you install the transformer / resistor.
All true! ORIGINAL: Shadow of The Wind
- The output impedance of a guitar is complex and it depends on how you adjust the pots. This is one of the reasons why changing the volume changes the sound.
Also true, and why I bypassed the potentiometers on several of my guitars, only to learn that I did in fact use small changes on those controls to create different sounds. DANG! So I had to put them back in the circuit<G>! ORIGINAL: Shadow of The Wind
- All high impedance microphones have phantom powered amplifiers inside the microphone and symmetric outputs. I still do not understand why guitars don't.
You and me both!!! I have built Tillman's little "active cable" and used it for guitars and to tame the output of a couple of really nasty stomp-boxes. But again there is a price to be paid... things don't sound the same! I like the new sound some of the time, but not all of the time. I also spent a fair amount of time building input and output buffers for various pedals (wah-wahs and really old fuzz-boxes are the worst!), but I never quite found the perfect combination where I could preserve the sound I was looking for while improving the S/N ratio and frequency response. In some cases the difference was subtle, but in some cases it turns out that part of the "sound" I was looking for was the result of the interaction between a specific output circuit and the a specific input circuit - e.g. the output of a vox wah driving a fuzz face. The low pass filter created is insane, but yet that's what we are accustomed to hearing because it was part of so many great guitar solos<G>! In an ideal world connecting two pedals would not degrade the signal at all, and there would be an additional control or two to emulate those old interactions. But as Yep points out, we still don't know exactly what that interaction is. We can model it, we can measure it, but without the "golden" pedals we're still shooting blind. And to complicate matters, everone's definition of the golden pedals is different<G>! What a mess!
-- Bill Audio Enterprise KB3KJF
|
yep
Max Output Level: -34.5 dBFS
- Total Posts : 4057
- Joined: 2004/01/26 15:21:41
- Location: Hub of the Universe
- Status: offline
RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2008/07/20 18:46:00
(permalink)
ORIGINAL: wst3
...Specifications are not meaningless if they are done well. I can learn a tremendous amount from a fully qualified specification...
I don't want to beat a dead horse, so I'll drop it after this, and anyone is free to have the last word. Some of the most detailed and comprehensive specs published are outright wrong, in terms of manufactured audio equipment (electrical components are a different thing). This from manufacturers like Shure, Universal Audio, and other studio-standard equipment that I think anyone would consider "professional," but of course YMMV. Moreover if we refuse manufacturers who do not provide detailed impedance specs but only a single round number without frequency, then we have to refuse brands such as Avalon, Grace Designs, and countless others that are absolutely top-shelf professional audio equipment. It would be far easier to list the manufacturers who *do* provide detailed impedance specs than to list those who don't, even among high-end professional equipment. My argument is not that detailed specs are bad, just that they are either unreliable or unavailable so often that only a vanishingly small number of components can be fairly assessed based on them, and moreover, there is no real way to know what that number is, since there is no obvious way to distinguish the good data from the bad, and since a lack of specs in no way implies subpar equipment. Specs on stuff like THD+N is certainly a lot easier and more common to test than impedance, but I think it's safe to say that even in a fair test, many cheap Behringer or Mackie type pieces will beat, say, a Telefunken or Neve input module on these measures. Does that mean that the el cheapo one is better or more professional than the vintage kilobuck preamp? Moreover, I think anyone who has worked in a professional audio environment doesn't need me to tell you that plenty of devices, even reputable ones, are often considerably noisier and more distorted than the spec says when you plug them in. The environment and application are huge factors, and not always predictable ones. I.e. device X that spe'c out better in a test lab may actually perform worse in as many or more real-world applications than device Y which spec'd out worse. Were the tests faulty? Is spotty build quality to blame? Or is it just normal variance when we're talking about such exacting and minute differences? And when you've got to finish a new album every week by a band you've never heard before, what's the difference? Technical design specs are the lifeblood of equipment designers, but they are in the service of the still very trial-and-error process of trying to make better-sounding stuff. For an end-user, trying to select a preamp or DI by reading the impedance specs is like trying to decide which restaurant is going to have the best bread by reading the ingredient list. We may be able to rule out a small handful of obviously bad examples, but the differences between the really good stuff are as numerous as the similarities. Anyhow, having said my piece, I leave it to the rest of you fine folk to settle this once and for all. cheers.
|
The Maillard Reaction
Max Output Level: 0 dBFS
- Total Posts : 31918
- Joined: 2004/07/09 20:02:20
- Status: offline
.
post edited by mister happy - 2017/06/29 00:35:39
|
funtimesman
Max Output Level: -90 dBFS
- Total Posts : 19
- Joined: 2009/09/07 15:52:52
- Status: offline
Re: RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2009/11/07 03:22:57
(permalink)
You guys are great, thanks for your thorough responses. I have a new question, I don't think it is related to the emails between us copied below. Basically, how good is the emu 1820, 1820m and the 1616/m for recordign guitar direct? I have the 1820. the 2 preamps on it are apparently mic/line or maybe mic/line/instrument inputs... Please take a thorough look at this thread (which says emu 1820's only have a 500 kohm impedance, which is half the accepted 1 mohm for Hi-Z, thus they really are not very good for guitar inputs.... http://forum.cakewalk.com/tm.aspx?m=1436553 its a damn interesting thread, but towards the end it just gets theoretical and boring.... basically the 500 kohm on the 1820 (is this the actual spec?) is half what it should be. Is the 1820 better in that regard. what the 1616/m an improvement there. Can one just plug in a guitar to a Boss pedal like the os2 distortion pedal (the thread comments on this: " A band aid for everyone that works!!!  : There are 2 types of pedals : True bypass and Not True bypass. As the names implies, true bypass means that when the FX is OFF, there is a JUST CABLE from input to output. You can recognize those for its switch, like a wha wha or a RAT (Pro co) distortion. Not tru bypass is like any BOSS pedal. When they are off they STILL GO thru circuit. That is why you can not use without a battery!!!! HOLLY SOLUTION: Take ANY pedal (BOSS preferable  ) and put a good power supply. Put it in off (Bypass). Connect your guitar to its input, and its output to any line level input. BINGO!!!!! you have a SUPER DECENT DI with propper 1 mohm input impedance!!! I guaranty you will see the difference. I record my guitar traxs like this thru my EMU1820 (it has a 'fake' hi z input)." FWIW, for this purtpose a simple transistor preamp like those found of boss pedals is just perfect. The sound is goos thru the whole guitar spectrum. the idea describen in this topic is enter the guitar signal into the computer as clena s posible. I tried to plug my guitar in to the line in in the back of the 1820 with a boss os2 distrotion pedal (the yellow one) between my guitar and the dock....i had to engage the pedal, so that the red "check" light came on....and i turned the drive setting to zero so there was no distortion, ie basically a clean tone (so I can keep my options open for later) and that seemed to work really well in terms of getting a hotter signal into my recording software via the 1820....the quote from the thread above seems to think i would not need to stop on the pedal and get the red "check"light (which to me means you have stomped into the distortion and out of the clean area) but I find if I click the pedal out of the distortion zone, the signal is as weak as if there was no pedal (ie, I click onto the distortion part of the pedal, and turn all the distortion setting so low that the tone is essentially clean, but i think the pedal acts as a preamp or di or something when i do that, thus making my emu have an acceptable hi-z input, in the line in in the back (ie not using the preamps in the front of the 1820) no less. Would it be even better to utilize this boss pedal with the preamps on the front of the emu? Thanks, for you help and insight. also, what about this thing: "ank you all for your interest and kind replies! A couple of things useful for all 'Direct Recorders' out there: 1) Build a preamp. http://www.till.com/articles/PreampCable/index.html They explicitly say that 1mohm is the minimum acceptable. This (or similar things) are actually a BUFFER: an preamp with 0dB of gain, but adapt impedances: In -->1mohm. Out--> 10Kohm (like a line out level). This leads to 2) Sincerely, Me
|
bullo
Max Output Level: -89 dBFS
- Total Posts : 57
- Joined: 2008/07/17 01:07:45
- Status: offline
Re: RE: Hi-z input (instrument in) on various interfaces: Lies from manufacturers
2009/11/07 21:26:27
(permalink)
hi.
I also have an emu 182o in my desktop. I made test by running the guitar directly into the mic/inst mic and passing thru a boss metal zone in bypass mode.
yes, the boss in bypass is much better. has a better freq response and less noise , due to its 'real' 1 mohm input impedance. in mot sure i understood you question, but you should use those pedals in bypass, meaning 'without the red light on'. In bypass, the signal still goes thru its input thru output so it effectively goes thru the desired 1 mohm input impedance. In this case that is what we want, but if you chain 10 pedals, that would deteriorate the signal because every epdal in bypass goes thru its in-->out circuit. that is not the case with true bypass pedals
Danny
PS: the 1820 is an excelent interface. I only had a few problem with 'all light one', where the interface creashes randomly. check taht in the emy forum
|