What are the ACTUAL specs for the 828 MKIII?

[thashobs]

First off I am posting this to match up setting in the eq for vocal micing technics with premium mics such as the U87 or tlm103.
(mics in the $1000 and up range.) The specs are important to know when using mics of this caliber. For instance the Neumann U87 frequency response is 20hz-20Khz with a dynamic range of 105db. Now when plugging this mic in to the 828 what parts of the audio spectrum am I losing due to differences in the in the gear. say if the 828 had only a 90db dynamic range and frquency response of 20-18khz I'd be losing subtle details picked up by the mic which I would need to compensate for with EQ.

If anybody knows the specs I would like to know them ie....

Headroom
Frequency response
THD
dynamic range
DSP speed ???MHz
Signal to noise ratio
input/output impediance
Mic pre specs


How am I supposed to push my gear to the limit without knowing these specs



please keep the question on the gear page. I'm not complaining about any kind of MOTU "policy" to keep their tech specs and industry secret.
I'm demanding that the information that Audio Engineers like myself would like to know.

look at mackies fullblown tech sheet for their competing interface.
How does the 828 stack up?

How much better compared to the mkII?
what are the specs for the MKII?

why isn't there a white paper or spec sheet?
http://www.mackie.com/products/400f/specs.html

AND the Neve portico mic pre specs
http://rupertneve.com/products/portico-5032/specs/

I can find no such data for my brand new 828MKIII

The 100 hundred page manual? nope no specs.

[conleycd]

I don't think specs are going to help you on much of anything. MOTU interfaces uses the same AD conversion chips (as do a bunch of interfaces) as Digidesign interfaces.

Specs are often measured in different ways so comparing one piece of gear to another is sometimes like comparing apples to oranges.

MOTU interfaces are as decent as most what you would find in RME, M-Audio, Digidesign and other like devices.

The frequency spectrum is likely stable to 20khz but would also extend beyond that as well. I'm not sure what the stability would be like +/- 1db or +/- 3db beyond. I don't think any human can really hear 20khz.

We may perceive the ways in which frequencies above 16khz distort. This may impact our perception about sound a bit.

Eitherway, the frequency specs of your MOTU interface do not impact your choice of microphone. While splitting hairs about this kind of stuff... the capacitance in your microphone cable likely has more impact than your MOTU guts.

You would probably be wise to consider a good quality microphone preamp which is responsible for the amplification of your microphone hundreds of times over before it his the A/D of your interface.

All that to say... I don't think MOTU has ever published their specs. It is good stuff to use.

If Tom Lord-Alge can mix in 44/16 bits, make great mixes, and win grammies - you should be ok with a system that can do 192/24 (although I doubt you'd hear any different at 192 vs. 44.1khz).

CC

[monkey man]

192 vs 44.1? I reckon many folks would be able to tell the difference.
For one thing, thousands of engineers can tell the difference between 48 and 44k, going way back to the early days of DAT.

On the spec issue, there's no hope IMHO; the emperor has pronounced access to said information verboeten. LOL

[conleycd]

192 vs 44.1? I reckon many folks would be able to tell the difference.
For one thing, thousands of engineers can tell the difference between 48 and 44k, going way back to the early days of DAT.

On the spec issue, there's no hope IMHO; the emperor has pronounced access to said information verboeten. LOL

Well, I did read another thread saying that MOTU would disclose the info over the phone. I don't think it is important.

I still don't agree that people can hear the difference between 192khz and most other sample rates. Human hearing doesn't extend much beyond 18khz. Roger Nichols wrote about one engineer who heard a Neve channel strip that had been capped at some frequency (i.e. 100khz) compared to the other channels by error (150khz). That fellow reportedly heard a difference when no one else did. The only logical argument I have read for "sounding different" is that beyond the hearing spectrum frequencies distort in a manner that our brains can interpret.

I think much above 48khz is a waste of hard drive space. You also introduce "digital weirdness" when you record at high sample rates and then dither down to 44.1khz.

Again, if Tom Lord-Alge can produce the kick ass mixes he does with a 16 bit ancient Sony digital tape to tape at 44.1khz (he converts all Pro Tools formats down into 16/44.1 format) - then we should all focus more on our mix skills and good equipment than lofty specs.

I think the difference between 16bit and 24bit is far more apparent than sample rate.

CC

[monkey man]

Firstly, I was joking about the emperor thing, conleycd, hence the "LOL".

The problem with the "human hearing limit" tests, IMHO, is that they test frequencies in isolation. This approach, IMHO, is irrelevant to the argument. I'm not contesting man's ability to hear any frequency above 20k in isolation. In fact, I believe that sum and difference frequencies come into play when perceiving a mix, and that this is only one of a possible number of explanations for the phenomenon.

I still think that 192 vs 44.1 should be an "easy" distinction to make.
88.2 or 96 vs 192kHz would be a different matter altogether, you'd think, but the currently-popular maximum vs the minimum, IMHO, is a vastly different comparison. I agree with you about s/rate conversion - don't like 'em myself, so I work at 44.1 all the way, and always have. Sounds fine to me.

I agree on the 24 vs 16 bit difference; it should be much more obvious than the former comparisons.

EDIT: I meant also to say that because a frequency can't be heard in isolation, it doesn't mean it won't be missed when removed from a mix or an instrument; this in an entirely different situation, IMHO, and not one that's addressed by the aforementioned tests.

[James Steele]

please keep the question on the gear page. I'm not complaining about any kind of MOTU "policy" to keep their tech specs and industry secret.
I'm demanding that the information that Audio Engineers like myself would like to know.

Ah, yes... I remember, you're DEMANDING something here. Well, it looks like you're going to keep at it and keep at it, so I shall just have to give in and let you keep the post here.

This one is phrased more like a question, but to my mind the last time you posted this, it was more like a complaint about MOTU policy and that's why I moved it, and explained it to you, and took flack from you for it. And so here you are posting it yet again. Okay.

So understand this. This board is not run by MOTU or affiliated with them. And in the other thread you were told that MOTU gives specs out over the phone, etc. So you might want to ask them, rather than here, as I don't know how effective asking it here will be.

[singproinc]

Pretty much the definitive technical view on the effectiveness of high sampling rates is Dan Lavry's white paper on Sampling Theory for Digital Audio. Here's a link to a pdf of Dan's paper:

http://www.lavryengineering.com/documen ... Theory.pdf

Dan's exhaustive writing on the subject basically comes down to the fact that higher sampling rates (like 192khz) are actually worse than 44.1/48 because of instability, and there is no audible quality increase with higher sample rates. Read it with a good strong cup of coffee in your hand to keep awake, but it's loaded with gold info.

Most engineers that I've talked to seem to think that 24-bit word length is more important than high sampling rates. I liked what Elliot Scheiner (Google him if you're unfamiliar with his resume) said to me when we were talking about 96k and 192k sampling rates. He said (paraphrased, because I didn't write it down at the time), "I've never heard anything up that high, I've never seen a system outside a control room that can reproduce that high, and I'm not sure how to eq anything above 20khz. What would I do at 32khz or 48khz anyway?"

My own experience as a Grammy and Emmy nominated engineer and member of the NARAS Producer & Engineers forum hasn't demonstrated an audible or other compelling reason for high sample rates.

YMMV; but that's what's been my experience.

[fokof]

Isn't the sampling rate frequency the quantity of information INSIDE one second of data instead of a frequency response ?

I think that a 192khz "sampling rate" means that there is gonna be 192 000 samples in one second. A 48khz "sampling rate is 48 000 samples in one second.

Nothing to do with the range of the D/A........

You can have a unit capable of 192khz but that would have a 30-18khz frequency response because of a deficient analog section.........

[fokof]

....sorry double post

[EMRR]

Probably applicable:

I measured the actual MATCHING output impedance of the 2408mkIII tonight, and found it to be in the 150-200 ohm range. This, of course, will probably be at a diminished maximum output voltage level and an increased distortion figure. Seems to be no real problem driving 600 ohms. Testing of a 200 ohm passive mic mixer with Spectrafoo shows no significant erratic frequency response issues related to low-Z loading.

I have not opened a unit up to see what build out resistance, of any, exists to set a minimum output impedance, while protecting the output driver circuit. There will surely be something, even if it is imbedded within an IC and consequently not viewable.

[singproinc]

Isn't the sampling rate frequency the quantity of information INSIDE one second of data instead of a frequency response ?

I think that a 192khz "sampling rate" means that there is gonna be 192 000 samples in one second. A 48khz "sampling rate is 48 000 samples in one second.

There are of course, 2 major pieces of math to describe a digital sample; the number of samples per second, and the resolution of each of those samples; AKA the word length or number of bits in each of those samples.

Sample rate relates to the frequency range. Sample rate must be 2x's the highest frequency you want; thus to accommodate a really good ear with 18khz hearing range, the sample rate is more than doubled to 44.1khz or 48khz to achieve a frequency response of 22.5khz or 24khz, beyond the hearing of most humans.

Word length (24 bit/16 bit/8 bit) relates most directly (in our day-to-day use) to signal-to-noise ratio. 8 bit is really noisy. 24 bit reaches the theoretical limit of signal-to-noise. 16 bit works well for lots of people, as we've seen with CDs, and the first mass released digital recording systems.

Read Dan's paper; he quickly defines those terms and their nuances.

Re: the ability of the D to A to output a wide swath of frequencies, check the above referenced white paper. There are devices that can put out 0hz to light frequency, and we've had them for a long time; including radio frequency and infrared frequency transmitters.

Frequency response on the analog output side is not a limiting factor in sample based audio products. Dan's paper discusses those issues. Again, it is a widely accepted treatise, and accepted by very smart people. The major points can be understood by most people in a couple of read-throughs.

Ultimately it comes back to this: what is effective? If I have an SM-57 with a 3-db down point at 14khz, micing a Fender Princeton with a 3-db down point at about 8khz, how will the ability to tweak the eq at 48k improve that? Or the same for a Neumann mic on a voice with a 3db down point at 18khz? Maybe the math and the manufacturing should concentrate on other factors that are more audible like jitter and clocking. The paper discusses that.

It's a must-read for anybody that wants to know more about sample rates and word length. It certainly is a must read for those that wish to speak with conviction and, at the same time, with a handle on the vocabulary and basic physics.

[jp009]

First off I am posting this to match up setting in the eq for vocal micing technics with premium mics such as the U87 or tlm103.

No offense to MOTU (I own a MkII), but a U87 deserves better than a stock MOTU preamp.

That aside, it does appear that MOTU is try to hide something by not talking about their preamp and converter technical details. I can understand not putting the info in ads, but not on their site or in the manual??

[fokof]

Sample rate relates to the frequency range. Sample rate must be 2x's the highest frequency you want; thus to accommodate a really good ear with 18khz hearing range, the sample rate is more than doubled to 44.1khz or 48khz to achieve a frequency response of 22.5khz or 24khz, beyond the hearing of most humans.

Didn't know that my HD192 was producing 96khz to my speakers.

Even if it does , I'm not a bat......

[singproinc]

Didn't know that my HD192 was producing 96khz to my speakers.

Even if it does , I'm not a bat......

That's how digital sampling works. Your HD192 puts out a theoretical frequency response of 96khz; which as you indicate is audible to some dogs and to most bats; but not any humans. It's called the Nyquist theorem, or more accurately, the Nyquist-Shannon theorem, and it's pretty much the foundation of PCM sampling as we currently use it.

Really, for anybody who uses PCM digital technology (that includes all of us on this board), Dan's paper is a must-read; if only to get a grasp of those fundamentals. What it explains in depth is pretty mind-blowing for people with high sample rate systems.

I apologize for hijacking this thread off the OP's subject. It won't happen again.

[thashobs]

analog = infinite sample rate

How do we strive to make digital gear as good as analog? higher sample rates..... new products coming in the next few years include 32-bit 256k sample rates. pro tools as well as all daws will have this.

can you hear a difference between an mp3 at 128k vs. 320k if you answered no than your probably over 30 or are almost deaf.

why do people relate the sample rate as being the frequency response. With all do respect: if you say you cant hear a difference between 192k and 44:1 sample rate your an idiot. First the;192k means that the song is sampled 192 thousand times per second 44.1 is only 44 thousand times per second. If you were to draw an analog waveform it would be perfectly curved. A digital waveform under the microscope looks like jagged stairs. the number of jagged edges depends on the sample rate. the higher the sample rate the smoother the digital representation of the analog waveform. this is why analog sounds so much better. also analogs comparative sample rate of analog gear would be infinite(yes infinity)
the reason for can hear a difference in sample rates is because even though we cant hear past 20khz: fundamental frequencies still exist in music way beyond 30khz .(they are often called "even order harmonics" that go up to 100khz and beyond) if you don't care about fundamental frequencies or harmonics than you should keep recording at 44.1 but if you do care you should record at higher sample rates. These high frequencies although not audible still resonate and you can still feel and perceive the high frequency content. they also add to the color of the sound.

also you cant hear bass below 20 hz but can you feel it?
the answer is yes
so why wouldn't frequencies past 20khz make a difference on the clarity of the HF content