Edirol R44 'sensitivity vs 'level

[Wayne]

A few questions if anyone knows regarding the R44.
'Sensitivity is shown pre-A/D in the block diagram, 'level post A/D (as is the limiter for that matter). I see 'level has finer adjustments and they refer to that as a way to trim record balances but is it indeed after the A/D? Since they don't show where the input metering takes its signal in the path shall I presume that when they say the level knob 'straight up = '0db’, then that would show the actual (at the A/D) record level?

[tnjazz]

Originally Posted by Wayne

A few questions if anyone knows regarding the R44.
'Sensitivity is shown pre-A/D in the block diagram, 'level post A/D (as is the limiter for that matter). I see 'level has finer adjustments and they refer to that as a way to trim record balances but is it indeed after the A/D? Since they don't show where the input metering takes its signal in the path shall I presume that when they say the level knob 'straight up = '0db’, then that would show the actual (at the A/D) record level?

Yes to both of your questions

[Ozpeter]

Assuming you are going to be putting the recording through some kind of postproduction process, the inner smooth knob does nothing useful - set it to the 'noon' position, keep an eye on it in case it's got accidentally moved, but otherwise forget about it.

The outer clicky knob handles the all-important preamp gain. With the smooth knobs on 'noon' indeed the metering is showing preamp metering in essence. To the left of the meters there's an area which lights up in the event of preamp overload as well.

Increasing the setting of the inner knob above 'noon' needlessly risks digital overloads. Reducing it could mean that you were recording preamp analog clipping at a lower level.

The outer clicky gain controls change the gain in 6dB steps, except in the case of a certain position (I think about the 4th click from the bottom) which is actually more like 12dB. Test that yourself.

[Wayne]

Thanks for that. It makes sense and its simple enough but I swear their verbiage made it less than straight forward first few times out.
Not that I’ll be using the limiter, but 'preventing overs'? Seems then they forgot that little tidbit of needing to knock twenty or so off the record level for it to be of any use.
Oops.
Clarity in writing please.

[wshaw]

I'm doing a live recording soon, and a videographer wants to take a stereo signal from my r44. I believe the only way to do that is via the headphone out, using the level dials to "mix" the outputs. Does that sound right to y'all?

[Prick Up UR Ears]

Originally Posted by wshaw

I'm doing a live recording soon, and a videographer wants to take a stereo signal from my r44. I believe the only way to do that is via the headphone out, using the level dials to "mix" the outputs. Does that sound right to y'all?

yes, thats the only way if you record on all 4 channels at the same time - I had the same situation once too.
But be aware of possible delays between channels when you mix direct signals and ambient microphone signals

[wshaw]

My sources will be a pair of Beyer mc930s, set up at the lip of the stage, and direct outs from the desk--one for vocals, one for instruments. My guess is that time smear shouldn't be a huge problem for the on-site "mixdown." And I can correct any smear later, while doing my real mix at my home studio. Anyone more pessimistic?

[Prick Up UR Ears]

Originally Posted by wshaw

My sources will be a pair of Beyer mc930s, set up at the lip of the stage, and direct outs from the desk--one for vocals, one for instruments. My guess is that time smear shouldn't be a huge problem for the on-site "mixdown." And I can correct any smear later, while doing my real mix at my home studio. Anyone more pessimistic?

sure you can correct it afterwards (and possibly have to) but for the on site mix for the video, you may better just use the 930s.
Hope you will have sufficient time to test and adjust it beforehand.
Good luck for the recording

[Tommy-boy]

Originally Posted by wshaw

I'm doing a live recording soon, and a videographer wants to take a stereo signal from my r44. I believe the only way to do that is via the headphone out, using the level dials to "mix" the outputs. Does that sound right to y'all?

That's probably the easiest way, but not the only way. You can also use the digital out, which has the same audio signal as what goes through the headphone jack. The third choice is to take the four lines out into a separate little mixer to get to 2 channels.

-Tom

[wshaw]

Right. I thought of those too. Question is whether the sonics would be improved enough to make it worthwhile, given that the signal will be run into and recorded by a video camera.

[Tommy-boy]

Back to the OP's comments - I also find the manual and block diagram a bit confusing sometimes.

I really don't understand the logic behind putting the level (little inner knob) after the A/D converters, unless their converters can take more than 0dBFS or the monitoring is somehow working off of different base.

Even more baffling to me, is the limiter, and how it's after the A/D converters. I really don't understand this logic and question whether the block diagram is right.

Even more confusing to me about the limiter is that it doesn't impact the internal mics. The other day I sat in the audience of a wind ensemble concert and recorded with the internal mics. there were approx 6 bands that came and played. I adjusted the senstivity/level settings as each band warmed up. I nailed all of the settings, except for one band. When it came time for the bass drummer to make his hits, he wound up like he was swinging a bat and cranked the sh*t out of the drum. My r-44 was lit up like a christmas tree. The over lights were blinking like crazy as this guy pounded the crap outta his drum. I expected the recording to be ruined, but it wasn't! It was as if there were no overs at all. I don't understand how this is possible, but I didn't hear any clipping despite the fact that the r-44 clearly indicated that I was clipping (and not by a small amount) when i tracked it.

Go figure. If somebody can explain this stuff to me, I'd be grateful.

Despite all of the above comments - I really love this device. It is a sweet 4 track recorder.

-Tom

[Ozpeter]

There is indeed not much point in the digital level controls unless you can't post-process the signal in a DAW later.

The limiter causes the set analog input level to be lowered by 12 dB, and then digital limiting applied followed by 12dB of makeup gain. Record with 24 bits at a more conservative level instead.

I think you are the first person to use the built in mics so your experiences with them constitute original research!

See pages 18 and 69 of the manual for output assignment setting.

Bear in mind that the is no control of panning during monitoring, with the channels 1 and 3 going to the left output and channels 2 and 4 going to the right output, which might not be what the videographer wants. The only way to get a properly panned and balanced monitor mix from the R-44 is to feed its individual outputs to an external mixer. In my case I use one designed for the purpose by Naiant. The new Tascam DR-680 addresses this problem by providing a proper monitor mixer on board.

[A3rd.Zero]

Originally Posted by Prick Up UR Ears

yes, thats the only way if you record on all 4 channels at the same time - I had the same situation once too.
But be aware of possible delays between channels when you mix direct signals and ambient microphone signals

You can also set the "monitor" output to the RCA outputs on the side. In system setup > output setup you can set all four outputs to run out RCA 1&2 as L&R respectively.

EDIT: As Ozpeter said, sorry to repeat, it's on page 69 of the manual. I use a Sound Devices MP2 in front of the R44 and I feed the LR RCA monitor of the R44 into the SD's tape return so that I can monitor both the SD and the R44 through the SD's headphone jack, which while making it easier to switch back and forth, has the added benefit of not sucking donkey nuts.

[Mark A. Jay]

Originally Posted by Wayne

A few questions if anyone knows regarding the R44.
'Sensitivity is shown pre-A/D in the block diagram, 'level post A/D (as is the limiter for that matter). I see 'level has finer adjustments and they refer to that as a way to trim record balances but is it indeed after the A/D? Since they don't show where the input metering takes its signal in the path shall I presume that when they say the level knob 'straight up = '0db’, then that would show the actual (at the A/D) record level?

Hey all...

I wrote the (now, hopefully) attached Excel sheet as a sort of quick and dirty reference check for interfacing mics with front ends. It's pretty self-explanatory, but in a nutshell, what this does is allows you to enter the microphone sensitivity either the old fashioned way (i.e. 'x' dB re 1V at 1 Pa) or as well as mV/Pa.

There's a provision for the user to enter the maximum anticipated SPL and based on that and the microphone sensitivity, the corresponding dBu, dBm, and dBV values are returned.

Finally, I made a table that starts at 0 dB and runs up to 140 dB SPL. The table gives you the corresponding pressure in Pa (for each 1 dB step in the table) as well as the corresponding dBm, dBu, dBV, and things like Volts RMS proper, millivolts RMS...millivolts peak...peak to peak etc.

It's not like this is specific to the Edirol, but I do think it can be a handy spreadsheet application.

One last thing...I did NOT protect the worksheet - too lazy this morning. There are only three user input cells (which are explained in the text box on the sheet), so if you accidentally change a formula or formulae, just close it and (obviously) don't save the file...or...you could protect the sheet yourself.

Anyway, let me know if it helps.

Mark

[Mark A. Jay]

Originally Posted by A3rd.Zero

....I use a Sound Devices MP2 in front of the R44 and I feed the LR RCA monitor of the R44 into the SD's tape return so that I can monitor both the SD and the R44 through the SD's headphone jack, which while making it easier to switch back and forth, has the added benefit of not sucking donkey nuts.

The 'not sucking donkey nuts' comment had me literally laughing out loud. Thanks for that...

[Modo]

I seem to be rather confused as well with this topic, I have an r44 and after reading what's discussed here... I did a sound test

I recorded some dialogue while raising and lowering the smaller inner level knob to tape. When I played it back in my DAW, the level knob did change how much signal was going to tape.

I just bought my r44 less than two weeks ago. Maybe I have a newer r44 that is updated? But as far as I know, the level knob is an input trim.

[NorseHorse]

Hello Modo,

Yes, that is correct. The level trim does effect the final file that is written.

To understand our concern about the difference between the two knobs, try these basic recording experiments.

Number 1: Set the level knob to NOON (no change). While talking into a microphone, raise the sensitivity knob until you see that the you are clipping. Now turn the level knob COUNTER-CLOCKWISE (negative change) till the clip indicator is no longer being triggered.

When you play back this file, you will hear that yes, the volume does change when turning the level knob, but more importantly ---> the clipping (distortion) is still there. This is because the clipping occurs before the level knob.

Number 2: Set the level knob to NOON (no change). While talking into a microphone, raise the sensitivity knob until you see that you are close to clipping, but not clipping. Now turn the level knob CLOCKWISE (positive change) so that you are clipping.

When you play back this file, you will hear that yes, the volume does change when turning the level knob, but more importantly --> you now have clipping (distortion).

Lesson learned? Keep the level knob set to noon (12 o'clock, unity). It is the only way to accurately reflect the analog gain. If you have it to the left, you might be clipping and not know it. If you have it to the right, you might cause clipping when you didn't need to.

[Mark A. Jay]

Originally Posted by NorseHorse

Hello Modo,

Yes, that is correct. The level trim does effect the final file that is written.

To understand our concern about the difference between the two knobs, try these basic recording experiments.

Number 1: Set the level knob to NOON (no change). While talking into a microphone, raise the sensitivity knob until you see that the you are clipping. Now turn the level knob COUNTER-CLOCKWISE (negative change) till the clip indicator is no longer being triggered.

When you play back this file, you will hear that yes, the volume does change when turning the level knob, but more importantly ---> the clipping (distortion) is still there. This is because the clipping occurs before the level knob.

Number 2: Set the level knob to NOON (no change). While talking into a microphone, raise the sensitivity knob until you see that you are close to clipping, but not clipping. Now turn the level knob CLOCKWISE (positive change) so that you are clipping.

When you play back this file, you will hear that yes, the volume does change when turning the level knob, but more importantly --> you now have clipping (distortion).

Lesson learned? Keep the level knob set to noon (12 o'clock, unity). It is the only way to accurately reflect the analog gain. If you have it to the left, you might be clipping and not know it. If you have it to the right, you might cause clipping when you didn't need to.

Have a look-see at this image - you can see that the 'sensitivity' controls the gain in a pre-A/D sense whereas, according to the block diagram, the 'level' is post A/D.

So, Norse is right - if you saturate the first stage (by having the 'sensitivity' too high) then no matter what, clipping ensues - and you can't un-do that.

I found this in the maual with regard to the clipping indicator(s); this should help clear matters up for you (see the points with asterisks next to them (and they are in bold italic font)):

Adjusting the input levels

1.

Turn the Input level knob [SENS] (outer) all the way left.

2.

Turn the Input level knob [LEVEL] (inner) to the center position. This position is 0 dB.

3.

Hold down the [PAUSE] button and press the [REC] button.
The R-44 enters recording standby mode. During recording standby, the [REC] button blinks
and the Display alternates between REC and STANDBY.

4.

Play the sound to be recorded into the microphone.
Gradually turn the Input level knob [SENS] toward the right.
Sensitivity has 11 steps of +4, -2, -8, -14, -20, -26 ,-32 ,-38, -44, -50, and -56 dBu.

* When the channel name display reverses, the input level has reached the clipping level. Set the knob
to the position just before the display reverses.

5.

Adjust the Input level knob [LEVEL] so that the Level
meter shown in the display reaches a point slightly
before C (clip level). If the recording level is too low,
quiet sounds will not be recorded. If the recording level
is too high, loud sounds will be distorted, producing a
crackling noise in the recording.

* When the channel name is blinking, clipping has occurred at
the sensitivity adjustment stage. In this case, clipping
occurs even if the Level meter does not reach the clip level.
* The Level meter displays the clip level (C) in values relative to 0 dBFS (FS = full scale).

For example, 12 indicates -12 dBFS.

Anyway, here is the image (block diagram):

[Wayne]

Originally Posted by NorseHorse

...Lesson learned? Keep the level knob set to noon (12 o'clock, unity)..

Yes. Frankly I don't see the point of their step 5.
And I wish; They didn't refer to both as 'Input Level', then call one 'Sens then the other-- 'Level'.
I'm getting better, but every time I have to go through that 'Let's see, just remember 'always.. no, never.. no 'always..'

(nod to Steve Martin there...

[Mark A. Jay]

Originally Posted by Wayne

Yes. Frankly I don't see the point of their step 5.
And I wish; They didn't refer to both as 'Input Level', then call one 'Sens then the other-- 'Level'.
I'm getting better, but every time I have to go through that 'Let's see, just remember 'always.. no, never.. no 'always..'

(nod to Steve Martin there...

First, thanks for the Steve martin reference - that was one funny bit.

Second, I'm in total agreement regarding the controls; in my other job (presently an Acoustics / NVH / Sound Quality Group Manager) I am used to dealing with multi-channel front ends where the only sensitivity control that one has is the full-scale sensitivity of the front end. That is, something like 10 mV, 20 mV, 50 mV, 100 mV and so on. I personally like this approach because there is no ambiguity that could be borne of a secondary gain control. However, in these systems, each transducer's sensitivity is part of the scale factor along with the engineering units specified - for acoustic acquisitions, these units are (of course) Pascal. There seems to be a bit of a gulf between the data acquisition and recording worlds, and I would argue that there's no real reason for this; it would be very handy to (at all times) know just how much acoustic pressure (in Pascal) is present at any mic at any time. Granted, if you know the dB re FS and the transducer sensitivity then you can ascertain that, but it would be nice to always see either Pa or dB SPL as a display option in all mixing software (this is how it is done in the noise and vibration software world).

In other words, if you are using a garden-variety B&K microphone (i.e. type 4189, 4190 etc) then you know its nominal sensitivity is 50 mV/Pa; the software 'knows' this (either by TEDS or by manual entry / calibration), and thus, whatever voltage is seen by the front end equates to a given acoustic pressure in Pascal - all that you need do during acquisition is to ensure that the A/D are properly ranged (no under or overload) and thus, there is no ambiguity borne of 'downstream' gain controls. Anyway, I like having to deal only with that one variable (ranging the front end). Mind you, some of the newer hardware (slated for noise and vibration analysis work) used in test labs does not really require any ranging of the A/D converters (Bruel & Kjaer's Dyn-X front ends, for example).

In a perfect world, the Edirol would simply have the sensitivity controls and omit the level controls entirely (and it would have been less expensive to build as well). Looking at the block diagram (if it is accurate), it would seem that the level controls actually control magnitude in the digital domain. As such, it would seem that a firmware version could be released that simply forces the level controls to be static at 0 dB (unity), thus negating their control. That would be a nice feature to have (in my opinion) and would mean that only the sensitivity control would be used, allowing the A/D's to be properly ranged based on one control with no ambiguity.

[kevinsinnott]

Hello, I'm reading this thread in interest. I have an R44 and use it constantly. I've been setting the inner knob to maximum under the idea that it's best to use the least analog gain I can to keep noise down.

I've not had problems with overload. Am I wrong to do this, soundwise that is?

[Wayne]

Originally Posted by kevinsinnott

Hello, I'm reading this thread in interest. I have an R44 and use it constantly. I've been setting the inner knob to maximum under the idea that it's best to use the least analog gain I can to keep noise down.

I've not had problems with overload. Am I wrong to do this, soundwise that is?

Looking in the manual Level' adds +8 digital gain (post A/d), then I guess that means you're sort of forcing a reduction of the analog gain about the same going into the a/d (plus what ever safety margine below 0dBFS?
Presuming default to the idea of nominal record level is -18 -22db range = good line level design in the front ends and a/d's, hopfully that already puts it in the good range.
The digital boost may be harmless enough, just seems a round about way to get here.

add.. Would you get to the same result by setting for additional -8 on your record target range ( -26 -30dBFS?

[Ozpeter]

I've been setting the inner knob to maximum under the idea that it's best to use the least analog gain I can to keep noise down.

It's the other way round. In most if not all preamps, the signal to noise ratio is best at highest analog gain (which is why it's normally measured at that point). Yes, you hear a lot of noise subjectively, but the gain will have increased more than the noise does each time you click the control clockwise. So the signal to noise ratio will have improved.

So you could say, lower the digital gain as far as practical and raise the analog gain to compensate. However, you'd be very likely to get analog clipping which the lowered digital gain will not improve.

So - leave the inner smooth-turning knob at the "noon" position. The only point in using it is if you don't have any means for undertaking post production normalisation of the recording, which these days is pretty unlikely.

[kevinsinnott]

Originally Posted by Ozpeter

It's the other way round. In most if not all preamps, the signal to noise ratio is best at highest analog gain (which is why it's normally measured at that point). Yes, you hear a lot of noise subjectively, but the gain will have increased more than the noise does each time you click the control clockwise. So the signal to noise ratio will have improved.

So you could say, lower the digital gain as far as practical and raise the analog gain to compensate. However, you'd be very likely to get analog clipping which the lowered digital gain will not improve.

So - leave the inner smooth-turning knob at the "noon" position. The only point in using it is if you don't have any means for undertaking post production normalisation of the recording, which these days is pretty unlikely.

Thanks to both of you. I assumed otherwise, but I'll accept you understand it. I made a recording this morning using the 12noon positions on the inner knob. I'll know soon if it is quieter or noisier.

Thanks again.

Revised @9:30pm 8-2-11 I did several tests using the 12PM digital level positions and it works fine. I guess my fears were unwarranted. I'm not sure it sounded substantially better, but my goal is to start with a proper baseline. It took me a while to "get" their thinking. Frankly, I'm still not sure I understand it, but I'm recording. I've had the best reults with two mikes: An XY stereo Audio Technica 4035 recommended by Doug Oade and a Superlux ortf stereo mike. Both do very well through the Oade-modified mic pres.