Proper Gain Staging with a Tascam DA-38 and DP7

[Rhetro]

Hello.

I couldn't find anything specific to the Tascam DA-38, so I thought this was the best place to post this. I also posted this on the Tascam forum so please excuse the redundancy with the DP7 comment.

I recorded a session on a DA-38 with an MA-AD8 (man, i still love the sound of this thing!!!). I used all 8 channels and need to output this tape to an iMac running DP7 via a Presonus StudioLive 16.4.2 board.

The recording process went well i suppose, however some of the tracks were less than optimal. Everything went through the MA-AD8 with no external gear (no, i didn't have the StudioLive board at the time). I got some red overs on some of the channels. On playback, everything sounded good though. I didn't hear any distortion/clipping. I'm a little confused to these overs being actual clipping (even though i didn't hear it), or just over sampling at the output digital-to-analog stage??? Something like that?

Anyway, from what i've read, to proper gain stage, I should set my faders on my board to halfway to 2/3 the way up. (the Presonus people actually told me -10db which about half way). On my audio monitor ( DP7 level meter), i should look for my levels going no higher than -6dB.

Now my question: The RCA outputs on the DA-38 are unbalanced at -10dB. The DB25 (sub-D) output is balanced at +4dB. I can use either or; however. Since i got some red overs on some of the tape channels, won't i have LESS headroom if i use the +4 (DB25), even though it's balanced (less noise)? Or. Won't i have MORE headroom with the RCA outputs even though they're unbalanced (more noise)? Do i try them both and see? Do i have the right idea?

If anyone is familiar with DP7 here, I'm also thinking about setting up a 16 channel template that has the trim plugin at the top of the chain (on all the channels). This would allow me to leave the faders at unity (or -10dB), and use the trim plug to approximate -6dB on the level meters.

Please tell me i'm not THAT far off base!!

Thanks !

[rodger1811]

Wow, I really don't know where to start...... Well, first of all, with regard to your digital overs, you may or may not here anything glaring with regard to it going over. Sometimes it's transient detail that clipped and you may not notice it which may explain why it sounds ok to you. However I submit, that it would have sounded even better if it hadn't gone over.

Secondly, there is no difference per se` in sound with regard to balanced and unbalanced. The main thing is to make sure that you're matching output and input impedance. for example, one of the most common problems that newbies (NOT SAYING YOU"RE ONE) have is sending a -10dBu output source to a +4dBV and not understanding why the levels are so hot! The fact of the matter is that there is between (depending on the math) a 11db-14db difference between the two.

-10dBu devices, (unbalanced) are HIGH Z or high impedance and +4 dBV are LOW Z or low impedance. So, if you connect a high impedance output device to a low impedance input port, you'll overburden the port and not be able to get an acceptable level because it's just too darn hot. If you plug a low impedance device in the a port that's configured for high impedance, it will seem that you can't get a high enough level. Generally speaking, unbalanced output devices are considered to be consumer grade and balanced output devices are considered to be professional grade.

Again, there is NO difference between the sound so to speak. If you have long runs then you may be forced to go the balanced route with cables. So for example, if you have a keyboard that has -10dBu output and you're well beyond 12ft or so, to avoid noise etc. you would need to use a DI and go through a preamp.

Motu's "Trim" pluggin is a great tool but it can't prevent you from clipping at your input or even before that at your preamp. You must learn how to get the best signal from point A all the way through to the input on the interface. There really is no magic answer other than to stay well within the limits of everything in the chain.

[Rhetro]

Hey Rodger1811!

Thanks for the reply.

No, i'll wear that hat -newbie is fine! Gotta start somewhere!

I understand what you mean about matching the impedance; looking for nominal output of the DA-38, to nominal input of the Studiolive board (which is what i need to figure out next)-proper gain staging, right? So does the Studiolive board prefer high or low impedance?; i'll see what the manual says.

I wasn't looking at it from an impedance perspective, and was concerned about the headroom limit i imposed on myself with those peaks -the two having nothing to do with each other!!!

The peaks are there, nothing i can do at this point (again, there is nothing noticeable -but i'd really wish they weren't there as you said!). I really, really, really hope i don't have to retake. I'll see what it sounds like once it's in the computer.

In essence: High Z to High Z ; low z to low z, right?

Thanks again.

[rodger1811]

In essence yes, you've got it but there really is so much more. Just getting the physical chain wired up correctly is half of the journey to appropriate gain structure. The other tricky part of correct gain structure, or shall I say appropriate gain structure, is truly understanding your gear and everything in your signal chain.

Every hardware piece has what I call a "sweet spot" which is where it's really designed to operate. Sort of like a guitar amp sounds best in some cases when it's master volume is at the mid point, or maybe two-thirds from max. Although we all know that it can be turned up to max, it rarely if ever sounds good. Every compressor, eq, mic pre, and audio interface input, has a sweet spot. Even software plugins have a sweet spot if they're modeled well.

The short answer to what headroom is, is the difference between what your recorded level is, and how much is left before you reach your digital maximum of 0db. This brings about another thought...... The meters in most DAW's are dBFS and the meters on most outboard gear are dBVU. This one issue adds quite a bit of confusion! dBFS meters are what's referred to as "Peak" meters and Dbvu meters are "average" meters. So, when you are averaging 0Db on your outboard mic pre, you should be reading roughly -18Dbfs on the meters within your DAW. Again, understanding this is critical to proper gain structure.

I don't mean to go on and on with this but I get more questions on this than ANYTHING else so I hope I haven't inundated you to badly.

[Rhetro]

Yep.

Just read in the Presonus manual and even called to double check:

+4dBu output (DB25 pin) balanced on the DA-38; +4dBu input (TRS) balanced on the StudioLive board. Faders at -10

Let's see how it sounds! Yikes. This could be scary. But if i need to retake, i know what to do now -better late than never!

[1nput0utput]

The main thing is to make sure that you're matching output and input impedance. for example, one of the most common problems that newbies (NOT SAYING YOU"RE ONE) have is sending a -10dBu output source to a +4dBV and not understanding why the levels are so hot!

You are wrong. Most professional equipment uses a reference level of +4 dBu, and some consumer or low-cost equipment uses a reference level of -10 dBV.

-10dBu devices, (unbalanced) are HIGH Z or high impedance and +4 dBV are LOW Z or low impedance.

I don't think this is true. Line level outputs will most often have a low source impedance regardless of the reference level used. To put it another way, the nominal reference level specified for a particular device and the device's input or output impedance are never necessarily correlated.

So, if you connect a high impedance output device to a low impedance input port, you'll overburden the port and not be able to get an acceptable level because it's just too darn hot. If you plug a low impedance device in the a port that's configured for high impedance, it will seem that you can't get a high enough level.

An impedance mismatch does not always result in an unexpectedly high or low signal level. Refer to these excerpts from the Yamaha Sound Reinforcement Handbook by Gary Davis and Ralph Jones. It is considered a standard reference for studio and live sound engineers:

In most modern, line level audio circuits, it is considered beneficial for the output's source impedance to be as low as practical, and the input's load impedance to be as high as practical within limits …

A circuit where the input termination impedance is a minimum of some ten times the source impedance of the output driving that input is said to be a bridging input, and the output is said to be bridged.

In some modern equipment, and in most older equipment, it was desirable for the input's terminating impedance to be roughly the same as the output's source impedance. Such circuits are said to be matched.

It is important for you to know whether the output of a particular piece of equipment is supposed to be matched or bridged, or whether that doesn't matter. When there is an impedance mismatch (which means the source and load are not right for one another, whether matched or bridged), the results can range from improper frequency response to excess distortion to incorrect operating levels to circuit failure. In terms of specifications, it is important to know what impedances were used when measuring the specs in order for the specs to be reproducible.

So for example, if you have a keyboard that has -10dBu output and you're well beyond 12ft or so, to avoid noise etc. you would need to use a DI and go through a preamp.

More specifically, what you mean here is that long, unbalanced cable runs for low voltage signals should be avoided whenever possible. The purpose of using balanced cable runs is to minimize interference.

The answer to the original poster's question is that it doesn't matter whether you use the DA-38's unbalanced -10 dBV outputs or its balanced +4 dBu outputs if the signal on tape is already clipping. The clipping indicator is alerting you that a digital sample on that track has the maximum possible value (0 dBFS) or is close to that value. (Some devices' clipping indicators come on "early" to warn you that clipping may occur soon.) This will result in clipping at the D/A output that drives both the balanced and unbalanced analog outputs. However, if you don't hear a problem when you listen to the recording, then there are few enough clipping samples, and you shouldn't worry about them. (Many clipping samples in a row would be heard as distortion, but a few generally aren't that problematic.)

You should choose which DA-38 analog outputs to use based on the analog inputs of the mixer or audio interface you're using. If the mixer inputs are referenced to -10 dBV, then the DA-38's +4 dBu outputs will be too hot. To compensate, you would need to apply attenuation at the mixer inputs, which will lower the signal-to-noise ratio. In that case, it would probably be better to use the -10 dBV outputs from the DA-38, use short cables to avoid interference, and apply no attenuation at the mixer inputs. If the mixer inputs are balanced, then that probably implies that they are also referenced to +4 dBu. In that case, you should use the DA-38's balanced outputs. The signal will benefit from interference rejection, and the signal-to-noise ratio will not be compromised by attenuation in the signal path.

[1nput0utput]

+4dBu output (DB25 pin) balanced on the DA-38; +4dBu input (TRS) balanced on the StudioLive board. Faders at -10

Why do they recommend setting the faders to -10 dB? If the reference level of the two devices is the same, then no gain or attenuation should be needed. I would assume the faders should be set to 0 dB. I think the recommendation of -10 dB for the faders would be correct if you were mixing down to one of the mixer's stereo outputs. But you're not doing that, right?

[rodger1811]

The main thing is to make sure that you're matching output and input impedance. for example, one of the most common problems that newbies (NOT SAYING YOU"RE ONE) have is sending a -10dBu output source to a +4dBV and not understanding why the levels are so hot!

You are wrong. Most professional equipment uses a reference level of +4 dBu, and some consumer or low-cost equipment uses a reference level of -10 dBV.

This is obviously a typo

-10dBu devices, (unbalanced) are HIGH Z or high impedance and +4 dBV are LOW Z or low impedance.

I don't think this is true. Line level outputs will most often have a low source impedance regardless of the reference level used. To put it another way, the nominal reference level specified for a particular device and the device's input or output impedance are never necessarily correlated.

You my friend are Totally incorrect on this issue

So, if you connect a high impedance output device to a low impedance input port, you'll overburden the port and not be able to get an acceptable level because it's just too darn hot. If you plug a low impedance device in the a port that's configured for high impedance, it will seem that you can't get a high enough level.

An impedance mismatch does not always result in an unexpectedly high or low signal level. Refer to these excerpts from the Yamaha Sound Reinforcement Handbook by Gary Davis and Ralph Jones. It is considered a standard reference for studio and live sound engineers:

I didn't say always if you read it correctly! But you will definitely have that problem plugging in a -10 into a +4. If you don't believe me, try it!!!!

In most modern, line level audio circuits, it is considered beneficial for the output's source impedance to be as low as practical, and the input's load impedance to be as high as practical within limits …

A circuit where the input termination impedance is a minimum of some ten times the source impedance of the output driving that input is said to be a bridging input, and the output is said to be bridged.

In some modern equipment, and in most older equipment, it was desirable for the input's terminating impedance to be roughly the same as the output's source impedance. Such circuits are said to be matched.

It is important for you to know whether the output of a particular piece of equipment is supposed to be matched or bridged, or whether that doesn't matter. When there is an impedance mismatch (which means the source and load are not right for one another, whether matched or bridged), the results can range from improper frequency response to excess distortion to incorrect operating levels to circuit failure. In terms of specifications, it is important to know what impedances were used when measuring the specs in order for the specs to be reproducible.

So for example, if you have a keyboard that has -10dBu output and you're well beyond 12ft or so, to avoid noise etc. you would need to use a DI and go through a preamp.

More specifically, what you mean here is that long, unbalanced cable runs for low voltage signals should be avoided whenever possible. The purpose of using balanced cable runs is to minimize interference.

The answer to the original poster's question is that it doesn't matter whether you use the DA-38's unbalanced -10 dBV outputs or its balanced +4 dBu outputs if the signal on tape is already clipping. The clipping indicator is alerting you that a digital sample on that track has the maximum possible value (0 dBFS) or is close to that value. (Some devices' clipping indicators come on "early" to warn you that clipping may occur soon.) This will result in clipping at the D/A output that drives both the balanced and unbalanced analog outputs. However, if you don't hear a problem when you listen to the recording, then there are few enough clipping samples, and you shouldn't worry about them. (Many clipping samples in a row would be heard as distortion, but a few generally aren't that problematic.)

You should choose which DA-38 analog outputs to use based on the analog inputs of the mixer or audio interface you're using. If the mixer inputs are referenced to -10 dBV, then the DA-38's +4 dBu outputs will be too hot. To compensate, you would need to apply attenuation at the mixer inputs, which will lower the signal-to-noise ratio. In that case, it would probably be better to use the -10 dBV outputs from the DA-38, use short cables to avoid interference, and apply no attenuation at the mixer inputs. If the mixer inputs are balanced, then that probably implies that they are also referenced to +4 dBu. In that case, you should use the DA-38's balanced outputs. The signal will benefit from interference rejection, and the signal-to-noise ratio will not be compromised by attenuation in the signal path.

Personally, I don't get the point of your reply? Ok, so there are some mistakes but NONE of them have anything to do with the issue of gain staging. If you want to talk dbv dbu that's great but it's not at all the issue.

[1nput0utput]

Personally, I don't get the point of your reply? Ok, so there are some mistakes but NONE of them have anything to do with the issue of gain staging. If you want to talk dbv dbu that's great but it's not at all the issue.

The purpose of my reply was to correct some of your points so that other people reading the thread would have accurate information and to downplay your over-emphasis of input and output impedance with regards to the original poster's question about gain staging. I addressed the original question in better detail in the last two paragraphs of that reply:

The answer to the original poster's question is that it doesn't matter whether you use the DA-38's unbalanced -10 dBV outputs or its balanced +4 dBu outputs if the signal on tape is already clipping. The clipping indicator is alerting you that a digital sample on that track has the maximum possible value (0 dBFS) or is close to that value. (Some devices' clipping indicators come on "early" to warn you that clipping may occur soon.) This will result in clipping at the D/A output that drives both the balanced and unbalanced analog outputs. However, if you don't hear a problem when you listen to the recording, then there are few enough clipping samples, and you shouldn't worry about them. (Many clipping samples in a row would be heard as distortion, but a few generally aren't that problematic.)

You should choose which DA-38 analog outputs to use based on the analog inputs of the mixer or audio interface you're using. If the mixer inputs are referenced to -10 dBV, then the DA-38's +4 dBu outputs will be too hot. To compensate, you would need to apply attenuation at the mixer inputs, which will lower the signal-to-noise ratio. In that case, it would probably be better to use the -10 dBV outputs from the DA-38, use short cables to avoid interference, and apply no attenuation at the mixer inputs. If the mixer inputs are balanced, then that probably implies that they are also referenced to +4 dBu. In that case, you should use the DA-38's balanced outputs. The signal will benefit from interference rejection, and the signal-to-noise ratio will not be compromised by attenuation in the signal path.

[rodger1811]

Personally, I don't get the point of your reply? Ok, so there are some mistakes but NONE of them have anything to do with the issue of gain staging. If you want to talk dbv dbu that's great but it's not at all the issue.

The purpose of my reply was to correct some of your points so that other people reading the thread would have accurate information and to downplay your over-emphasis of input and output impedance with regards to the original poster's question about gain staging. I addressed the original question in better detail in the last two paragraphs of that reply:

The answer to the original poster's question is that it doesn't matter whether you use the DA-38's unbalanced -10 dBV outputs or its balanced +4 dBu outputs if the signal on tape is already clipping. The clipping indicator is alerting you that a digital sample on that track has the maximum possible value (0 dBFS) or is close to that value. (Some devices' clipping indicators come on "early" to warn you that clipping may occur soon.) This will result in clipping at the D/A output that drives both the balanced and unbalanced analog outputs. However, if you don't hear a problem when you listen to the recording, then there are few enough clipping samples, and you shouldn't worry about them. (Many clipping samples in a row would be heard as distortion, but a few generally aren't that problematic.)

You should choose which DA-38 analog outputs to use based on the analog inputs of the mixer or audio interface you're using. If the mixer inputs are referenced to -10 dBV, then the DA-38's +4 dBu outputs will be too hot. To compensate, you would need to apply attenuation at the mixer inputs, which will lower the signal-to-noise ratio. In that case, it would probably be better to use the -10 dBV outputs from the DA-38, use short cables to avoid interference, and apply no attenuation at the mixer inputs. If the mixer inputs are balanced, then that probably implies that they are also referenced to +4 dBu. In that case, you should use the DA-38's balanced outputs. The signal will benefit from interference rejection, and the signal-to-noise ratio will not be compromised by attenuation in the signal path.

There ABSOLUTELY needs to be emphasis with regard to matching proper impedance. How can you honestly get PROPER gain staging without a thorough understanding of this issue? I'd be willing to bet that MANY of us has run into the issue of plugging a -10 device into a +4 input and didn't have a clue why we couldn't get appropriate levels.

Isn't it important to know that you might have a piece of gear that allows you to change the impedance of its inputs so that you can match appropriately? Emphasis needs to be placed on ANYTHING and everything that can impact proper gain structure. For example what good would it be to make sure you're not clipping your input if you're clipping at the mic pre? Many mic pre's don't have Vu or peak meters. What you'll end up with is a bunch of garbage and good levels at the DAW.

Don'd get me wrong, I absolutely don't have a problem with being wrong and I appreciate you pointing out my error; I just think that they're immaterial.

I don't think that I confused the original poster at all and judging by his response, I think he gets it.

[Rhetro]

Thanks guys!

Whew. I've learn more in the past several posts than i bargained for -this is a good thing! I've actually printed everything out! I need to get that Yamaha book as well.

Rodger, no you haven't inundated me too badly. I'm still in the kiddie pool so it's not too bad! Very helpful once i read all the posts a two or three times. It will take me a little while.
It's also nice to know that i may have a little room with the peaking on the DA-38.

1inputOutput, to you as well. I just need to educate myself a little better. To answer your question about setting the -10 on the fader: You're right, i'm not using the board for mixdown (although i suppose i could try that later...hmmmm). I don't know why the guy told to use -10 on the faders -although he did sound tired. At this point i'm just using the mixer as the input source for the DA-38 into the computer. So it would make sense to leave the faders at unity -don't add/don't subtract, right? And yes, i'm going to be using the balanced out (DB25pin) of the DA-38 @ +4 with the balanced line level inputs (TRS) of the StudioLive @ +4. Thank God i have one of those DB25pin to 8 channel TRS snakes!

Just when you scratch the surface, you realize that it's only the tip of the iceberg. Thanks for keeping the heat down! You guys have been very helpful/patient!

Cheers.

[rodger1811]

Thanks guys!

Whew. I've learn more in the past several posts than i bargained for -this is a good thing! I've actually printed everything out! I need to get that Yamaha book as well.

Rodger, no you haven't inundated me too badly. I'm still in the kiddie pool so it's not too bad! Very helpful once i read all the posts a two or three times. It will take me a little while.
It's also nice to know that i may have a little room with the peaking on the DA-38.

1inputOutput, to you as well. I just need to educate myself a little better. To answer your question about setting the -10 on the fader: You're right, i'm not using the board for mixdown (although i suppose i could try that later...hmmmm). I don't know why the guy told to use -10 on the faders -although he did sound tired. At this point i'm just using the mixer as the input source for the DA-38 into the computer. So it would make sense to leave the faders at unity -don't add/don't subtract, right? And yes, i'm going to be using the balanced out (DB25pin) of the DA-38 @ +4 with the balanced line level inputs (TRS) of the StudioLive @ +4. Thank God i have one of those DB25pin to 8 channel TRS snakes!

Just when you scratch the surface, you realize that it's only the tip of the iceberg. Thanks for keeping the heat down! You guys have been very helpful/patient!

Cheers.

My pleasure! For the record, I don't have a problem with my other fellow member; it's just a healthy discussion and it's good to keep each other on our toes. I have to be more careful.

[Rhetro]

Ok.

So i started playing around with it. I.E. playing out from the DA-38 into the Presonus StudioLive board, and then into DP7.

To recap: DA-38 balanced outputs at +4, Presonus mixer balanced inputs at +4, unity gain on the mixer faders (0) -not that the audio monitor in DP7 cares.

Audio Monitor in DP7: Ouch, very hot. Clipping and the little red box show it! I guess this is what you guys are talking about; now i see what you mean.

But...It sounds great!!!! Again, no distortion. That little tape inside the DA-38 doesn't lie! (at least i can't hear it, and i've listened to it enough times now).

However, I guess there's no way to pad the input to keep it at approximately -6dBs (or some better value) in the audio monitor?
I can't do anything with the DA-38.
I can't do anything with the mixer
Can i do something in DP7? I know that if i record it at 16bit, it runs hotter. What about the other way around. Recording it at 24bit to cool it off a little? Ok, ok...I'm reaching for straws here. Just trying to go one step at a time and brainstorm from the little bits and pieces that i've heard -right or wrong (i'll leave that up to you experts)!

If it sounds good, and i can't really do anything to pad it, I suppose the only thing i can really do is ignore the audio monitor (or just close the damn thing!) and just dump it into the computer, and tweak it in the software i.e. trim plugs etc.

This is quite the challenge. But i must admit, i'm enjoying this.

Thanks again gentlemen

[1nput0utput]

Audio Monitor in DP7: Ouch, very hot. Clipping and the little red box show it! I guess this is what you guys are talking about; now i see what you mean.

But...It sounds great!!!! Again, no distortion. That little tape inside the DA-38 doesn't lie! (at least i can't hear it, and i've listened to it enough times now).

However, I guess there's no way to pad the input to keep it at approximately -6dBs (or some better value) in the audio monitor?
I can't do anything with the DA-38.
I can't do anything with the mixer

I would be wary of clipping DP at the input. The first thing I would try is a few dB of attenuation at the mixer inputs if possible. Do the line inputs have a trim? Maybe you could trim down 1 or 2 dB. As I mentioned before, attenuating the signal will affect the signal-to-noise ratio, but that isn't as bad to me as having little to no headroom for the recorded tracks in DP.

Can i do something in DP7? I know that if i record it at 16bit, it runs hotter. What about the other way around. Recording it at 24bit to cool it off a little? Ok, ok...I'm reaching for straws here. Just trying to go one step at a time and brainstorm from the little bits and pieces that i've heard -right or wrong (i'll leave that up to you experts)!

Recording 24 bits instead of 16 bits might help. It doesn't hurt much to try.

[rodger1811]

I STRONGLY urge you to record at 24bit and reduce your record levels. Recording at 16bit will require you to record hotter because of the noise floor. Recording at 24bit virtually does away with that problem. I'd shoot for a recording level riding between -14dBFS and -12dBFS. That will give you room for your transients without the likelihood of clipping at your record input. When you get to the mix stage, you'll have plenty of headroom to do what you'll need to do without concern for adding noise by pushing up a fader etc.