low output using my DIY compressors (1176,LA2A)

I have been recording with my compressors for over a year now and I'm not pleased with the results.  When recording direct from the Neve to Logic I get a great sound.  When I use either of the comps, I cannot get enough volume out of them without distortion and pushing the VU into the red. The difference in wave file renderings is dramatic in size. 

I'm certain its my fault so I'm here to get answers to a few questions this scenario has brought up.  First, I'll start with this one:If I leave the Neve at the volume I like and then go into either compressor, they can't take the signal apparently.  The VU needles slam up into the red zone.  Is it normal to have to turn down the mic pre when using compressors or is there something wrong?

Holiday wishes,

A

We would be shooting completely in the dark without knowing what levels you are talking about. In this forum the least you could do is measure some sinewaves with your multimeter and tell us exactly what you are seeing, and where you measured them. Just how "low" is the "too low output" from your compressors. There needs to be numbers, and not shady descriptions of "slamming into red zone".

Then we will talk about gain staging. Sounds to me like this concept is pretty much alien to you.

And what AD do you have. did you know you don't actually need to record at full blast to your AD inputs to capture the whole dynamic range from those compressors?

Kingston said:

We would be shooting completely in the dark without knowing what levels you are talking about. In this forum the least you could do is measure some sinewaves with your multimeter and tell us exactly what you are seeing, and where you measured them. Just how "low" is the "too low output" from your compressors. There needs to be numbers, and not shady descriptions of "slamming into red zone".

Then we will talk about gain staging. Sounds to me like this concept is pretty much alien to you.

And what AD do you have. did you know you don't actually need to record at full blast to your AD inputs to capture the whole dynamic range from those compressors?

Click to expand...

...Apogee converter. If I record without clipping the VU I have to boost the fader and then of course the noise to signal ratio isn't great.  The name is abecedarian by choice.  Ok, Ok,  thought I'd ask a simple question but as usual this place doesn't appreciate what mistakenly appears to be lazy man's requests. I assure you I'm not and by the way, I have donated money to this forum.  I'll get measurements and then we can talk about gain staging. 

LA2's in particular like to be hit fairly light unless you're looking for gobs of compression. They will distort easily if you turn up the output with a hot signal.

Try this with your 1176 using a signal like guitar or a vocal. Input on "30" output on "18", 4:1 ratio, attack 3, release 7. Turn down your mic pre all the way and insert it into the 1176. Bring up your level on the mic pre until you start to see some compression, 3-4db is plenty. If your A/D is calibrated for -18dbfs you should see a healthy level into your DAW.

Try it and report back.

Regards,
Mark

abecedarian said:

...Apogee converter. If I record without clipping the VU I have to boost the fader and then of course the noise to signal ratio isn't great.  The name is abecedarian by choice.  Ok, Ok,  thought I'd ask a simple question but as usual this place doesn't appreciate what mistakenly appears to be lazy man's requests.

Click to expand...

It certainly did come off as a pretty lazy question on a complex topic. Perhaps you're unfamiliar with providing data on the level required for electronics troubleshooting.

For example:

which apogee converter? (so I could quickly check if there's something unusual about this converter that might affect the situation)

what fader? (something within the DAW, or perhaps related to the converter or its drivers, or something on the analog side?)

signal to noise ratio where? within your DAW? Recording even 20dB below A/D clipping threshold is perfectly fine with a modern 24bit converter. It's unlikely that either the LA2A or 1176 would have their own S/N ratio above even 80dB. Recording them closer to the clipping threshold would achieve nothing since all you are doing is recording more noise at LSB's (least significant bits of the converter).

Just a general comment without knowing all your gear and circumstances, comps can have huge output level gains -eg the la2a. I know this has been mentioned above. Does your preamp(s) have a pad db switch on it? You may find lowering the preamp output signal gives you a bit more freedom with the comp without blowing your recording levels.

Also, and this might be obvious, but one off the most embarrassing experiences I had once was during a session where the audio was distorting perpetually and for the the life of me I couldn't work out what had become faulty in my outboard gear. Later in the afternoon I realized that the person who borrowed my gear the night before had set the apogee software settings to mic instead of line! My point?  - check the obvious before you embark on serious gain staging data

Indeed, we're all a bit in the dark re:

1.  What "Neve" unit are we talking about?  A vintage 8068 desk, or some sort of reissue/clone box?

2.  As was asked by Kingston, what Apogee, and what fader are you referring to?

Also, which make/models are your DIY compressors?

Bri

If you are set to VU on the meter it will definitely be slammed tracking hot to the average +18dBu modern input.  VU meters in that case are pretty useless unless they have an attenuator, which is not usually the case. 

Then I have to wonder why they are set on VU, rather than gain reduction. 

In a properly set up 24 bit system, a nominal analog operating level of +4 dBu (approx. 1.23 VAC) should result in a digital level somewhere in the range of -16 dBfs to -20 dBfs (the user picks their poison).  Some (but definitely NOT all) converters provide level trimmers to "tune" the digi operating level.  The Avid/DigiDesign 192, older (16 bit) 888, etc. systems come to mind.

Other converters force you to use some random analog operating level. since they have no analog I/O adjustments....which I consider as highly unprofessional.  But,  it's a reality.....

Bri

Brian, -16 dBfs to -20 dBfs is reading it at RMS levels though, isn't it and not peak?

canidoit said:

Brian, -16 dBfs to -20 dBfs is reading it at RMS levels though, isn't it and not peak?

Click to expand...

I was referring to using a 1 kHz sine wave at +4 dBu (the typical 0 VU reading from the mechanical meter) from the analog source and then setting the digi box to read -20 (or whatever) dBfs with that test signal.

Most any piece of pro analog gear can produce at least +20 dBu output levels before clipping, meaning there is at least 16 dB of analog headroom above 0 VU/+ 4 dBu output.  Thus, the analog source gear and digi destination both clip at around the same point.

With 24 bit gear, I've been using analog 0 VU/+ 4 dBu = -20 dBfs for a number of years.  0 VU = -16 dBfs (and "hotter") is sort of a hold-over from 16 bit digi gear.

Bri

Gentlemen, Thank you for taking your time to chime in.

Biasrocks said:

LA2's in particular like to be hit fairly light unless you're looking for gobs of compression. They will distort easily if you turn up the output with a hot signal.
Try this with your 1176 using a signal like guitar or a vocal. Input on "30" output on "18", 4:1 ratio, attack 3, release 7. Turn down your mic pre all the way and insert it into the 1176. Bring up your level on the mic pre until you start to see some compression, 3-4db is plenty. If your A/D is calibrated for -18dbfs you should see a healthy level into your DAW.
Try it and report back.
Regards,
Mark

Click to expand...

Mark, thank you very much for helping. I followed your instructions while speaking into the mic. Healthy level, yes, into D.A.W. was achieved.  Compression was averaging around 3dB but when I switched over to see what the output dB was, it was going far into the red often. 

emrr said:

If you are set to VU on the meter it will definitely be slammed tracking hot to the average +18dBu modern input.  VU meters in that case are pretty useless unless they have an attenuator, which is not usually the case. 
Then I have to wonder why they are set on VU, rather than gain reduction. 

Click to expand...

And I have to wonder why they are there then?  I thought in the red meant distortion even though I couldn't hear it myself. Thanks for bringing that up. 

Brian Roth said:

Indeed, we're all a bit in the dark re:
1.  What "Neve" unit are we talking about?  A vintage 8068 desk, or some sort of reissue/clone box?
2.  As was asked by Kingston, what Apogee, and what fader are you referring to?
Also, which make/models are your DIY compressors?
Bri

Click to expand...

1. 1064 with e.q. - not a reissue, original 100% (well, recapped,etc).
2. Duet, fader=in Logic
3. 1176LN rev D - I forget the LA2A revision number.

deuce42 said:

Just a general comment without knowing all your gear and circumstances, comps can have huge output level gains -eg the la2a. I know this has been mentioned above. Does your preamp(s) have a pad db switch on it? You may find lowering the preamp output signal gives you a bit more freedom with the comp without blowing your recording levels.
Also, and this might be obvious, but one off the most embarrassing experiences I had once was during a session where the audio was distorting perpetually and for the the life of me I couldn't work out what had become faulty in my outboard gear. Later in the afternoon I realized that the person who borrowed my gear the night before had set the apogee software settings to mic instead of line! My point?  - check the obvious before you embark on serious gain staging data

Click to expand...

A very good point.  I have four settings for the input on the Apogee;  +4dBu, -10dBV, XLR mic and instrument. The Neve has 20 to 80 -dBm output. I have been using the +4 dBu setting on the Apogee and the Neve for this test ended up at 35 and the -10 dB pad on the U87 was off. 

Kingston said:

...signal to noise ratio where? within your DAW? Recording even 20dB below A/D clipping threshold is perfectly fine with a modern 24bit converter. It's unlikely that either the LA2A or 1176 would have their own S/N ratio above even 80dB. Recording them closer to the clipping threshold would achieve nothing since all you are doing is recording more noise at LSB's (least significant bits of the converter).

Click to expand...

  If I recorded the output of the comps at an output value where the VU on the comps did not go into the red, as I thought it shouldn't, the ratio was not great.  Obviously, turning up the volume fader in Logic would result in magnifying the noise floor as well as the vox or gtr, etc,

abecedarian said:

Compression was averaging around 3dB but when I switched over to see what the output dB was, it was going far into the red often. 

emrr said:

If you are set to VU on the meter it will definitely be slammed tracking hot to the average +18dBu modern input.

Click to expand...

And I have to wonder why they are there then?  I thought in the red meant distortion even though I couldn't hear it myself.

Click to expand...

They are there because in 1970 we worked at lower levels, leaving headroom.  There were no commonly used peak reading meters adjusted to an obvious brick wall distortion ceiling like we get out of digital today.  In the red only means you are above the meter calibration point, which is likely arbitrary relative to your system.  I never have anything with a VU meter switched to VU, they are almost always slammed far into the red in today's world.  A fancy broadcast limiter might have a VU attenuator with 1 dB steps, for calibrated deviation from +4 standard, and with one of those you can recalibrate the meter to a different working reference level. 

abecedarian said:

...signal to noise ratio where? within your DAW? Recording even 20dB below A/D clipping threshold is perfectly fine with a modern 24bit converter. It's unlikely that either the LA2A or 1176 would have their own S/N ratio above even 80dB. Recording them closer to the clipping threshold would achieve nothing since all you are doing is recording more noise at LSB's (least significant bits of the converter).

Click to expand...

  If I recorded the output of the comps at an output value where the VU on the comps did not go into the red, as I thought it shouldn't, the ratio was not great.  Obviously, turning up the volume fader in Logic would result in magnifying the noise floor as well as the vox or gtr, etc,

Click to expand...

It won't "obviously magnify" the noise floor, but it's a very common mistake to think it does. We're no longer recording on C-cassette.

The dynamic range of your A/D: 100dB
The dynamic range of a common LA2A/1176 clone: 80dB

Like I was trying to explain, recording noisy gear (by modern standards) like your 1176 or LA2A blasted to your max A/D levels will only record more noise, compared to setting your input levels to something more sensible and not worry about clipping. "turning up the fader" in logic will not change anything within the noise floor.

The above specs mean that even if you have to "turn up the fader" +20dB in logic, you still have an optimal recording. I urge a quick test on your rig.

Which actually makes your +4dBu calibrated VU meters valuable. Trust them, and never worry about peak DAW recording levels.

Just a short addendum. The La2 and 1176 were designed to drive a 600 ohm load, so they need to be terminated in order to work and read correctly into modern equipment. A 1/2 watt 620R across pins 2/3 will do the job. Some folks make it switchable to get the proper loading when your running into a 600 ohm input.

Also, you may have to adjust your DUET'S input level to get optimal levels in your DAW. Try inputting a 1khz sine wave at +4 (0vu on your 1176) and adjust your duet until you get the input reading -18 on your DAW inputs. You should then be calibrated properly.

PAGE 17
http://www.apogeedigital.com/pdf/Duet2-UsersGuide-print.pdf

Regards,
Mark