Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Joseph D. Macry]

I installed a new sound system in an elementary school cafeteria. During commissioning, the design consultant set compressor/limiter in DSP at 100:1 ratio, +5dbu threshold. (The DSP is Biamp Nexia SP.) I thought the ratio was rather excessive, but hey he's the consultant.
Today, while investigating a report of distortion in the microphones, I found the comp/limiter hitting a few db of GR sounded rather ugly. (Had to get close and loud on the mic to hit the limiter.) So I eased back on the ratio to 20:1, which sounded much better. Still, I backed off the mic trims a tad.
The distortion was not reported by the users, but by the consultant. The school staff don't use the mic this loud.
Now the consultant wants me to put the ratio back to 100:1, but raise the threshold by "+10 or 15dB". I can do that, sure.

The question: What would you use for a limiter ratio, when the purpose is the protect the system from damage from transients?

[Ivan Beaver]

I installed a new sound system in an elementary school cafeteria. During commissioning, the design consultant set compressor/limiter in DSP at 100:1 ratio, +5dbu threshold. (The DSP is Biamp Nexia SP.) I thought the ratio was rather excessive, but hey he's the consultant.
Today, while investigating a report of distortion in the microphones, I found the comp/limiter hitting a few db of GR sounded rather ugly. (Had to get close and loud on the mic to hit the limiter.) So I eased back on the ratio to 20:1, which sounded much better. Still, I backed off the mic trims a tad.
The distortion was not reported by the users, but by the consultant. The school staff don't use the mic this loud.
Now the consultant wants me to put the ratio back to 100:1, but raise the threshold by "+10 or 15dB". I can do that, sure.

The question: What would you use for a limiter ratio, when the purpose is the protect the system from damage from transients?

The threshold could be anything-the individual numbers mean NOTHING-unless you know the rest of the system-like the rest of the gain in the system-the ratings of the loudspeakers and so forth.

I hope SOMEBODY realizes that by raising the output to +15 you are allowing TEN TIMES the power to go to the speaker system???

Is it rated for that-can the do it without clipping???/

There are bigger questions than the ratio.  I think something else is going on.

[Jason Lavoie]

I installed a new sound system in an elementary school cafeteria. During commissioning, the design consultant set compressor/limiter in DSP at 100:1 ratio, +5dbu threshold. (The DSP is Biamp Nexia SP.) I thought the ratio was rather excessive, but hey he's the consultant.
Today, while investigating a report of distortion in the microphones, I found the comp/limiter hitting a few db of GR sounded rather ugly. (Had to get close and loud on the mic to hit the limiter.) So I eased back on the ratio to 20:1, which sounded much better. Still, I backed off the mic trims a tad.
The distortion was not reported by the users, but by the consultant. The school staff don't use the mic this loud.
Now the consultant wants me to put the ratio back to 100:1, but raise the threshold by "+10 or 15dB". I can do that, sure.

The question: What would you use for a limiter ratio, when the purpose is the protect the system from damage from transients?

In the strictest terms, for a "limiter" you want the ratio as high as possible, otherwise it's just a compressor.

Jason

[Steve M Smith]

In the strictest terms, for a "limiter" you want the ratio as high as possible, otherwise it's just a compressor.

Surely, it's equally a function of the threshold.

If you have a signal averaging 0dB going in and the threshold is at +3dB, it's going to limit.  If the threshold is down to -20dB with the same signal, it's going to compress.


Steve.

[John Roberts {JR}]

Surely, it's equally a function of the threshold.

If you have a signal averaging 0dB going in and the threshold is at +3dB, it's going to limit.  If the threshold is down to -20dB with the same signal, it's going to compress.


Steve.

Sorry to disagree but no... a limiter with the threshold set for -20dB will limit at -20dB.

Compression describes reducing signal level changes proportionately reducing the gain X dB for every Y dB signal change above a given threshold so the sound is similar to before just less dynamic. Limiting describes reducing the gain whatever it takes to prevent the signal exceeding the limit threshold.

Compression is used as an effect to improve the sound quality. Limiting is generally used to protect gear from overpowering.

Of course this is the short description but close enough.

JR

[Steve M Smith]

Sorry to disagree but no... a limiter with the threshold set for -20dB will limit at -20dB.

Yes.  You're absolutely right.

I generally think of compression as something which is active all the time and limiting as something to catch the peaks.

I could be wrong though as I have been wrong in the past.  I think it was once in 1984 and again in 1997!


Steve.

[John Roberts {JR}]

Yes.  You're absolutely right.

I generally think of compression as something which is active all the time and limiting as something to catch the peaks.

There are different flavors of compression involving variation in behavior around the threshold (like soft-knee). There are also full range compressors that do not have a threshold (like 2:1 compression used in old DBX tape noise reduction encoder.)

I could be wrong though as I have been wrong in the past.  I think it was once in 1984 and again in 1997!


Steve.

I thought I was wrong once, but I was mistaken.... 

JR

[Joseph D. Macry]

To clarify:

The signal level in DSP peaks between +5 and +8 dBu, producing a pretty loud program in the school Cafetorium. The limiter was set at +5, so a loud voice was causing the limiter to kick in, and it sounded lousy. I found it to sound better with the milder 20:1 ratio, but the consultant wants to keep the stiff 100:1 ratio while raising the threshold up to where something unusual (say, a dropped mic) would be required to hit the gain reduction.

The block in DSP is called "Compressor/Limiter". A limiter is simply a compressor with a rather high ratio and a threshold set above normal operating range.

[John Roberts {JR}]

To clarify:

The signal level in DSP peaks between +5 and +8 dBu, producing a pretty loud program in the school Cafetorium. The limiter was set at +5, so a loud voice was causing the limiter to kick in, and it sounded lousy. I found it to sound better with the milder 20:1 ratio, but the consultant wants to keep the stiff 100:1 ratio while raising the threshold up to where something unusual (say, a dropped mic) would be required to hit the gain reduction.

Either the system needs to be limited at +5 or it doesn't. Turning down the ratio, or raising the threshold reduces the amount of system protection.

If it sounds bad from 3dB of limiting  (+8 dB - +5 dB) there may be something else wrong. In fact limiters do not have to sound bad, even when they are seriously limiting. 

The block in DSP is called "Compressor/Limiter". A limiter is simply a compressor with a rather high ratio and a threshold set above normal operating range.

Perhaps in their mind. While they both are similar in that they reduce gain above a threshold, limiters (for protection) generally use faster attack and faster release times than compressors which serve a somewhat different purpose (gentle sound shaping). 

If vocals are hitting the system limiter hard enough to sound bad, maybe turn down the vocal, or buy a bigger system that plays louder?

JR

[Mac Kerr]

If vocals are hitting the system limiter hard enough to sound bad, maybe turn down the vocal, or buy a bigger system that plays louder?

Yes, this.

[Joseph D. Macry]

Thank you for all the input.
Back to the original question: What do you usually use for a compression ratio, when the purpose is limiting for system protection?

[Ivan Beaver]

Thank you for all the input.
Back to the original question: What do you usually use for a compression ratio, when the purpose is limiting for system protection?

Very steep.

Otherwise it is not limiting, it is "allowing" some things to go through.

As others have said-there is more to the "sound" and how much, where, how, for how long etc than just the ratio.

Since you have a nexia-then why not have a hard limit (at the proper point) for system protection, and then add some compressors to help smooth things out a bit" on the vocal channels.

Then you have the best of both (assuming they are each setup properly-that is often a BIG IF).

[Jonathan Kok]

Very steep.

Otherwise it is not limiting, it is "allowing" some things to go through.

As others have said-there is more to the "sound" and how much, where, how, for how long etc than just the ratio.

Since you have a nexia-then why not have a hard limit (at the proper point) for system protection, and then add some compressors to help smooth things out a bit" on the vocal channels.

Then you have the best of both (assuming they are each setup properly-that is often a BIG IF).

Yup.

The point of the limiter on an output is 'system protection'. That's it. Short of spending a good chunk of change on an external limiter, it's probably not going to sound good, no matter which processor you're using, especially if you're pushing a lot of level into it.

You could, of course, use a smaller ratio set at a lower level to achieve the same 'system protection' results. For example, assuming your processor max input is +24dB, and your max output needs to be +4db, you could set a compressor for 2.1:1 at -20dB, and achieve the same system protection results. However, you're typically trying to maximize sonic quality prior to limiting. Any compression that hits the audio prior to the limiter is 'coloring' the sound.

Note the phrase 'needs to be'. Limiters aren't typically set at some arbitrary point between +5 and +15db; they're usually the result of calculations and/or measurements.

[Bob L. Wilson]

To clarify:

The signal level in DSP peaks between +5 and +8 dBu, producing a pretty loud program in the school Cafetorium. The limiter was set at +5, so a loud voice was causing the limiter to kick in, and it sounded lousy. I found it to sound better with the milder 20:1 ratio, but the consultant wants to keep the stiff 100:1 ratio while raising the threshold up to where something unusual (say, a dropped mic) would be required to hit the gain reduction.

The block in DSP is called "Compressor/Limiter". A limiter is simply a compressor with a rather high ratio and a threshold set above normal operating range.

Unless the specific block you are using in the Nexia "looks ahead" (by delaying the signal one or two microseconds) and I don't think that one does then a big enough fast enough transient like a dropped mic can and will still go through it and wreck stuff.

[Scott Carneval]

A limiter is to protect the speakers.  You should rarely be hitting the limiter.  If you're consistently hitting the limiter and it's affecting the sound quality then you need a bigger system.

I would recommend a 4:1 compressor on the mic INPUT to reduce peaks in the mic signal. 

THEN I would agree with the consultant that a 100:1 or even an INF:1 (if available) ratio on the OUTPUT limiter.  If it's just spoken word then I would use a short attack time and a relatively quick release time (maybe 50db/sec).  This is typically considered a peak limiter. If there is music playback, especially modern Hip-Hop or EDM, then I would use a very long release time.  This is typically considered a RMS limiter. 

[ProSoundWeb Community]