Digital recording snare transients

F

frott

New member
Hi -

I have been wrestling with getting a nice, present snare sound with my digital setup.

Unfortunately, everywhere I've looked for advice either talks about saturating tape which I don't have, or that digital distortion is to be avoided like the plague.

Therein lies my problem.

I use a single SM57 to record the snare (have tried a bunch of different positions) this goes straight into an Onyx, and from there to a delta 1010 and then audition on the 'puter.

The transient clips, so I lower the gain on the Onyx. Now the transient doesn't clip but the entire track is too quiet.


I hesitate to use my rack compressor because I've read that recording 24/96 digital gives enough to get a full image without doing so (this confuses me).

But when I do compress this is my signal chain:

mic -> onyx -> compressor -> delta 1010

And, unfortunately, the wave form is still clipping the transients even when by all metering on the onyx and compressor it should not be.

I've torn my hair out tweaking the tons of settings at my disposal but I can't seem to find a good balance between non-clipping transients and non-quiet everything else.


I don't actually hear any distortion, I just see the waveform clearly clipping at the transients.


The compressor is a dbx 166XL. Perhaps my settings are wrong on it? Everything coming out of the compressor seems correct but it still is not pulling the transient of the snare down enough on the waveform.

And again it still bugs me that I read that I shouldn't "need" the compressor to record a non-clipping, non-quiet signal.




Should this beasty transient be ignored? Should I be using a pad? Should I leave the recording as is and be using compression AFTER the wave is recorded?

Thanks in advance for any advice or for pointing me towards where I can learn about this.
 
1. Stop looking at waveforms.
2. A compressor removes the peaks of a signal. It would create a waveform that could be interpreted as "clipped looking."
3. Recording at 96kHz is a waste of computing speed. IMO YMMV.
4. If the recorded track is too quiet, turn it up. If need be, clone it, link the tracks and turn them both up.
5. You're using an Onyx mixer, with it's own converters and digital outs? If not, what are you using for a converter. If not, are you sure you are not clipping after the Onyx?
 
If you record your snare so it doesn't clip, but is still peaking anywhere above -6db, and it's "too quiet".... that means you have recorded everything else too loud.

The average recording level should be around -18dbfs (in your computer). Don't fall for the "record everything as loud as you can without clipping" BS, it's untrue and wll give you more problems than you can shake a stick at.


As far as compressing the snare: do that in the mix using a plugin..
 
Mix around the snare level. Get it so it's not clipping and adjust the volumes of the other stuff around that. Some compression or limiting on the snare track during mixdown is pretty common.
 
If you want to stay with the raw snare tone (more or less) use the clipper on the 166. You can likely catch the strays' and some amount of the tops w/o a lot of change in character.
The compressor section would be more adapt to shape changing I'd think.
Then there's the question of whether you're actually pursuing the 'loud fat percussion thing. This is not the sound of the raw snare.
 
Thanks for all of the replies.

I cleared out a lot of the cobwebs and used a modified "recorderman" / Glyn Johns setup for the drums and I'd just like to say that it is also a really good benchmark for how good your gear and setup can sound.

I've been caught by the "if only I had this or that gear/gadget" bug often, and it was really refreshing to finally do something quickly that had great results relative to the time investment.

Anyhow, the snare transient was still clipping and I'll try to address various points in the responses so far:

~~~

ermghoti - this isn't clipped looking so much as the transients literally go off the edge of the little graphic and the red light comes on signifying it was at 0(not sure how many consecutive 0db samples is needed for audition to trigger the light).

I understand that 32 (24) is good for dithering - well, that's about my extent of the knowledge - but why would 96 not be better. I mean, it's ~52k more!

It's an Onyx 800r for recording. Just twiddling gain knobs on it, and outputting to the delta 1010.

~~~

farview - Thanks for those numbers, I will try and see if there's a way to check that out.

I assume that "record everything as loud as you can without clipping" would be the act of essentially me playing the loudest part of the drums (finale fills and crashes, lets say) and basically gunning all of the gain so that the Onyx isn't showing any red.

Obviously when I play that back in the recording device I'll have to compress or lower everything because the tracks adding up will clip.


That's the part that's bugging me, I am recording the snare as hot as possible and while the onyx shows no clip the software does. If I'm supposed to not look at the waveforms, well, the software is :)


~~~~

Tex - I've read that you shouldn't need to compress snare pre-recording with 24 bit and you'll be able to capture it fine without clipping. Why is this possible with 24 but not 16?

I'm comfortable with using compression to change sound qualities subtly or bring things in and out of the mix but was getting frustrated trying to squash the transients to make the software happy.

~~~~

mixsit - good idea, I'll give it a shot. I have that eerie feeling that there's something simple I'm missing (somewhere a -10 and +4 are confused and I'm not catching it). I'm happy to blame Audition as well :p as justification to finally grab a copy of cubase.




Anyhow, thanks all for the Q&A -- if there are any sources I should go to for a general overview (besides here, of course) post them here. Its just very easy for some exception or specific issue to not be addressed and lead to confusion, where I'm randomly poring over PDFs and guides, etc.
 
Tex - I've read that you shouldn't need to compress snare pre-recording with 24 bit and you'll be able to capture it fine without clipping. Why is this possible with 24 but not 16?

I'm comfortable with using compression to change sound qualities subtly or bring things in and out of the mix but was getting frustrated trying to squash the transients to make the software happy.
Click to expand...

It doesn't matter what bit depth you use, just turn down the preamp gain until the snare no longer clips. You are making this way too hard.

96khz sampling rate is "better" since it can record high frequencies up to 48khz. Chances are this is much higher then your mics, preamp or monitoring system can even produce. If you have enough computer power then keep it at 96k but once you add a bunch of tracks the project is going to get bogged down pretty quickly.

48khz sampling rate @ 24bit is pretty standard for professional quality recording. If you want to try a different DAW program then check out Reaper www.reaper.fm
 
frott said:
I understand that 32 (24) is good for dithering - well, that's about my extent of the knowledge - but why would 96 not be better. I mean, it's ~52k more!
Click to expand...
You seem to be confusing bit depth with sample rate. You should record at 24 bit. There is no compelling reason to record at 96k, all that allows you to do is record frequencies that you can't hear and use twice as much disc space.


frott said:
I assume that "record everything as loud as you can without clipping" would be the act of essentially me playing the loudest part of the drums (finale fills and crashes, lets say) and basically gunning all of the gain so that the Onyx isn't showing any red.
Click to expand...
Yes, but I'm telling you NOT to do that!

frott said:
That's the part that's bugging me, I am recording the snare as hot as possible
Click to expand...
Stop that!
frott said:
and while the onyx shows no clip the software does. If I'm supposed to not look at the waveforms, well, the software is :)
Click to expand...
You need to be watching the meters in the computer. The onyx obviously has more headroom than the Delta does. And you are recording way too hot.
frott said:
Tex - I've read that you shouldn't need to compress snare pre-recording with 24 bit and you'll be able to capture it fine without clipping. Why is this possible with 24 but not 16?
Click to expand...
With 16 bit, the noise floor is 48db louder than with 24 bit. That means you have to record hotter to stay above it.
 
In a nutshell, you need to turn down the gain in the Onyx until your loudest snare hits no longer reach up to 0dB in your software. Really you don't want it anywhere near 0dB. Don't worry, your snare will be plenty loud. Recording as hot as possible is the wrong idea in digital recording. It's kind of a hold over from when everyone recorded to tape, which could sound good when you slammed the heck out of it and it had a high noise floor to overcome.

If you like the sound of the compressor, then use it. But don't use it to control the peaks in the software. This would kinda (but not exactly) be like pressing the gas (gain) and brakes(compressor) at the same time. Just lay off the gas (gain).

Also, don't bother buying any more gear to fix this problem. Save it for something else.
 
frott said:
I use a single SM57 to record the snare (have tried a bunch of different positions) this goes straight into an Onyx, and from there to a delta 1010 and then audition on the 'puter.

The transient clips, so I lower the gain on the Onyx. Now the transient doesn't clip but the entire track is too quiet.
Click to expand...

Are you using overheads as well?

How does the snare sound in them?

You should be able to get a nice snare sound by getting the balance right between the 57 with the overheads, along with some parallel compression for more punch
 
Thanks for the continued advice --

Farview,
yes, I understand the diff and will be pulling things down to 48.
i understand to not gun everything.
Thanks for the info on the noise floor. I "knew" that but didn't recall it until reading it again. I just wanted to make sure I wasn't erring in assuming the bigger number is always better.

Tex,
Yep, I'm really interested in Reaper. I was a fan of Cool Edit and honestly I feel Adobe did a hatchet job with functionality - rather than beefing up the interface and adding customization options, they made it their "here's how you get sound into premiere" package. No thanks!

Philoe,
Thanks - the drum miccing setup has me jazzed to invest money in a 3rd large diaphragm condenser. I had been using a budget pair in XY and they were lacking to say the least.

Bulls - Well, the precise way I was doing it had the overheads. Those didn't pick up as much of the transient and with the recorderman setup it was pretty easy to pull everything together.

Note: Everything sounds OK as far as I can hear (let's not go into drummer's ear) but red lights blinking at me just made me think something was wrong.

I'll record a 4 mic sample and post it sometime this weekend.
 
To repeat what everyone seems to be trying to tell you: Don't record the snare "as hot as possible". There's no reason to peak any higher than -6. If it's then "too quiet".....Turn it up....If you have no room to turn it up, turn everything else down.

Pretty simple really.
 
Apparently the 1010 has a button on the back to switch between +4dbu and -10dbV. You want the button out for +4, which I believe is what your Onyx should be sending out if you have everything cabled up properly. If you are stilling clipping at the A/D, then you may need to get an inline pad for your mic. Not unusual on drums at all. Chances are, your preamps will sound a little cleaner if the signal coming in from the mic does not approach the spec'd clipping point.
 
Its a calibration problem between your mixer and soundcard. get a professional cd player with test tones and make sure that what your mixer is showing is the same as your interface. I assume your taking direct outs on each channel, and therefore its a level problem between your equipment (mixer outputting "+4dB" while interface is expecting "-10dB" ). Check in the delta control panel (the page with the samplerates) and make sure its set correctly
 
You must log in or register to reply here.
Back
Top