Shure FP42 Mixer Arrives

[GaryNattrass]

Ha ha we do still use analogue levels in TV and that is why PPM's are still used here in the UK. 0 or PPM 4 for line up and +8db or 6 PPM for max! ;0)

 

[Noiz2]

Levels agin, huh?

Actually actually dbFS is different that analog dbVU. First off they should really be written as dbFS16 or dbFS24, etc. As they vary depending upon the bit depth. What 0 dbFS is is the point where you sound level is using all the bits available. If you think of it you can record a louder sound with 24 bits than you can with 16 bits, and FS means Full Scale, so 0 dbFS is the maximum level you can record period. Levels above that do not clip, they do not distort, they just aren't there at all. In effect, they disappear.

On the other hand dbVU is measured at a reference level, because in general you do not know how much headroom your analog circuit has, so a safe reference level has been chosen for 0 db. Which is about 4 db down from the worse case setups, really good setup may be able to record fine at +20 dbVU which is the equivalent of 24 bit digital, by the way.

OK, those signals brought to a comparable level 16 bit is -12db, and 24 bit is -20db, which are pretty much equal to 0dbVU. The conversions are not exact but are plenty close enough to be considered the same in the real world.

I do not understand why in the world they did not all use the old analog reference and just say that 16bit has 12db of headroom and 24bit has 20db of headroom, and not confuse everyone. Well, I actually do understand why. Jargon is used to to confuse those who are not in the know and prove to yourself you are better than they are. That is why kid's gangs love to use made up or misused words. In other words, childish nonsense.

Some of this is misleading at best. 24 bit is not louder than 16 bit and -20dBfs is the same level with both. Digital fs metering is a very different thing than analog VU and digital recorders have very different needs that analog recorders.

Analog meters are about measuring the average average "loudness" of a signal because analog recorders like a strong average. Plus they have a soft top whee clipping starts a very soft and sometimes desired distortion. And analog tape naturally acts as a compressor near the top. So you want a strong signal and you are not too worried about clipping because you have to hammer it pretty hard before its bad. Plus you have tape hiss so you want to always be hot enough to mask that.
Mic pres for analog didn't need to be super quiet, just quieter than tape the hiss. Everything below that would be masked by the inherent noise in the system.

Digital meters are all about instantaneous peak level. They usually also do some integration so you get a feel for average level also but their main goal is to tell you if you ever hit zero. Digital recordings have no wiggle room, at zero you are done. There is no natural compression and the slightest over sounds really bad instantly. They don't care about average signal, they don't need a minimum level to overcome tape hiss. The worst that happens when you get too low is quantization issues, assuming your mics and preamps are quiet enough. Quantization is an issue if your signal is too low, effectively lowering your bit depth and creating a low res gritty sound. This BTW is why it's not a great idea to record really low and then try and boost it in post.

With analog you calibrated with tone zero dBvu, but a professional mixer could handle +25 dBvu or so. Theoretically your tape would start saturating at zero dBvu but it's a soft top so hitting +6 - +12 was not often a problem.
Which is why digital is calibrated at -20 dBfs. There is no "+" in a dBfs scale so setting your "zero" at -20 gives you a similar headroom to analog. But a lot of this is to accommodate old gear and old workflows. If, as is usually the case these days, you are not going analog till you play back over speakers then you are largely free to record at any level that doesn't clip. The exceptions are when you are working for a place that has delivery specifications for your recordings. Studios probably still tone digital files because it's part of the work flow and makes it easy to check your signal flow and levels. Back in the day it was really pretty critical so that the system could be calibrated since different studios often had somewhat different preferences so having a zero reference was important. In a digital fs world zero is always the same.

We could go way down the rabbit hole with metering systems. Averaging meters come in lots of flavors with different integration times and scales. The PPM meter that is mentioned looks like a VU meter but has as low an integration time as a mechanical meter can have and is closer to a digital dBfs meter in practice than a VU meter.

BTW since it might have gotten confusing, while the tone is calibrated to -20dBfs you are looking to have your average signal on the recorder at -12 or so with peaks at -6 to -3. The -20 is get the analog gear set up to work well with the digital recorder.

 

[Alex H.]

... so 0 dbFS is the maximum level you can record period. Levels above that do not clip, they do not distort, they just aren't there at all. In effect, they disappear.

Not true.

Digital recordings have no wiggle room, at zero you are done.

And "done" means digital clipping. Try to push your digital levels past 0: any data above what the signal can hold is simply sheared off, effectively creating a square wave. This digital distortion is in no way pleasant, like analog distortion can be in some cases, and there's no way to repair it.

 

[drapeama]

My question was, with a clean signal (from a good preamp), having the audio at an average of -20dBFS and peaking a little louder, can we bring that average (with a compressor) to something more audible in post? Will the quality be as good as if it's been recorded around -12dBFS?

I assume that, this previously asked question (asked with in mind, a Shure FP33 with a very decent preamp) can be answered with this:

With analog you calibrated with tone zero dBvu, but a professional mixer could handle +25 dBvu or so. Theoretically your tape would start saturating at zero dBvu but it's a soft top so hitting +6 - +12 was not often a problem.
Which is why digital is calibrated at -20 dBfs. There is no "+" in a dBfs scale so setting your "zero" at -20 gives you a similar headroom to analog. But a lot of this is to accommodate old gear and old workflows. If, as is usually the case these days, you are not going analog till you play back over speakers then you are largely free to record at any level that doesn't clip.

BTW since it might have gotten confusing, while the tone is calibrated to -20dBfs you are looking to have your average signal on the recorder at -12 or so with peaks at -6 to -3. The -20 is get the analog gear set up to work well with the digital recorder.

So, I'd be using the Mixer/Preamp (Shure FP33) slate to set the level at -20dBFS on the recoder (Tascam DR-40), and when recording audio, have the average signal at -12dBFS?
If the average signal is lower than -12dBFS, let's say peaking at -12dBFS, can it be "boosted" in post just as well, or it's preferable to have it recorded at the right level, even if the headroom is smaller?

 

[graywolf]

Not true.



And "done" means digital clipping. Try to push your digital levels past 0: any data above what the signal can hold is simply sheared off, effectively creating a square wave. This digital distortion is in no way pleasant, like analog distortion can be in some cases, and there's no way to repair it.

And do any of you see that we are saying the same thing? Not there, done, digitally clipped all mean there is nothing above 0 dbfs.

And I am correct -12db 16bit, and -20db 24bit is approximately the same as 0dbvu. It really does depend upon your bit depth!

My notions are only based upon a BSEE and 55 years of experience, plus double checking them.

 

[graywolf]

drapama, none of this actually answers your question exactly.

There are two points of reference that audio engineers use. Peak, and Average.

Peak is used with a bit of safety factor to make sure the loundest sounds do not distort. If your peaks are 3-4db down from that point you are pretty safe.

Average is the working level, for me it is usually a converstational voice, for a music recordist it is somewhere in the middle of the mix, although the best guys work to bring the softest notes to a base level and let the rest fall where it does unless the peaks are dangerously high, in which case they will use compressors and limiters. When you listen to stuff mastered by those guys you do not need to touch your volume control once set as the levels all match perfectly.

Now, what I think these guys here are trying to say, is that a VU meter tells you nothing about your equipment. But familiarity gives you a good idea of the setting you want for what. Digital meters tell you exactly what your digital equipment is going to do. Note that if you are feeding that digital signal into a piece of analog gear you are back into that best guess territory. It winds up being a combination of theory and experience that works best.

BTW, that -18/-20/-24 stuff actually refers to below the equivalent of 0dbVU, and is for that average recording level. It is altogether different from the equals 0dbVU stuff. transom.org has a fairly good article about this.

 

[Noiz2]

And do any of you see that we are saying the same thing? Not there, done, digitally clipped all mean there is nothing above 0 dbfs.

Well that is not quite how you put it

... so 0 dbFS is the maximum level you can record period. Levels above that do not clip, they do not distort, they just aren't there at all. In effect, they disappear.

I think the issue we have is with the later portion. On one very lawyerly level you could argue that when you say no distortion you are referring to the non existent over zero sounds. But on a real world level when we are talking about distortion it is the recording we are talking about and a signal that tries to go over zero on a digital recorder is VERY heavily distorted, just as a totally saturated analog tape will be very distorted. The point is analog eases into saturation over zero and digital is completly distorted AT zero.

And I am correct -12db 16bit, and -20db 24bit is approximately the same as 0dbvu. It really does depend upon your bit depth!

I think you are mixing systems. dB is a referential system Digital scales are FS and reference zero -20dBfs to zero dBfs is the same relationship no mater how finely you slice the steps which is what bit depth is. 0dBvu is not a digital scale so not really relevant to digital signals. Plus it is generally calibrated to a reference tone and so when it was in common use in analog it didn't universally refer to a specific voltage, it was referenced to the calibration tone.

My notions are only based upon a BSEE and 55 years of experience, plus double checking them.

You got me beat by a few years but I'm not sure you went through the digital transition era. I learned on analog and was working regularly when everything started changing to digital. It gets very confusing when you are thinking analog but dealing with digital. All of the old references to a moving target (calibrating your deck and mixer etc to the tones on the incoming tape) are really gone. That some people still do it is habit and policy. You can "fake" analog with a digital system and pretend that you are living in an analog system but under the hood it's all digital. There are no voltages to calibrate to or adjust. Over some voltage you have a "1" and under it you have a "0" and that is your sound, a long string of 1's and 0's. When you use up all your bits you have nothing you can salvage.

Things may change. Stuff like bit stream recorders can be more "analog" like in how you work with them so?

No offense, remember I'm only a few years younger, but a BSEE 20+ years before personal computers is not necessarily a plus. I'm a tad younger but when I was in high school the electronics classes were still teaching students how to build tube circuits. Ten years later the university was still using punch cards and the computer I built to control my keyboard was programed in hex. It's another fifteen years till digital recording comes into play much. So while I fully respect the knowledge and effort that degree represents, it also has very little to do with digital signals. I'm saying this based on how much what I had grown up with didn't help when digital recordings and editing came in.

 

[graywolf]

Once again you are making assumptions. I was originally trained by the Air Force as a radar/computer technician. I was into personal computers before they were called personal computers. I was programming computers before there was a microprocessor. (That's kind of out of order)

There is no distortion of the digital signal, you cannot distort bits. The distortion comes when you convert the digital back to analog. And if you are dealing with an analog/digital setup like I, and apparently drapama are, you have to figure out how to interconnect them. In purely digital setup you can ignore VU, but not in a combined analog/digital setup.

 

[Noiz2]

Um... In a very technical way you are correct. The digital signal compared to itself is not distorted. The digital signal compared to the original sound however is very distorted.

And that is really the point because nobody really cares what the digital signal looks like as long as in the end what comes out is the same as what goes in. And so you don't try to split that hair I'm talking of what comes in as the original source sound and what comes out is what ever playback you can actually hear.

My apologies about assumptions you predate (pre 1968 programing experience) what I assumed, but again that is not as useful as it might seem.

Everyone is dealing with an analog/digital set up, that is not the issue. And you most certainly can ignore VU. VU is kind of pointless in a system that ends up on a digital recorder. It was designed to monitor radio modulation and that happens to work well for analog tape, it's pretty pointless in digital and it is certainly not needed for an analog mixer, except if that mixer is headed to an analog tape deck. VU with out a peak monitoring FS meter is a disaster waiting to happen. A PPM is almost as useful as a FS meter. The problem with VU going to a digital recorder is that it can be way off what the FS meter reads and that difference is frequency and transient dependent so it's not even just an off set, you really have no idea what the recorder is seeing.

To get back OT though. What I use is not super far from what you have. The MixPre uses dBu metering and there is some integration time so while I have an idea of what the signal is I could not be sure I didn't clip by just using those meters. But the MixPre also has great limiters, so if you set up so that you can't clip and lock down the recorders levels you can use the MixPre meters to have a feel for how hot the signal is and the recorders meters periodically to see where you peaked.

You can do basically the same thing with VU meters. If you don't have good brick wall limiters though you will need to record lower so you have a bigger safety margin. I would probably just use the digital meters but they are harder to read at a glance. Basically you can think of the digital meters as the only ones that count and the analog ones as indicators of about where your signal is.

 

[Rick R]

Not to get more off topic and confuse folks, but when I was a music recording engineer in the 80s (not to reveal my age), we used to set the tape machine's VU meters to MRL calibration tape and go 6 and sometimes 9dB over the 185 and 250 nanowebers flux level, so the meters where... read lower, but the tape was getting hammered. (tape saturation sounded great on some instrument) Can't really do that today, but there's a lot of plug-ins that try to duplicate that sound.

 

[GaryNattrass]

Not to get more off topic and confuse folks, but when I was a music recording engineer in the 80s (not to reveal my age), we used to set the tape machine's VU meters to MRL calibration tape and go 6 and sometimes 9dB over the 185 and 250 nanowebers flux level, so the meters where... read lower, but the tape was getting hammered. (tape saturation sounded great on some instrument) Can't really do that today, but there's a lot of plug-ins that try to duplicate that sound.

When I worked at AMS Neve in the early 90's I was once told that when digital recorders such as the Sony Dash machines came out the engineers used to do the same and lay down tracks to the max on the digital full scale, that is why the 0db reference level changed from -18 to -20 to allow two db of headroom on the metering scale to 0dbfs.

 

[drapeama]

... that is why the 0db reference level changed from -18 to -20 to allow two db of headroom on the metering scale to 0dbfs.

Interesting!

As to get back to using an analog mixer with a digital recorder (if I understand correctly) use the slate function (mixer) to set the recording level (recorder) at -18/20dBFS is a good start.

 

[graywolf]

Well, what I do, now that I have the mixer, is use the oscillator to set 0db on the mixer, and -012db (16/48) on the camcorder. Turn off the oscillator. Then I use the mic gain control on the to give me -20db on the VU meter for average conversational level talking. That setting is going to change with the particular mic, distance, and how loud the persons voice is. The old "Testing 1 2 3" (grin). Then I turn on the limiter, just in case.

That gives a clean signal across the normal recording range on my DV tape. And each tape has pretty much the same audio level.

In post mix, I can adjust things more exactly, but I find that I often do not have to adjust it in post.

Now this is all (video) pretty new to me, but I do have a background in sound, photography, and computers that help out a lot. Plus I have a good background in the theory of all that. I certainly do better than those guys on youtube whom you cannot hear over the background noise, or those whose final mix is at tracking levels (way too loud).

It works, so even if it is different than how others say to do it, it is not wrong.

 

[drapeama]

Well, what I do, now that I have the mixer, is use the oscillator to set 0db on the mixer, and -012db (16/48) on the camcorder. Turn off the oscillator. Then I use the mic gain control on the to give me -20db on the VU meter for average conversational level talking. That setting is going to change with the particular mic, distance, and how loud the persons voice is. The old "Testing 1 2 3" (grin). Then I turn on the limiter, just in case.

That's pretty much what I had in mind. My only concern was if the average sound is too low (below -12dB), would it introduce noise when adjusting in post?

In post mix, I can adjust things more exactly, but I find that I often do not have to adjust it in post.

Is -18dB/20dB too low for an average audio? Haven't tested it yet, just wondering if it would be better to use the limiters and get the average signal a little more hot?

It works, so even if it is different than how others say to do it, it is not wrong.

Yes, as long as it works for one.

 

[graywolf]

What I am doing, drapeama, is setting things to zero gain, then adjusting the input for sound level I am recording.

Thing to remember is that you would normally consider +12db as the clipping level in analog -30db is just above the noise floor. In fact Audacity shows my digital noise floor to be -42 to -45. Also remember that my old Sony PDX10's only have 16bit sound.

With 24bit you will have even more room. Think of really soft dialog as being about -30db, which leaves it at least 12db above the noise floor. That means 4x louder than the noise floor. Now if you are dealing with room noise, you are talking about something else entirely, I am talking about the electronic noise the equipment itself makes.

Sound (room) noise has to be dealt with some other way because it is coming in the mic. If it is very regular, like a fan, you can filter it out in post (not the best way to deal with it), but if it is irregular you have to get rid of it before it hits the microphones (room treatment, turning things off, etc).

The best way to figure all this out is by intelligent experimentation.

 

[Noiz2]

drapeama, as was said if it works for what you are doing then don't worry about it. Speaking strictly from a perspective of the meters on your recorder, so dBfs, averaging your levels at around -20dBfs I would consider low. But it also depends on your dynamic range. If your "normal" level is -20 dBfs and your peaks are at -6 dBfs then you are about right you just have very dynamic content that you will probably need to compress in post.

Assuming your analog chain is not particularly noisy boosting the level isn't going to add a lot of "noise", but it will add it in proportion so balance between dialog and noise is going to stay the same.

However every 3dB you don't use effectively reduces your bit depth by 1. If you are recording your average dialog at about -20 dBfs then you are down to 10bit resolution, assuming you are starting at 16 bit. Which is another reason to record at 24 bit BTW. The difference between your peaks and -3 dBfs is essentially waste. The reason I used -3dBfs and not zero is that when the digital signal is changed back to an analog one you can have "over shoot" at the peaks that can be as high as 3 dB over what the digital meter read. You want some buffer zone between -3 dBfs, unless you have limiters that are really good, but you don't want to waste a lot of range for no reason.

Which wraps back to what I posted awhile back. Generally it is this low effective bit depth that is the problem boosting low signals not so much noise. The lower the bit depth the more stairstepy the sound is. It starts getting "gritty" and digital sounding. How big a problem this is depends on a lot of factors including what your final product is. Generally for dialog you want as much resolution as possible. First because people are very aware of what the voice sounds like and will pick up on the low bit depth faster than some other sounds. Related is that usually dialog is giving the audience crucial information so they are even more focused on the dialog. And secondly a lot of time dialog gets a fair amount of processing in post. The levels get adjusted and some EQ gets put on, maybe some compression and de-essing or noise reduction, etc. All of these things in the digital realm are math operations and the more data you start with the better the result is.

For your particular case you will probably, like greywolf and I, be looking mostly at your mixer meters because they are easier to keep an eye on. SO you really have to just go out and use the system a bunch till you get a feel for how the bouncing needles relate to the recording levels. With a mixer designed to go to analog tape any limiters it has are probably not fast or steep enough to help a lot with a digital recorder. And most affordable digital recorders have not great to completely useless limiters. Tascam's are not the worst but are almost certainly digital so won't really save the day. My main recorders built in limiters act more like an AGC and so are a disaster and I never use them. But the only way to really find out is to go out and do some recording.

With out bullet proof limiters you are going to want to record lower to have some safety margin, but how much you need depends on what you are recording. A sit down interview is not likely to have any sudden loud sounds (that you want) so you can probably push higher, other things you may have to go lower because they are very dynamic.

 

[graywolf]

You are sitting in your studio before your 128 channel digital mixer board, in the sound booth is your current musician playing his instrument. You will do this separately with each instrument in the band and maybe using a dozen mic's on that one instrument. You want to track that at the maximum safe level you can, the highest peaks ought to almost touch that 0dbFS line, but not actually touch it. You gate, limit, compress, equalize, and reverb everything just the way you want it. Everything will be fixed just right in the final mix.

Does that sound like how you want to record your film dialog into the camera? Maybe? If you have a Hollywood budget.

 

[Noiz2]

Does that sound like how you want to record your film dialog into the camera? Maybe? If you have a Hollywood budget.​

​

? Not sure what that relates to. That is not how a film with a budget would shoot and it doesn't look like any music studio I have seen.

I must be missing something because I don't see how it relates to the topic.

 

[Noiz2]

We have probably gotten too technical for the question.

As to levels what I do is:
Production work/ dialog.

I use tone to set the recorder so the MixPre's limiters will prevent clipping and I try to record fairly hot, so my normal dialog is probably higher than -12 but not a lot.
My priority is not clipping and getting the most usable signal I can. I do a LOT of post so I have a pretty good idea of what will work and what will be a problem. In tracks I get low levels are not the worst problem, though at times they have been. The worst problems come from recordists not paying attention to sounds other than what they are trying to record. All that "other" sound is a major problem with a lot of low budget tracks.

FX recording.

Here you almost always want to be as hot as you can get, and you often know in advance if there will be any loud sounds so you can really push it. The exceptions are things that might have a surprise sound in them and that you won't get a second chance at. In that case I will go with a production setting.

Concerts, live shows.

If I can be there for a final rehearsal I will set levels then. Sort of like dialog/ production I will have the recorder set so that the limiters are clip proof and I'm peaking as close to -3 as possible with out the limiters activating.

For all things I try to never go above -3 dBfs. I also try never to change levels during a recording. Both of these practices are counter to traditional production mixing where you would be aiming at closer to a -6dBfs peak and mixers actually mix. But both are based on my post experience in the digital world. I would always prefer to lower a tracks level than boost it. Having levels change in a track is a real PITA in post. It is done traditionally for a couple of reasons. First most recorders were mono till shortly before DAT. Stereo Nagra's were more of the FX recordists tool, though they became more popular just before DAT came in. With few tracks you have to mix if you have more than one mic. And the production tracks were used raw for the dailies so they needed to sound decent or you were out of a job. On a real film the production mixer these days is probably using an 8+ channel recorder and they are recording individual tracks for every mic and then mixing to "mix tracks" for dailies. This way the dailies get a decent mix and post gets tracks with levels that aren't a moving target.

Also since the mixer doesn't have to fight with a hefty noise floor you can let tracks get quieter than you could with analog.

So that is me, it's not particularly standard except FX probably. I mention limiters a lot because I use them pretty much always but I try to never hit them. I use them as a safety net not as a level tool.

 

[drapeama]

Assuming your analog chain is not particularly noisy boosting the level isn't going to add a lot of "noise", but it will add it in proportion so balance between dialog and noise is going to stay the same.

Ok, that sounds good to me. I as afraid to record too "safely" and boost in post where it would add too much noise. But I'll try to find the sweet spot between recording hot enough and adjusting in post.

With out bullet proof limiters you are going to want to record lower to have some safety margin, but how much you need depends on what you are recording. A sit down interview is not likely to have any sudden loud sounds (that you want) so you can probably push higher, other things you may have to go lower because they are very dynamic.

From what I've tested quickly, the FP42's limiter aren't best, especially compared to the FP33, hence the price. I've tested the FP33's limiter with the DR-40 and even with very loud and aggressive sound, the recorder won't peak. Sound is crushed by the limiter (this was an extreme wind blowing test) but the recorder gets -3dBFS or -6dBFS, I don't remember the right number. All that to say, that on a "normal" recording (dialog), the limiter should be plenty good to get a clean track on the DR-40 as it doesn't sound limited like with a brick-wall limiter.

I use tone to set the recorder so the MixPre's limiters will prevent clipping and I try to record fairly hot, so my normal dialog is probably higher than -12 but not a lot.

So, the average sound/dialog should throttle around -12dBFS on the recorder. I'll aim for that, knowing that the limiter will activates without ruining the audio.

If I can be there for a final rehearsal I will set levels then. Sort of like dialog/ production I will have the recorder set so that the limiters are clip proof and I'm peaking as close to -3 as possible with out the limiters activating.

Last time, using the feed from a console (without the mixer) I asked the sound guys to feed the music at it's higher volume, so I set the camera's input accordingly. Then I compressed in post and it was pretty good that way.

For all things I try to never go above -3 dBfs.

Good advice, always trying to record on the edge, but never above that. I would have play safer, like -6dBFS. I'll experiment, trying to know the gear a little better!

Also since the mixer doesn't have to fight with a hefty noise floor you can let tracks get quieter than you could with analog.

Depends, the FP42 has a little noise floor passed 7/10 on the volume knob. Below that, it's pretty good, but above that you can hear it easily. Sure, noise reduction can be applied, but still.
I wonder with a console feed if I would have to "push" it up to 7/10 though. My guess is that 5 or 6/10 would be pretty sufficient.

I mention limiters a lot because I use them pretty much always but I try to never hit them. I use them as a safety net not as a level tool.

I understand. Always nice to know there's a safety. After a couple times, you get used to know how to set the gear, so you don't even have to rely on it.