Topic: Bought high ohm headphones, and realized my recordings are low...

[Brian E.]

So I usually master on my Shure SE530 earphones, which sound great - but they're 30ohm or so, so they are easy to drive.  I just got some Sennheiser HD650's, which are 300ohm.  Listening to them with an preamp with studio recordings and then switching to my recordings shows they they are very low on volume in reality.  I'm wondering if there is anything to be done, or should I just leave them?  I wouldn't want to replace all my archive.org stuff, nor should I.  But I'm a bit embarrassed that seemingly it's very hard to drive these recordings on large phones.....

:sigh:

[bugg100]

Well, what you have proven is that your new cans are underdriven by your playback gear.... Nothing here speaks of the levels of your shows..... Your cans, your playback, both have zero to do with levels of your flac/wav files.  For that check RMS average levels in wavelab or the like.

 But you really shouldn't squash the dynamic range out of your masters just so somebody doesn't need to use their volume control.....

[Brian E.]

trust me, they're not underdriven.  I'm using a NuForce Icon HDP.  It has more than enough power.  I'm just noticing that my studio albums are much louder than my recordings at the same level.

[rastasean]

That's usually the case. Do you have a live recording released by a band and a studio album? It is obvious that the studio will be much louder.

[StuStu]

Loudness Wars

[page]

trust me, they're not underdriven.  I'm using a NuForce Icon HDP.  It has more than enough power.  I'm just noticing that my studio albums are much louder than my recordings at the same level.

If the only change was the headphones, then either your RMS values were really low to start with (anything more then -20RMS would probably qualify) and you were turning up the volume on your headphones (or they were really sensitive outputs), or it really is that the senns don't have the juice (which a lot of people complain about). My westone um3x were blazing hot sensitivity while my denon 7000s arn't nearly as much but they have the similar low-ohm ratings. I just turn the amp up. 

[Belexes]

Loudness Wars

Agreed.

The recordings are likely right where they should be.

[page]

trust me, they're not underdriven.  I'm using a NuForce Icon HDP.  It has more than enough power.  I'm just noticing that my studio albums are much louder than my recordings at the same level.

If the only change was the headphones, then either your RMS values were really low to start with (anything more then -20RMS would probably qualify) and you were turning up the volume on your headphones (or they were really sensitive outputs), or it really is that the senns don't have the juice (which a lot of people complain about). My westone um3x were blazing hot sensitivity while my denon 7000s arn't nearly as much but they have the similar low-ohm ratings. I just turn the amp up. 

Shure SE530: 119 dB SPL/mW
Sennheiser HD650: 103 dB SPL/mW

Yeah, you have less sensitive headphones, just turn up the amp (unless as mentioned earlier, your RMS values are really low).

[Brian E.]

I know I have less sensitive headphones with the Senn's vs. the Shures.  That's not the issue.  I'm saying the 530's are good around 11A to 12P on the dial, and the Senn's are good around 2-3PM.  But with my recordings vs. silver discs, my recordings have to be turned up all the way on the dial and it's still not as loud.  I feel like I need to increase the volume or something.

I'll try a few more.  The one thing that tipped me off to it was a friend saying the recording was too low when I sent him one 6 months ago.  I shrugged it off, but now I'm wondering if he had lower sensitivity phones than what I'm used it with the IEM's - and if he does, most people probably do, which maybe means I should master them higher.

[Chuck]

Can you post one of the WAVs somewhere?

[bugg100]

Or a mp3 sample...

[Brian E.]

http://www.archive.org/search.php?query=taper:%22Brian%20Emerick%22&sort=-publicdate



maybe it's just all in my head.  Or maybe the recordings I was testing out were just far away to begin with and I didn't increase volume so as not to clip

[notlance]

There are a couple of issues here, and hopefully I can explain them well enough to not add confusion.

First of all, if you are concerned about the "loudness" of your recordings compared to commercial recordings, the headphones you use do not matter.  Headphone sensitivities vary, but the relative level of different recording played through the same headphones will remain the same.  You may have to turn up the volume more for a particular set of headphones to achieve a comfortable loudness level, but that does not explain why your recordings don't sound as loud as commercial recordings played through the same headphones at the same volume control setting.

Secondly, loudness is a psychoacoustic phenomenon that is connected to the average sound level.  Sounds that have high peak levels will not sound loud if the average sound level is low.  Loud music tends to get noticed more than soft music, especially when that music is played in a noisy environment such as a car.  Music that gets noticed sells more.  Bands and record companies think that loud music sells more than soft music.  Therefore, commercial recordings are often made as "loud" as possible.  The way to make the music sound loud is to increase the average level of the music, and they do that by compressing the music's dynamic range.  The amount of compression in commercial recordings has increased over the last 15 years or so, and this trend has been called "The Loudness Wars".  Each band or label is trying to get their CD to sound louder than everyone else.  Whereas live music may have a difference between its average level and its peak level of maybe 20 dB (this difference is called its Crest Factor), commercial recordings, particularly popular music, may have a Crest Factor of only a few dB.

Since you probably do not compress your recordings, they will not sound as "loud" as a highly compressed commercial recording.

Thirdly: I downloaded a few of your recordings and saw that your peak levels were at about -6 dBFS.  That is a really nice level to record at, but you can amplify the recording in post to get the peak levels up to about 0 dBFS, making your recordings sound "louder" without decreasing the quality of your recordings.  This amplification will not change the Crest Factor, nor the dynamic range, nor anything else in your recording.  It will just set the level of your recording to the maximum that can be represented digitally without clipping.  Because some D/A converters may not handle a 0 dBFS level properly, I set my max level to -0.3 dBFS, but there is no noticeable loudness difference between -0.3 dBFS and 0 dBFS.

Often times in live recordings the peak level is due to applause, particularly "The Loud Clapper".  The Loud Clapper is the guy who sits right next to your mic stand and whacks his hands as hard as he can.  These clap peaks can either be edited out, or limited, or ignored (who cares of clapping peaks are clipping?).

I hope the above explanation helps.

[Patrick]

How hot are you running your levels when taping a show?  It depends on if you're running 16 or 24bit, but either way your levels should be hot but never clipping.  Even still, studio recordings go through 4-5 stages of compression and limiting to squeeze every bit of resolution they can from what's available.  As someone mentioned before, this has become a whole new issue thanks to the Loudness War over the past few years.

I love listening to audience recordings for exactly this reason-- it's a recording of the space and should have a fairly wide dynamic range.  The best audience recordings are very pleasing to listen to and never sound fatiguing after long listening.  I certainly can't say the same thing about many commercial releases these days...

EDIT: Spelling

[Brian E.]

Thanks for the huge writeup, and everyone else for their input.  As long as you guys agree that my stuff is loud enough for the source material, I'm good with it.  I'll take the dB into account too.

Spelling - I run my R09-HR at around 60-65 level (out of 80) on that side, and I basically run my 9100 pre-amp until I get to about 80% on the meter on my R09-HR.  Not sure the dB levels, but I visually check it.  That way I'm never clipping, and I can always boost a bit in post.

I always run 24-bit now

[acidjack]

^^ The point being made though, is that in post you should ALWAYS amplify/normalize to 0dB regardless of where your levels were during recording. And, as noted, you should try and isolate out any weirdly loud peak sounds (I usually find it's a snare hit, or on an SBD, sometimes a particularly loud shout or something if that's part of the vocal).

Your recordings will still be quieter than commercial releases, but they'll at least be as loud as everyone else's live recordings.

[rastasean]

It's funny that this loudness war and over-compression has actually made everything seem quieter. I doubt that is the case but music recorded in the 1970s just doesn't seem as loud as todays music. maybe it is over-compression or engineers, producers, and musicians being more conscientious about their hearing.

http://www.youtube.com/watch?v=3Gmex_4hreQ

[Belexes]

I boost to about - 3 or -2dB. At -1 or 0 I feel the signal is far too hot for my ears.

[Patrick]

^^ The point being made though, is that in post you should ALWAYS amplify/normalize to 0dB regardless of where your levels were during recording. And, as noted, you should try and isolate out any weirdly loud peak sounds (I usually find it's a snare hit, or on an SBD, sometimes a particularly loud shout or something if that's part of the vocal).

Your recordings will still be quieter than commercial releases, but they'll at least be as loud as everyone else's live recordings.

Couldn't disagree more.  Normalizing is not something that you always want to do in post production.  Plus, normalizing doesn't always work well with audience recordings, since hand claps, shouts, and PA noise can "fool" the normalizing plug in when its analyzing the signal for peaks.  I agree with getting your RMS as hot as you can, but that can be done while recording and not in post production.  I definitely have applied gain to recordings that need it, but never with a normalizing plug in.

Just how I do things    Doesn't mean it's the right way.

[Brian E.]

It's funny that this loudness war and over-compression has actually made everything seem quieter. I doubt that is the case but music recorded in the 1970s just doesn't seem as loud as todays music. maybe it is over-compression or engineers, producers, and musicians being more conscientious about their hearing.

http://www.youtube.com/watch?v=3Gmex_4hreQ

nice link

I always leave my recordings as is - I'll boost dB but only until just before it clips, and never compress

[acidjack]

I should not have added "normalize" as I understand that can do some unusual things to the waveform.  But "amplify" is simply a straight-line gain increase such that the highest peaks will be 0dB or whatever you set it at.  What possibly can be the harm in doing that?  I've heard that sometimes you want to stay a hair under 0dB for some obscure mathematical reason having to do with old A>Ds, but....

^^ The point being made though, is that in post you should ALWAYS amplify/normalize to 0dB regardless of where your levels were during recording. And, as noted, you should try and isolate out any weirdly loud peak sounds (I usually find it's a snare hit, or on an SBD, sometimes a particularly loud shout or something if that's part of the vocal).

Your recordings will still be quieter than commercial releases, but they'll at least be as loud as everyone else's live recordings.

Couldn't disagree more.  Normalizing is not something that you always want to do in post production.  Plus, normalizing doesn't always work well with audience recordings, since hand claps, shouts, and PA noise can "fool" the normalizing plug in when its analyzing the signal for peaks.  I agree with getting your RMS as hot as you can, but that can be done while recording and not in post production.  I definitely have applied gain to recordings that need it, but never with a normalizing plug in.

Just how I do things    Doesn't mean it's the right way.

[page]

I should not have added "normalize" as I understand that can do some unusual things to the waveform.  But "amplify" is simply a straight-line gain increase such that the highest peaks will be 0dB or whatever you set it at.  What possibly can be the harm in doing that?  I've heard that sometimes you want to stay a hair under 0dB for some obscure mathematical reason having to do with old A>Ds, but....

Conversely, on the D>A part of the chain; clipping can be really apparant, or non-existant. Different DACs can resolve a clip (audibly) differently, my portable dac/amp from ibasso was rather forgiving with my westones, while my denon's with the m901 is not nearly as much.

I record at -12db (limiter at -6), amp it the rest of the way to -0.1 in post, run a soft limiter at -3db (which results in a new overall peak of about -1.4) to take care of an errant peak or so and amp that result to -0.3 or -0.1.

[Patrick]

I should not have added "normalize" as I understand that can do some unusual things to the waveform.  But "amplify" is simply a straight-line gain increase such that the highest peaks will be 0dB or whatever you set it at.  What possibly can be the harm in doing that?  I've heard that sometimes you want to stay a hair under 0dB for some obscure mathematical reason having to do with old A>Ds, but....

It all depends on how much gain the normalizing process is adding.  Normalizing can bring out the noise floor of the gear and make it quite noisy.  Adding a db or two of gain is fine, and as long as it doesn't push the signal into clipping, there's no noticeable harm done. 

Also like I said before, a close clapper that creates a sharp peak in the recording can confuse the normalizing plug in, and it won't properly analyze the sound to apply the right amount of gain.

[newplanet7]

^^ The point being made though, is that in post you should ALWAYS amplify/normalize to 0dB regardless of where your levels were during recording. And, as noted, you should try and isolate out any weirdly loud peak sounds (I usually find it's a snare hit, or on an SBD, sometimes a particularly loud shout or something if that's part of the vocal).

This is exactly what I do.
Sometimes I run hot and don't add anything.

I also find the high peaks from a clap or a sharp snare hit etc, and reduce those significantly.
When normalizing I use peak normalize which takes the guesswork out of amplifying to a certain db.
All peak normalizing does is scan the waveform for the highest peak and amplifies the whole file from that point to the limit I specify, which is usually -.25 db.
That is where reducing those sharp peaks from a clap etc. beforehand works great.
Say the average music peaks around -6db in the set. However, there were a couple huge claps that reach -1db.
Without fixing those I would only be adding .75 to to the set since my limit was -.25db.
Tone down those claps to around -6 and peak normalize will add roughly 5.75db to the set.

[Brian E.]

I do the same with the hot claps, etc