Topic: newbie to post processing

[icebox]

I recorded my first show yesterday with a sony M10, using the default mics because the mics I've ordered will be coming later in the month.  I recorded 24/48 with low mic sensitivity, limiter on, and the low cut filter on.  I had the level on the M10 at between 2 and 3 for the most part.  This kept the loudest parts peaking around -2 for the most part, from overloading, but softer sections are around -6 to -9, and the concert was soft for the most part.  To listen comfortably on the computer I have to turn the speakers up more than I'd like. 

I used to record with a Sony D6 when I was younger in analog days, but never did post-processing with those recordings at that time.  So I'm post-processing for the first time.  I've read through the "read this first" material at the top of the forum.  I have Audition, and I was thinking to use the hard limiter to knock the peaks down to -3, then amplify by 2.9 db.  I don't want to mess with eq because I don't know what I'm doing with that.  What sorts of problems might I introduce by doing this approach, and is there a better solution?  One other thing, while sound was OK at the venue, the crowd was very noisy, particularly in the softer sections.

Second q: Moronically, I managed to hit the stop button (yeah, I know I should have had hold engaged; I've learned my lesson there) during a song, and lost about 15 seconds of recording before I got the recording started again.  So I have a song that cuts out during the last verse, then picks up for the instrumental ending for 30 seconds.  How should I handle this?  Should I join the 2 parts together with a quick fade on each side of the join, or leave them separate as a part 1/part 2?  I'm thinking joining them would be least jarring, but the fade will lose more of the last verse on the first part. 

[Chimney Top]

I would not use the limiter and certainly not the low cut filter while recording (if the recording has too much bass or the low end is distorting use a high pass filter or low cut in post-production). As for post production, I haven't used Audacity so I don't know what features it has, but if your levels are peaking at -6db's to -9db's you need to normalize the wav file (either before exporting/bouncing, or before... depends on the software.  Some have both options).  Your playback level should be ok after you normalize.

[icebox]

What's wrong with using the limiter in recording?  I had the red lights go on at multiple points in the show, and when I listen to those peaks that must have invoked the limiter there's no audible anything to suggest the limiter came on, at least to my ears at a reasonable volume.  I don't know what better result I could have hoped for on that score. 

And another newbie q, what's the difference between applying a high pass filter in the recording vs. post production?  Both will reduce bass frequencies?

Re: normalizing, my levels aren't peaking at -6, the peaks are at -1, but most of the show is sitting between -6 and -9, so i don't think normalizing off the bat will help, because the peak has almost no room to move, so there will be a no more than 1 db increase for the quieter parts. 

[Chimney Top]

What's wrong with using the limiter in recording?  I had the red lights go on at multiple points in the show, and when I listen to those peaks that must have invoked the limiter there's no audible anything to suggest the limiter came on, at least to my ears at a reasonable volume.  I don't know what better result I could have hoped for on that score. 

And another newbie q, what's the difference between applying a high pass filter in the recording vs. post production?  Both will reduce bass frequencies?

Re: normalizing, my levels aren't peaking at -6, the peaks are at -1, but most of the show is sitting between -6 and -9, so i don't think normalizing off the bat will help, because the peak has almost no room to move, so there will be a no more than 1 db increase for the quieter parts.

-You can use the limiter, but depends on the limiter... some may be good, others not so much.  I have not used the Sony m10, but it's probably decent.

-Both will reduce remove frequencies, but if you remove them before they are recorded, they are not recorded, and you may not want to remove the frequencies.  I would only use that if you are recording a subwoofer from a few feet away... or perhaps an acoustic guitar (only) the lower frequencies 60hz and below or so is mostly noise, but again you could use a high pass filter in post-production.

-I'm not sure what you recorded, but if the levels are too low for quieter parts of the show you may want to amplify those a few db's accordingly.  Also depends on your playback system.

Also, Audacity is a decent free software.  You can 'highlight' the area you want to amplify, click on effects > amplify and then enter how many db's.

[icebox]

I have Audacity, which I have used in the past for spectrum analysis.  I tried using that program before Audition, but couldn't figure out how to split the raw file into tracks, and gave up and fired up Audition, and I was able to figure out the basiscs of Audition much faster so for now I'm sticking with it. 

The reason I considered the LCF on the M10 was that earlier in the day yesterday, I did a test recording of a track playing on the computer (Pendulum by Broadcast, which is a very bass heavy track, and with levels peaking at -4, the bass was heavily distorted.  With LCF on, it sounded a whole lot better.  Looking at the natural and LCF spectrum analysis, the LCF cuts about 20db off the lowest frequencies, and rises up to 0db reduction at 250 hz. 

Anyway, after seeing your comment, I tried to see if I could correct the distorted bass recording I had made, and ran it through the setting on Audition's Parametric EQ "Generic High Pass fllter" and the bass level was reduced, but it still rumbled, albeit at a far lower level, in a way the original track did not, and playing with the EQ curve, I couldn't get the rumble of distorition to go away without making the track sound thin from insufficient bass.  My failure might be my inexperience in working an EQ.  It did sound better after applying the filter in post, but I think the version I made of Pendulum with the LCF on, then with the EQ tweaked to amplify under 300 hz bass by 5-10db, sounded better than anything I could do with the distorted one. 

I don't think every flaw can be corrected in post, otherwise we'd be able to make good recordings out of any defective source, which is plainly not possible.  I see that modern tools can improve a distorted original, but it seems that they don't improve it to the quality you'd get from not having distorted at all. 

[Chimney Top]

The reason I considered the LCF on the M10 was that earlier in the day yesterday, I did a test recording of a track playing on the computer (Pendulum by Broadcast, which is a very bass heavy track, and with levels peaking at -4, the bass was heavily distorted.  With LCF on, it sounded a whole lot better.  Looking at the natural and LCF spectrum analysis, the LCF cuts about 20db off the lowest frequencies, and rises up to 0db reduction at 250 hz. 

I wouldn't use the low cut regardless of the band.  If it's too much bass I would attempt to filter in post production.  I use logic pro, but if you want to split a stereo track in Audacity click on the box (left side of track) and select split stereo track... then you can pan and edit L/R tracks.

edit) also, a low cut at 250hz is way too much.  120 is usually the max, maybe 60-80hz, but 250hz is (generally) too much.

[voltronic]

The reason I considered the LCF on the M10 was that earlier in the day yesterday, I did a test recording of a track playing on the computer (Pendulum by Broadcast, which is a very bass heavy track, and with levels peaking at -4, the bass was heavily distorted.  With LCF on, it sounded a whole lot better.  Looking at the natural and LCF spectrum analysis, the LCF cuts about 20db off the lowest frequencies, and rises up to 0db reduction at 250 hz. 

I wouldn't use the low cut regardless of the band.  If it's too much bass I would attempt to filter in post production.  I use logic pro, but if you want to split a stereo track in Audacity click on the box (left side of track) and select split stereo track... then you can pan and edit L/R tracks.

edit) also, a low cut at 250hz is way too much.  120 is usually the max, maybe 60-80hz, but 250hz is (generally) too much.

I second all of this.  Low-cut, especially one that high, should never be used for music recording, and you're better off notching out any problem frequencies with EQ or a much less extreme HP filter later on. 

In the future, I would set your levels so your peaks are no higher than -12.  Yes, that's a completely safe level noise wise on the M10.  Setting the overal level low, and I mean lower than you think you should, will head off so many problems later.  Go ahead and leave the limiter on just in case if you want, but you honestly don't want it to ever have to engage.  Speaking of, I suspect your level was set too hot in the first place and the loud bass was hitting the limiter pretty hard - that was the distortion you heard.  The -4 peaks were the result of the limiter doing its job.  If it hadn't been engaged, I bet your peak levels would have been pinned in the red. 

And if you're worried about having too quiet a recording later on, this is super easy to correct in Audacity:
1. If you have applause or other crowd noise that is much louder than the music, highlight those parts only (or the entire track if you want) and go to Effects > Hard Limiter.  Knock down the max level there to where your music peaks were.
2. Select the entire track and do Effects > Normalize.
3. Now you should have your recording up to full volume.

This may seem like extra work, but trust me it's worth it.

[icebox]

The reason I considered the LCF on the M10 was that earlier in the day yesterday, I did a test recording of a track playing on the computer (Pendulum by Broadcast, which is a very bass heavy track, and with levels peaking at -4, the bass was heavily distorted.  With LCF on, it sounded a whole lot better.  Looking at the natural and LCF spectrum analysis, the LCF cuts about 20db off the lowest frequencies, and rises up to 0db reduction at 250 hz. 

I wouldn't use the low cut regardless of the band.  If it's too much bass I would attempt to filter in post production.  I use logic pro, but if you want to split a stereo track in Audacity click on the box (left side of track) and select split stereo track... then you can pan and edit L/R tracks.

edit) also, a low cut at 250hz is way too much.  120 is usually the max, maybe 60-80hz, but 250hz is (generally) too much.

I second all of this.  Low-cut, especially one that high, should never be used for music recording, and you're better off notching out any problem frequencies with EQ or a much less extreme HP filter later on. 

In the future, I would set your levels so your peaks are no higher than -12.  Yes, that's a completely safe level noise wise on the M10.  Setting the overal level low, and I mean lower than you think you should, will head off so many problems later.  Go ahead and leave the limiter on just in case if you want, but you honestly don't want it to ever have to engage.  Speaking of, I suspect your level was set too hot in the first place and the loud bass was hitting the limiter pretty hard - that was the distortion you heard.  The -4 peaks were the result of the limiter doing its job.  If it hadn't been engaged, I bet your peak levels would have been pinned in the red. 

And if you're worried about having too quiet a recording later on, this is super easy to correct in Audacity:
1. If you have applause or other crowd noise that is much louder than the music, highlight those parts only (or the entire track if you want) and go to Effects > Hard Limiter.  Knock down the max level there to where your music peaks were.
2. Select the entire track and do Effects > Normalize.
3. Now you should have your recording up to full volume.

This may seem like extra work, but trust me it's worth it.

I'm less concerned about loud applause with the current recording and more about wide dynamic range in song, which is what the 2 of the 3 bands yesterday did.  Nearly all acoustic instruments, mostly not played loud, with emphatic parts of most songs that were briefly loud .  I was aiming for a -3 peak.  And except for a few cases where the limiter was engaged, that worked (and the spots where the limiter engaged peak at -1, not -4).  If I had set my levels to -12 peaks, that means that the occasional loud stuff in songs would be there and the quiet stuff, which was a good chunk of the show, would have been below -18.   It also means that I'd have to have greater than 10db of amplification to make the thing listenable at normal volume.  That's gotta have an side effect of some sort.  Though I have been very surprised at the lack of noise when amplifying 4.5 db. 

As for the LCF, at the club, the sound was very biased to the low end, which is why I kept in on from the afternoon.  I did a new test of Pendulum, this time with LCF and Limiter off, peaking at -12.  The raw file in winamp sounds unpleasantly boomy even at these low levels.  After playing around with the HPF, I was able to get something that sounded reasonably approximating the sound of the original, after applying it and then amplifying so that peaks were at -1. So I guess I see the point, though I'm being dragged kicking and screaming to this point.    I'm just scared of doing a crap job with EQing because I just am so inexperience with it.

[voltronic]

I'm less concerned about loud applause with the current recording and more about wide dynamic range in song, which is what the 2 of the 3 bands yesterday did.  Nearly all acoustic instruments, mostly not played loud, with emphatic parts of most songs that were briefly loud .  I was aiming for a -3 peak.  And except for a few cases where the limiter was engaged, that worked (and the spots where the limiter engaged peak at -1, not -4).  If I had set my levels to -12 peaks, that means that the occasional loud stuff in songs would be there and the quiet stuff, which was a good chunk of the show, would have been below -18.   It also means that I'd have to have greater than 10db of amplification to make the thing listenable at normal volume.  That's gotta have an side effect of some sort.  Though I have been very surprised at the lack of noise when amplifying 4.5 db. 
[/quote]

Aiming for a -3 peak is way too high, IMHO.  If the limiter is engaging at all, that's a sign you've set the levels too hot.  What I record is 100% acoustic / classical, and when I'm recording a cappella choir I sometimes have peak levels in songs of -18 or lower.  You might think there would be a lot of background noise when I amplify by that much later, but it's not the case with the M10.  The self-noise is that low.  Make sure you're recording in 24bit and you'll be fine. 

[icebox]

I'm recording 24/48.   What I wonder is why is it better for post purposes to have recorded in 24 bit compared to 16 bit?

[page]

also, a low cut at 250hz is way too much.  120 is usually the max, maybe 60-80hz, but 250hz is (generally) too much.

This. the only reason I've seen to use a HPF above 100hz is when you're in an environment that has a ton of wind (more than what dead-rats will help with) and you have to turn the audio over to a production house for processing asap (in otherwords; you don't have the time in post to fine-tune it from scratch). I'm not sure I agree with it, but thats the only reason I've seen.

I wouldn't HPF over about 60hz personally unless I was multi-tracking and knew what I was doing with the tracks later. I've done a steep rolloff at 40hz on individual tracks before because I knew I'd be rolling off over that later anyway so I used the extra headroom to prevent the limiter from kicking in. Which brings me to the second point:

Go ahead and leave the limiter on just in case if you want, but you honestly don't want it to ever have to engage.

I totally agree with this; the limiter is a safety net that you hope to never use but should be infrequent enough that it's ok. Some limiters are better than others, ymmv. Also, limiters engage at different points depending on how a company did the configuration (and usually thats not changable by end users).

I'm just scared of doing a crap job with EQing because I just am so inexperience with it.

Then get better at learning frequencies and hearing sounds (and get a playback environment which sounds good enough that you can hear what you're doing). I'm not trying to be pithy about it, just pointing out that being scared of EQ and accept HPF as an alternative is a paradox, they are both having an end result that affects the audio in terms of prominance of a set of frequencies. It's 2014, trying to "make the tape" that night isn't a requirement anymore. When in doubt, do less (actually, that's a good idea in general when doing EQ but point stands).

I'm recording 24/48.   What I wonder is why is it better for post purposes to have recorded in 24 bit compared to 16 bit?

Noise floor from the ADC and other equipment. That lets you add gain without having your quantization noise floor start at 96db (and instead will likely start around 110db, something that is dependent upon the ADC used, again, not all are created equal).

[Sloan Simpson]

Then get better at learning frequencies and hearing sounds (and get a playback environment which sounds good enough that you can hear what you're doing).

This is IMO the biggest "upgrade" a taper can do.