Topic: Help/Advice/Tips on Post-Processing a Recording?

[Bullet1979]

Hey all, I've done a fair amount of reading on this board and have tried to master many many recordings in the past but I feel like I've hit a ceiling with "self-teaching" and am reaching out to see if anyone would be willing to give me some direction on ways I can improve my skills.  I just went to a show on Tuesday night and, as stated, am trying to use the information I've gathered to make the recording sound as good as it can... Now, I understand a recording can only sound so good considering the source - but I feel like I'm not doing the best that can be done.

My setup was CA-11 cards to a MM-CBM-Mini battery box into the line-in of an Edirol R09.  I'm using Adobe Audition 3.0 but would certainly not refuse to use a different piece of software if something else is recommended.  First, I used the Dynamics Processing to compress (which I probably have no business doing because I still can't figure out exactly what I'm doing other than trying to get the compress around 2.0) then I used the Graphic Equalizer to adjust the EQ.  Once again, with the EQ, I really had no intuitive knowledge of what I was doing so I just sorta arbitrarily adjusted the sliders the try to get a sound I was happy with (mostly trying to achieve a balanced sound while raising the low-end a bit).  I'm fairly happy with the instrumental portion but vocals seem a little too sharp...

Anyway, I've uploaded an untouched version of a song from the recording as well as my attempt at mastering.  I uploaded them as flac24 but if anyone prefers I upload something a bit smaller I will...  If anyone is able to utilize their expertise and accomplish a better sound than I have, I'd love to know what kind of steps you took or what configuration you used for EQ (maybe a screenshot).  I know this is asking quite a bit but this is a skill I'd really like to improve so any advice would be immensely appreciated!

Untouched:
http://www.sendspace.com/file/t5qv49

"Mastered:"
http://www.sendspace.com/file/0bd4id 

[Scooter123]

--Processing Files is a Self-Taught
--It is Always Subjective
--Processes Will Depend on Your Ears, The Room & the Mikes

I can't remember ever boosting the bass--I'm always using highpass at 60,80 or100 to remove it.  I rarely add tones.  Usually just compression, highpass, and maybe some delay. 

My suggestion is to load up your favorite song and spend about 4 hours on Audition playing with the various settings to get a crude understanding what you are doing and what you like.  I can't teach you. 

[Bullet1979]

A big reason that I'm reaching out to see if someone else can show me what they're capable of is that it feels like I have hit a ceiling...  I can spend 5 minutes throwing my usual combination at a recording or I can spend an hour trying to comb through different effects/tools and both results will sound practically indiscernible.

I don't think I'm necessarily looking to be taught precisely how to attack each project, but at the same time it seems that a lot of my recordings sound pretty similar because I generally use the same setup and seek out the same spot(s) in the venues.  I mostly would be interested in hearing what someone who knows what they're doing can attain with one of my recordings so I have an idea of what to aspire to (and if I can get little pointers on how he/she did it, even better).

Thanks for the advice, though... any input is appreciated!

[Scooter123]

Bullet--

Its trial and error, Buddy.  You need to use your ears. And to some extent your eyes (the wave file will tell you a lot).  By the way, it takes me about 3 hours to master a tape--copying, compressing, highpass, boosting, joining the sections, dithering it down, naming the splits, reducing clapping, and splitting it. 

Seriously, about 5 years ago, I loaded up "Layla" and went through about 20 effects in varying degrees effects on Gold Wave.  Probably took me a good half a day. 

I came away with the idea that compression and highpass was all I really needed, although speed correction and delay were sorta cool.  Once you've mastered about 50 tapes, you'll start to get a grip on it.  Just get out and do it. 

You'll also figure out that the music, the mike and your position in the hall will have a HUGE effect on the sound, more so than mastering. 

[taperj]

Just get out and do it. 

[hieye]

Hi Bullet1979,

If you found any great answers to your question, please PM me.  I'm in a similar dilemma. 

You sound like you've done more post-processing than I have; I posted a similar question:

http://taperssection.com/index.php?topic=132591.0

I got a few concrete responses, but the whole question of post-processing seems to yield vague answers.  It would be nice to have a step-by-step sticky/wiki to "follow" and then expand upon.

I'm using Audacity simply because it seems like that is what most use.  I also have SoundForge. 

Good luck with the taping/processing. 

[ghellquist]

So here are my takes on post processing. Probably not what you wanted, but you asked for it.

First point is, why post process? What especially do you want to achieve? Unless you have this in place it will mostly be a non-directed process. One of the really important things to consider is when to stop.

And this brings me to the first main point -- can you hear when to stop? You then need a decent playback equipment that you are fairly used to. My personal guess is that you should aim for a specialized sound card (not one of those game cards, Soundblasters or whatever) and a good set of speakers or headphones. Given that good speakers start well above headphones and also are influenced by the room, open back headphones should be your first stop. Open backed because, in general, these are more true to the sound than close backed ( and as always, there are exceptions ) .  I like the higher number Sennheiser ones, HD600 is a good choice if you ask me and it will cost a bit of money. Budget is a limiting factor for us all, but do try to go a fair bit above the typical iPod earbuds.

Now, to why you want to post-process. I will not look into the very obvious things such as compressing to mp3 or FLAC or whater in order to save some bandwidth. I will also assume that you have an idea if you want to chop things up into songs, remove dead moments and so on. So why on earth then process. Only you can answer that, by I tend to do it because:
  - to make it sound better
  - to make it sound more even
  - to make it sound louder

Each of these takes its own part of the process. Lets start from the end. And this is actually what i do last in the chain, making it louder.

First, don´t try to make it sound like a commercial CD recording in loudness. This will most probably spoil the whole purpose of everything else. There is a volume control on all playback equipment, use that. But we could move things a little way up in volume.

The ideal is that the peaks goes almost all the way up to 0dB but never quite there. Reaching 0dB will sound bad on many types of equipment. So keep a good headroom left, somewhere around 1 to 3 dB is a good point to aim for unless you know your software really, really well. One tool you can use is normalizing. This will raise the level without otherwise changing the sound.

A small digression here. Doing all your processing of 32 bit float files in the computer is a good bet. This saves you from a lot of issues and keeps quality through the process. The very last step then is converting to 16 bit or whatever format you want  to present the results.

Another tool you can use to raise levels is called a limiter. Limiters comes in many different flavors and are used in different ways, but they all have one thing in common : use them too heavily handed and all sounds like shit. Easy does it. You also need to know if the limiter actually limits to a max level (say -1dB) or if it lets some peaks pass through. They are all different. Limiters work by keeping the odd peaks of volume down, allowing you to raise everything else. You need to listen to the result carefully when using a limiter.

continued in next post ....

// Gunnar

[ghellquist]

continued ...

To make it sound more even in loudness.

This can basically be done in two different ways, manually or automatically. The best way if you ask me is to do it manually.

Most music programs allows you to add something called "automation" of volume -- this is actually the manual way. You listen to the music and change the volume in the meantime. This is often shown as an automation curve. Best way to do this is using a really good computer connected fader. Low-price version might be the Behringer BCF 2000. Look it, up, no need to buy it as there are other ways.

The idea is that when volume goes down, you slightly raise the playback volume. When volume goes up, especially when the audience makes noises, you turn down the playback volume. The really good thing is that you have full control over everything and can do this as heavy handed or lightly as you want to. It does take time to do which may be a disadvantage.

It can be done in a lot of different ways depending on what software you are using, the basic idea is the same in all cases. Check your software and what alternatives there are. This feature alone might make you select a different software.

The automatic way is more lazy and less predictable. It is called a compressor. All software programs has some kind of compressor. They all have a bunch of knobs and you need to experiment with each to get the effect you want. Again, keep a light touch. A general advice may be to run the material through the compressor several times rather than trying to do it all in one sweep. A compressor always modifies the sound so you need to listen to the result with epupment that allows you hear any bad side effects.

Doing a taped performance what we want to use is a single band compressor, the most common type, suitable for full materials. There are other compressor that shine on snare drums or vocals alone, but these are not the most suitable in this kind of situation.

Ideally, I want to use completely different compressing on the music and on the audience noise ( applauds etc ) but it can be time consuming and difficult to get there.

I generally compress as the middle step, after getting things sounding good and before getting volume up, but your choice may be different. As always, once you get confident in what you are doing, rules are to be broken.

to be continued ...
// Gunnar

[ghellquist]

third part

Making it sound good.

Now this is the part that really takes taste and personal preferences into account. What sounds good to me might sound bad in your ears. There are a few basic factual things, but everything else is based on hearing the changes and either liking them or avoiding them.

The basic factual things is to remove all subsonic bass content. The things in the bass you cannot hear on your playback equipment probably is not music but spurious noise. This depends on the type of music you record, the equipment you use and what you intend to use the music for. Recording a really large church organ with a 32 foot bass pedal is different from recording say an acoustic bluegrass band. Those really low bass notes from the organ may go very low, even into single digit Hz and requires playback equipment very rarely seen in normal households even using a typical home theater subwoofer system.

Most music seldom goes that low, exactly how low it goes depends on a lot of things. Secondly, even if it was on the show maybe your mics did not catch it. If you run a typical cardioid or supercardiod mic there is very little sound left below 40 Hertz.

A safe bet then is to cut out the very lowest frequencys already from the start. This will be a so called high pass filter, perhaps set at 20 Hz. The very best filter for this may vary, but lately I have taken a liking for linear phase filters for this kind of filtering, but a normal filter will do the job. Typical test is to swith the filter in and out and listen for any effect. If I cannot hear any difference that is a sign that I have reached the lower limit of my listening equipment. Leave that filer in before going to the next step.

Next step might be to EQ the sound. Slightly. Less is more. I might hear some typical resonances in the room or in the PA. A slight notch here, decreasing that frequence a few dB will make things better. Maybe the bass did not come through as mighty as I wanted it to -- cutting down on mid and high frequencys is a good idea then. And maybe the cymbals come through a little too much, cutting a bit is a good idea.

It is very rare for me to raise any specific frequency range, I tend to lower all others instead. There are several reasons for this, but it basically boils down to it sounding better in the end.

Sometimes a very slight compression might be used to "glue" things together a little. In that case I set the compression rate as low as 1.1 or 1.2, too little to be heard but just about noticed.

Sometimes a little reverb may be called for. Adding reverb just right on a full recording is very difficult so you might want to leave it out.

It has happened that I have used the other tools in the programs. Very rarely, on very specific things, very lightly. Your choice, but only use it if you are confident that your listening equipment actually allows you to hear what is happening.

A few final words in this three part series. If anything comes across here as sounding as I would know the thruth, please pass by that. This was written not with a lot of afterthough, on a website and in a foreign language ( my native language is Swedish ) . I concurr with several of the earlier answers as I believe there is no truth here, only taste. And taste do differ which I really appreciate.


// Gunnar

[Bullet1979]

ghellquist, thanks so much for your input!  A lot of what you said sounds like good advice and I will definitely consider all of it when I'm experimenting in the software.  I'm especially intrigued by your method of EQ'ing as I've only been raising particular frequencies...  thanks again!

[Shadow_7]

It is a lot of trial and error and personal tastes.  There's probably more to be done on the input side (point of capture) than in post.  Of course you'll need some good monitoring equipment to even tell what you're doing for the most part.

You can do some data analysis of the EQ spectrum of what you have versus what you want it to sound like.  And other more scientific approaches.  But human hearing and mic responsiveness doesn't always work in calculatable ways.  I generally do an EQ pass to bump the low end, but that can make percussive sounds unruly.  And ultimately I'll address that issue with mics with a better low end. 

I also do a hard limiter to bring those unruly things in check.  And of course amplify / normalize to audible.  Beyond that, not much to do with a simple stereo setup.  On location has it's own ambience.  In the sterile studio environment however you can add ambience in post.  And many do with reverb and other things.  But that could sound horrid on a source that already has some.  Some like to ride faders on a track to add dynamics and the likes where the musicians didn't or didn't enough.  I prefer to trust the musicians, less work.

[Shadow_7]

I don't like the mastered one (on my laptop).  The untouched one has a bit of unruly peaks from applause on the Right channel.

In audacity, I would do this to it.  See if you like it or not.

Import audio -> untouched.flac

Effect -> Amplify ( 8.7dB to reach 0dB, do 7.0dB to leave some headroom )

Effect -> Bass Boost ( 200 / 12 defaults )

Effect -> 151 -> Hard Limiter ( -9dB / 1 / 0 )

Effect -> Amplify ( 9.0dB to 0dB)

And that's how my post flow generally goes.  I don't normally do bass boost, but that track seemed quite lite on the low end(untouched).  Normally I'll do Effect -> 181 -> Multiband EQ.  I still don't like the EQ attributes of the track, but I'm not going to invest the time to identify why.  And that's just what I perceive on my laptops speakers.  I hard limit to just graze the average high peaks, and that number varies depending on the content / levels used.

[Bullet1979]

Thank you very much for your advice and input, Shadow_7.  I'm liking the differences in how it turns out with your modifications.

As for the Multiband EQ - is it really just a matter of adjusting different frequencies and just listening the best you can or is there generally a benefit to having certain "shapes" in the EQ levels (i.e. a 'W' shape or a 'V' or '~', etc.)?

[Shadow_7]

Certain shapes are for purpose specific applications.  Like broadcast in MONO over AM radio.  And other want to sound like an old record / particular label / whatever attributes. 

As far as EQ, you want it to be true to the source.  Which requires that you either a) know your mic, and/or b) know your source.  And some intuition on how the recording sounds versus the real thing.  Some of which is dependent on the gear reproducing the sound, more so than what's in the recording.  If your little headphones or speakers cannot do > 16kHz frequencies or < 120Hz then filtering those OUT of the recording might help those things sound better by NOT trying to reproduce that which it can't.  And other tricks to sub in for whatever you feel is appropriate where bass boost was listed earlier.  Within reason, the more you edit, the more you ultimately degrade the source material.  i.e. always keep a copy of the unaltered original. 

Also take notes of how you got from A to B and put it in a .TXT file.  It's an ever evolving process and what you do next week might not have any relation to last week.  But the client / group might prefer last weeks version to this ones.  Having notes helps you make this weeks version sound like last weeks and so forth and so on...

[boojum]

Bob Katz is the guru on mastering.  He has a book out.  It is a bitch to understand.  But it has all you need to know.  It is damned tough going but better to start good habits than bad.  Bob will teach you the good ones.

Cheers

[live2496]

Gave it a quick listen and it sounds like the bass has been rolled off while recording. Use a low-shelf starting at about 160 or maybe 200 and boost the low end gently until it sounds more balanced to you.

I would guess that you could boost the low end quite a bit without doing any damage. Eq it first before trying to raise the overall level. Also be careful when doing a wideband boost of the lows as it can begin to mask the mids and highs.

Each recording is different, so just use your ears and make it sound balanced.