Your ideal processing sequence (Edit: and headphones advice)

Special Interest Groups Audio Processing
1

Although I’m still an Audacity “newbie,” as the kiddies say, I’ve been rather obsessed with recording my latest originals in 2.0.3, as it’s so much better (in general) than anything digital that I’ve used in the past. I’ve been experimenting quite a bit with the best sequence of processing events – which is quite different, if my ears are being as truthful as usual, from Nero Wave Editor, which I’m used to.

To be honest, I’m finding it difficult to amplify everything so that it’s reasonably loud, while still avoiding the constant “combination clipping” from all four or five tracks playing together.

They’re fine on their own, of course (amplified to the software’s suggested max, in each case). But when played simultaneously, the overall output exceeds the maximum decibels. So there have been hours of manual tinkering to try and combine everything while still coming just up to the threshold when they’re all playing together.

Any suggestions on how to handle “Amplify” and get everything adequately loud without simultaneous-play clipping/distortion? I haven’t messed with the defaults, DB-wise or meter-wise. As far as the sequence of processes goes, I think I’ve arrived at the best. Correct any glaring mistakes you might notice, if you would! Thanks!

After Amplify:

  1. Noise Reduction – I love Audacity’s Sonic Solutions-type “sampling the noise” method. It hardly leaves any artifacts. Of course, I’m not coming in terribly hissy to begin with, but it’s still a USB mic without any outboard processing, so there’s going to be some gray noise in there. Audacity has no problem handling it.

  2. Compression – on the vocals only. It’s around 3:1 and definitely helps with my quite varied singing levels. When they’re too extreme to be handled by compression without unpleasant brick-walling, I change the volumes of particular stanzas by exporting to WAV, altering the volumes in Nero Wave Editor, and then importing back in.

This is because it’s practically impossible to change the volume of just a few seconds in Audacity without hearing a pop, even if the difference is only, say, 2.5. One can certainly try using Repair, but as Audacity doesn’t keep track of the points in time that have been altered, like NWE does, how does one even find the microseconds in question? The highlighted section certainly returns to being un-highlighted – for good – if anything else is done to the track, including merely clicking the cursor in another spot.

  1. EQ

  2. Reverb and/or delay (I hate the unnatural sound of dry vocals)

  3. Combine and export as a WAV, and then import it back in as a single track and normalize the whole thing to -1?

Thanks very much for any suggestions regarding the above sequence. While I’m at it, I might as well add this little bonus question: If anyone can recommend a brand of headphones that offers trustworthy top AND bottom, I’d appreciate it. The only ones I’ve found that get good reviews from audiophiles run at about half my monthly rent…does anything somewhat affordable exist that anyone here likes?

Thanks for any input! As always, I really appreciate your time, guys.

2

Try the Envelope Tool Audacity Manual

3

The only ones I’ve found that get good reviews from audiophiles run at about half my monthly rent

Yes, well I’m afraid that’s the kind of thing that happens. It’s not that the lower priced headphones are bad, it’s just that they’re under the thumb of the marketing and promotional people and are forced to use such words as “Sonic Bass,” “Sparkling Highs” and “Award-Winning Presence.” That kind of thing makes my ears bleed.

“Graceful Sounding and Comfortable to use During Long Edit Days” doesn’t make good ad copy.

The Gold Standard for headphones has been the Koss Pro-4AA, but you can’t mix on those for long periods because It’s like wearing two Land Rovers on your head. The Pro-3AA is better, but I can hear the sound quality slipping.

I’m having a hard time besting the Sennheiser eH-150.
http://www.amazon.com/Sennheiser-EH-150-Dynamic-Evolution-Headphones/dp/B00067OF80

I got mine through a studio move. Some silly goose left them behind and I grabbed them on the way to the bin.

They’re perfectly comfortable and have good, long-edit-session reproduction with maybe a very slight bass hump. The boost is nothing like the Disco Thump Hump bad headphones I’ve used in the past.

They have some physical problems. You can’t fold them to do a one-muff listen, they don’t fold up and they have a Y-shaped, straight cord. Some audio types hate that. I don’t much care.

Koz

4

If you are in Europe, these are the best budget headphones that I have found: http://www.thomann.de/gb/the_tbone_hd880.htm
They won’t compete with Koss Pro-4AA, but they are fantastic for the price. Out of 8 sets of these (not all for me :wink: ) I’ve had one pair that were faulty, but Thomann replaced them with no quibble and they covered the postage charge. They are quite large and sit over the ear, which I find a lot more comfortable than having by ears squashed by foam pads.

I’d advise against wearing any kind of headphones for extended periods as I don’t think that it is good for the health of ears, so take regular breaks and take the headphones off when not listening to them.

5

A limiter to remove any occasional loud spikes … Limiter

If the delivery format is stereo , maybe add pseudo-stereo to the otherwise mono vocal …

6

Thank you for the suggestions, guys. It looks like it would behoove me to experiment with using the Limiter instead of the Compressor, or perhaps in addition to the latter, in extreme cases…? It’s always interesting trying to assess the differences between the old analogue compressors and limiters I used for years, vs. the often radically different digital equivalents. I appreciate the tips.

EDIT: I’m also looking at Audio-Technica ATH-M30 and M40 cans. Unless I’m mistaken, the M40s will require a pre-amp, instead of just being plugged directly into the PC. I can’t find anything online that clarifies whether or not I’ll need one with the M30s, or the Sennheisers recommended above. Any experience/thoughts in this vein? Thanks again.

7

You can simulate an FM broadcast compression with its automatic peak limiting, volume compression and loudness compensation with Chris’s Compressor.

http://theaudacitytopodcast.com/chriss-dynamic-compressor-plugin-for-audacity/

I change the first setting value, Compression Ratio from 0.5 to 0.77 and it gets remarkably close to one of the local FM stations with one-pass processing.

You can mess with the French Horn attack and release times if you want, but Chris does a remarkable job out of the box.

You would produce your echo/reverb , Eq and other production special effects and let Chris take care of volume variations and waveform processing.

Koz

8

Any experience/thoughts in this vein?

I found myself with a box of multiple different headphones and one day I went through them all plugged into my Mac and graded them. I did this on purpose because that was one of the ways I actually use them.

1000×779 101 KB

I was not after world class “Golden Ear” ratings with a panel of international experts. I was after real-world headphone use, not leaning back for three hours with a good Mahler.

So given that, I found several units with “disco bass boost” and several units that hurt my ears to listen to at normal volume. I used the Pro4AAs as the baseline for comparison. I still like my ratty, thousand year old Sennheiser HD414. I have to hold my head a certain way because of a broken cable. I understand they have restrained bass, but they’re remarkably pleasant to listen to for long periods. The eH-150s continue in that tradition only with crisper highs and an actual bass response. I can listen to these through a whole movie without bothering the neighbors (Not everybody in Los Angeles has air conditioning).

Given you’re even considering a headphone that requires a preamp for good performance means you’re probably in the Mahler camp and you should certainly try multiple different headsets rather than just watch what we’re doing. We may have seriously different goals.

Koz

9

Thanks for both suggestions. I’ve actually downloaded Chris’s Compressor and the attendant podcast, since its “all-in-one” process is highly appealing.

All I’m after in terms of headphones is a flat response that doesn’t lie to me about the top or bottom – i.e. solely for mixing, rather than Mahler-type passive enjoyment. My motive is to get the finished songs to sound good on any reasonably well-balanced stereo, rather than just mine; while this obviously depends on the listeners’ own settings and equipment, some mixes in my recent past have been too top-heavy or bottom-heavy in general, because I didn’t have reasonably flat/honest cans. I’m merely hoping to rectify that without spending extra money on a pre-amp and/or phantom power, if possible.

10

I have a travel story where I was forced to go to a book store cold and buy off-the rack headphones. I found out later they were very bass heavy and it’s a good thing I wasn’t doing anything critical.

Koz

11

The limiter acts much quicker than a standard compressor, (e,g, “Chris’s”), and will trim off any occasional loud spikes …
Limiter acts quicker than standard compressor.gif
then apply the compressor, (which wouldn’t have been quick enough to cope with these transient spikes).

12

A note here about the BBC Peak Program Meter. It is intentionally designed to ignore occasional transient peaks. If you’re worried about all peaks one at a time, maybe you should examine your goal.

Koz

13

Excellent tip. Thanks for that!

14

So in your opinion, it’s best to apply Chris’s Compressor after applying the reverb and EQ?
(Thanks again for all of your help.)

15

You need to be very careful if you apply compression after applying reverb because it will tend to make the reverb louder.

16

That’s what I’ve always concluded; I usually even wait until after compression to apply EQ, as compression tends to diminish the low end, and this usually has to be atoned for after the fact. But Koz’s advice made me wonder if this particular plug-in behaves differently; I probably just misunderstood his post.

17

(Bump)

I’m still extremely curious to hear about any alarm bells that might go off for those who are more experienced with Audacity than I, concerning the actual order of the processes that I apply to music tracks. (If there are none, then that’s awesome, of course; but it certainly seems useful to ask.)

For convenience, the original post is pasted below. Thanks very much for any input.

I’ve been experimenting quite a bit with the best sequence of processing events – which is quite different, if my ears are being as truthful as usual, from Nero Wave Editor, which I’m used to.

To be honest, I’m finding it difficult to amplify everything so that it’s reasonably loud, while still avoiding the constant “combination clipping” from having all four or five vocal tracks playing together.

They’re fine on their own, of course (amplified to the software’s suggested max, in each case). But when played simultaneously, the overall output exceeds the maximum decibels. So there have been hours of manual tinkering to try and combine everything while still coming just up to the threshold when they’re all playing together.

I haven’t messed with the defaults, DB-wise or meter-wise. As far as the sequence of processes goes, I think I’ve arrived at the best. Correct any glaring mistakes you might notice, if you would! Thanks!

After Amplify:

  1. Noise Reduction – I love Audacity’s Sonic Solutions-type “sampling the noise” method. It hardly leaves any artifacts. Of course, I’m not coming in terribly hissy to begin with, but it’s still a USB mic without any outboard processing, so there’s going to be some gray noise in there. Audacity has no problem handling it. The upper field is set between 12 and 14, depending on the level of hiss.

  2. Compression – on the vocals only. It’s around 3:1 and definitely helps with my quite varied singing levels. When they’re too extreme to be handled by compression without unpleasant brick-walling, I change the volumes of particular stanzas by exporting to WAV, altering the volumes in Nero Wave Editor, and then importing back in.

This is because it’s practically impossible to change the volume of just a few seconds in Audacity without hearing a pop, even if the difference is only, say, 2.5. One can certainly try using Repair, but as Audacity doesn’t keep track of the points in time that have been altered, like NWE does, how does one even find the microseconds in question? The highlighted section certainly returns to being un-highlighted – for good – if anything else is done to the track, including merely clicking the cursor in another spot.

  1. EQ

  2. Reverb and/or delay (I hate the unnatural sound of dry vocals)

  3. Combine and export as a WAV, and then import it back in as a single track and normalize the whole thing to -1?

Thanks very much for any suggestions regarding the above sequence. As always, I really appreciate your time, guys.

18

This is the workflow that we honed for LP transfers - you may find some useful tips here: http://manual.audacityteam.org/o/man/sample_workflow_for_lp_digitization.html

Re budget, flat response, cans - have a look at the Sennheiser PX-100, about 30 quid in the UK (the ordinary ones not the noise cancellers).

I use a pair of studio Sennheisers at home - and I’ve just spent a week or so on holiday with my travelling Px-100s/iPod - and this reminded me of how good these little fellas are.

WC

19

Terrific. Thanks, WC!