Seamless Edits: Help! [Split from the Audacity Problem thread]

[ArtemisBee]

Hello!

I'm pretty new to the recording world, and enjoying the heck out of it right now. But one problem I'm encountering is during the editing process. When I record (in Audacity) I commandeer the walk-in closet, set up, record, edit/process and then tear down. As expected, after I upload I can count on there being a couple of edits that need to be done, like I mispronounced a word or missed a word, you know. Obviously I don't want to record the whole thing over just to add in one word/sentence, so I set the booth back up, and then record the word and then plop it into the track where it belongs.

My issue is this: The newly added piece doesn't match the original, even after processing. Maybe it's a little bit quieter, maybe the noise reduction isn't quite the same.

Does anyone have any tips to make these kinds of edits more seamless? And what does your usual editing process look like before uploading? Normalise, THEN compress? Vice versa?

I'm really looking forward to your thoughts!

Thanks,

EJ

[TriciaG]

EJ - I split this off into its own thread so that you get more eyes on it. Plus, I think it's less of an Audacity problem as it is a general recording problem that can happen to anyone, regardless of software used.

I try to edit as soon after recording as possible, to minimize the differences in sound. I know this isn't always possible, though.

Mic placement is important, of course; try to place it in the same position in relation to your mouth every time.

Try recording about the same time of day, when your voice is as tired (or not) as the original recording.

Listen to a few sentences before the edit to be made, and try to match your tone, reading speed, etc. Also, don't just record and edit in the single word or phrase. Record the whole sentence (or thereabouts), and sub it all in, or cut out the phrase you need and sub in that portion. Cut and sub in at natural points, such as a breath or pause.

Other than that, you can amplify or reduce the volume on the edited part as needed to match the surrounding vocals. Usually 1-2 dB up or down is all I need.

I do all my editing (cutting, re-recording, etc.) first, then do any noise reduction, compression, and volume adjustments. That way when I notice I have a part I have to re-record, the background noise is generally the same up in my recording room, so I can just get it all in one pass at the end of my editing session.

I don't normalize. I determine what the volume is and use Amplify to get where I need it instead. Normalize doesn't give me as much control. (It adjusts the overall volume to match what the highest spike in your recording will end up at. If you have one big spike and the rest is really quiet, that will throw off the normalizing.)

I don't need to amplify, since my mic input volume and mic placement get me the volume I need. So I finish my editing (cuts, re-records), then do a quick compression (light; only -13 dB threshold), then noise removal, and export to MP3.

Noise removal vs. compression order isn't important, unless you're also making the compressor "normalize to 0 dB after compressing" (which I do not recommend, as I said what I did about normalizing). Compressor doesn't affect lower volume sound, so it doesn't affect the background noise. I would, however, do any noise removal before amplifying, because amplifying brings up the volume of the background noise, too, making more noise to clean out.

Err, guess that is about it from me. Hope this helps some!

[loon]

When I have to do a pickup after the fact I'll read a sentence or two before the line I want to fix. That gives me a chance to get my pacing, emotional tone, pitch of voice, etc. back where it needs to be to edit in nicely.

[Peter Why]

Especially if I'm trying for a particular character's voice or a particular emotion, I'll record the same fragment a number of times in slightly different ways/tones/accents, then try each one out for the best fit.

Peter

[Kazbek]

I make my edits in the final version. I apply noise reduction to the new inserts before copying, to make sure the noise profile is up to date. I often have to amplify just the inserted portion to match the surrounding volume. I listen to the earlier version before recording and try to imitate my vocal quality and intonation. Sometimes it takes a few tries. I replace the smallest passage surrounded by pauses. I've also found thats edit tend to sound more seamless if I keep the leading breath and trailing pause from the earlier version and insert only the vocal portion of the soundwave.

Michael

[pschempf]

As others have said rehearsing the bit to be replaced helps get your voice into the proper cadence, tone and inflection. I find it useful to listen to the section while I recite it several times before I try to record it again. Contrary to what Tricia said, I find it easier to patch the smallest piece possible, the smaller bit being less noticeable than a longer phrase if it's a bit off, but to each their own. I check the volume of the bit to be replaced for its volume. That gives me a starting point for correcting the volume of the new bit. I look for obvious breaks in the wave form to chose where I want to start and end the replacement. T's, p's and c's at the beginnings of words are usually easy to see in the waveform and can be cut very precisely. S's can also be seen and cut in the middle of the sound for a seamless join if the volume is matched. A trick I learned from PhilC is to use cntrl-z before you make a cut. It shifts the cut point slightly so it intersects the waveform where it crosses the centerline and avoids creating a click at the cut. If I make a poor edit, I can use the Edit/Undo function in Audacity to back up and try again.

I'm sure there are other things I do that I can't recall at the moment, but if I do I will post them.

At the end of the day, some edits seem to defy seamless correction and all of them may not end up to be perfect. No need to lose too much sleep over that.

[philchenevert]

A permanent recording place would solve most of your concerns. Recording and editing in the same place makes it much much easier to match everything even if it is days or weeks later. And I ALWAYS need to change or insert words when editing. tsk, tsk. This is useless advice if you don't have room but it needed to be said. Happy recording and don't be too fussy, most edits are fine.

[GraceBuchanan]

I'm so glad that others covered so much of your question. Here's one more tidbit:

Like you, I have to set up and strike my equipment each time I record, so I record everything at least twice. Sometimes my first take is slow and careful to make sure that I have every syllable clear and available to be edited into the final cut. Then I record what I intend to be the final cut, and don't worry about mistakes since I have that first take.

Sometimes I let my enthusiasm carry me through the first take, and then make sure the second one is enunciated perfectly so it's available for edits into the final cut.

By recording my reading twice, I usually avoid having to do retakes. Usually. Sometimes I splice syllables together to fix words, and that feels like doing micro-surgery as I zoom in to cut and paste precisely.

By the way, the keyboard shortcut on my computer is simply z, not ctrl-z, for making smoother edits. I might have changed it to make it that way; I'm not sure.

It's great to see that we all do things differently. We each do what works best for our unique situation, and based on our personal preferences.

[pschempf]

I think you’re right, Grace, just z.

[ArtemisBee]

Thank you, Michael! I hadn't even considered the breath lead in. And, for whatever reason, I hadn't even thought about editing just the fragment and amplifying to match before this thread. So much to learn! I appreciate the input.

I make my edits in the final version. I apply noise reduction to the new inserts before copying, to make sure the noise profile is up to date. I often have to amplify just the inserted portion to match the surrounding volume. I listen to the earlier version before recording and try to imitate my vocal quality and intonation. Sometimes it takes a few tries. I replace the smallest passage surrounded by pauses. I've also found thats edit tend to sound more seamless if I keep the leading breath and trailing pause from the earlier version and insert only the vocal portion of the soundwave.

Michael

[ArtemisBee]

Thank you, Tricia, for pulling this to it's own thread. I've gotten really great responses and this has been really helpful.

I always have to amplify afterwards. I guess I'm not a loud as I thought. The suggestion of amplifying the fragment to match the rest is a good one, and came up a couple times. I don't know why I didn't think of it, but it's probably to do with not being as comfortable with Audacity as others are. How do I know how many dB's need to change?

I've clearly been using the normalise all wrong. Live and learn! Thank you so much for the great input!


Best,

EJ

EJ - I split this off into its own thread so that you get more eyes on it. Plus, I think it's less of an Audacity problem as it is a general recording problem that can happen to anyone, regardless of software used.

I try to edit as soon after recording as possible, to minimize the differences in sound. I know this isn't always possible, though.

Mic placement is important, of course; try to place it in the same position in relation to your mouth every time.

Try recording about the same time of day, when your voice is as tired (or not) as the original recording.

Listen to a few sentences before the edit to be made, and try to match your tone, reading speed, etc. Also, don't just record and edit in the single word or phrase. Record the whole sentence (or thereabouts), and sub it all in, or cut out the phrase you need and sub in that portion. Cut and sub in at natural points, such as a breath or pause.

Other than that, you can amplify or reduce the volume on the edited part as needed to match the surrounding vocals. Usually 1-2 dB up or down is all I need.

I do all my editing (cutting, re-recording, etc.) first, then do any noise reduction, compression, and volume adjustments. That way when I notice I have a part I have to re-record, the background noise is generally the same up in my recording room, so I can just get it all in one pass at the end of my editing session.

I don't normalize. I determine what the volume is and use Amplify to get where I need it instead. Normalize doesn't give me as much control. (It adjusts the overall volume to match what the highest spike in your recording will end up at. If you have one big spike and the rest is really quiet, that will throw off the normalizing.)

I don't need to amplify, since my mic input volume and mic placement get me the volume I need. So I finish my editing (cuts, re-records), then do a quick compression (light; only -13 dB threshold), then noise removal, and export to MP3.

Noise removal vs. compression order isn't important, unless you're also making the compressor "normalize to 0 dB after compressing" (which I do not recommend, as I said what I did about normalizing). Compressor doesn't affect lower volume sound, so it doesn't affect the background noise. I would, however, do any noise removal before amplifying, because amplifying brings up the volume of the background noise, too, making more noise to clean out.

Err, guess that is about it from me. Hope this helps some!

[ArtemisBee]

But where will I put my laundry? But in all sincerity, thank you for the tip. Every time I do edits I think to myself how much easier it would be if I had a more permanent spot and could reduce the variables that go into the recording. Thanks, Phil!

A permanent recording place would solve most of your concerns. Recording and editing in the same place makes it much much easier to match everything even if it is days or weeks later. And I ALWAYS need to change or insert words when editing. tsk, tsk. This is useless advice if you don't have room but it needed to be said. Happy recording and don't be too fussy, most edits are fine.

[ArtemisBee]

Ohhhhh!! That is a great suggestion, Grace, thanks! I think I will start to do that for my shorter reads. Thank you!

I'm so glad that others covered so much of your question. Here's one more tidbit:

Like you, I have to set up and strike my equipment each time I record, so I record everything at least twice. Sometimes my first take is slow and careful to make sure that I have every syllable clear and available to be edited into the final cut. Then I record what I intend to be the final cut, and don't worry about mistakes since I have that first take.

Sometimes I let my enthusiasm carry me through the first take, and then make sure the second one is enunciated perfectly so it's available for edits into the final cut.

By recording my reading twice, I usually avoid having to do retakes. Usually. Sometimes I splice syllables together to fix words, and that feels like doing micro-surgery as I zoom in to cut and paste precisely.

By the way, the keyboard shortcut on my computer is simply z, not ctrl-z, for making smoother edits. I might have changed it to make it that way; I'm not sure.

It's great to see that we all do things differently. We each do what works best for our unique situation, and based on our personal preferences.

[ArtemisBee]

Thanks, Rich! I like the idea of giving myself a buffer. Appreciate the tip!

Best, EJ

When I have to do a pickup after the fact I'll read a sentence or two before the line I want to fix. That gives me a chance to get my pacing, emotional tone, pitch of voice, etc. back where it needs to be to edit in nicely.

[silverquill]

A lot of good suggestions from experienced readers!

I didn't get the whole editing thing when I started. The first thing I ever did was a Poe poem for one of Tricia's project. I made many recordings until I got it right!

Then I kept thinking if I could just learn to read better, I wouldn't have to edit so much. Still hasn't happened.

I'm of the, "record a longer sentence or phrase" tribe. Maybe I just don't have the finesse to cut in a word or something in the middle of a sentence. I also do a lot of PL work, and hear a lot of bad patches. We're pretty lenient, but with some effort and patience, we can get pretty good results.

Even if one has to set up each time, keeping things as much the same as possible helps. And, as Tricia pointed out, our voices change with the time of day or other activities, so sometimes we just can't match that quality perfectly. For initial editing, I am more and more trying to do that in the same sitting as the recording to minimize these differences. That is not always feasible, of course. The only thing I do before inserting the edit is noise reduction. I find it easier to do the volume adjustment after the edit is in place where I can listen to everything and hear how the volume blends in, or doesn't, and make the adjustments there.

You'll get more familiar with the sound of your voice and reading style and find out what procedure works best for you. Keep it as simple as possible, and don't let editing become a barrier between you and enjoying the experience.