Chris and Rosey discuss Cool Edit Pro
Chris: Using CEP as much as I have, I know it mostly second nature, and so I can describe how to edit efficiently.
Rosey: Excellent...I know who to ask about it then :)
Chris: It is still slow and tedious though.
Chris: There is no real-time DSP.
Rosey: OK.
Chris: Unless you consider overall volume and pan.
Rosey: What is pan, please?
Chris: The position in the stereo field. The spectrum left to right.
Rosey: Oh, OK, I know what you mean.
Chris: You can adjust those for each track without editing the waveforms, but then CEP still does annoying background mixing after any changes, because it makes adjustments to the temp files on the hard drive.
Rosey: Is it supposed to do that?
Chris: Yes, it's designed that way. The temp files are also why it requires at least twice as much space as the total file sizes to do anything. When you load files in the multitrack mode you will see in the right side of the status bar, the total free space diminishes. It gives you the figure in mb. And to save any changes for a mix you have to actually save the changes to the wav files. None of that stuff applies to Cubase, thats why I thought it would be very good for you and your son. Still, CEP will provide more versatility than his 4 track tape recorder.
Rosey: Oh...you can no longer save the original, you mean?.
Chris: You can overwrite the original but I don't recommend it. There is a bug or something in the program that sometimes causes files to be corrupted when you overwrite. So any time I make changes I save a new file with a new name. Usually I keep the raw version and delete intermediate revisions. If after some time I am satified with the revisions I delete the raw tracks and burn everything to a cd.
Rosey: I always save the originals of things I mess around with.
Chris: With just the 2-channel sound card your son won't be able to tranfer four tracks independently at once. But there is a method I have found to deal with that.
Rosey: Oh?
Chris: He can record a sharp tone or other sound at the head and tail on the 4 track tape, before and after each song or whatever, then record two tracks to the computer, then the other 2. Then find the peaks at the end and beginning and select inbetween. Any discrepancy in the number of samples can be adjusted with the Time/Pitch transform in CEP.
Rosey: OK, I know where that is :)
Chris: First you get the 100% figure for two of the tracks, then copy to clipboard the figure it reads out for the time. Then with the corresponding portion selected for the other two tracks, paste the figure you copied for the correct running time into the Length field of the Constant Stretch tab. The figure for the Ratio will change automatically. Select Resample, High Precision, since you want to preserve neither the pitch nor the time.
Rosey: So they are the same length when finished.
Chris: Yes.
Rosey: That's a great idea!
Chris: I have used this technique and it works. Even though I have an 8-channel sound card, I have cassettes that were recorded with a machine I no longer have. It used Dolby B noise reduction, and the only way I can get that is to use a stereo tape recorder.
Rosey: Ah, ok.
Chris: In that case, I play two of the tracks forward, flip the tape over, and play the other two in reverse, then reverse the two backwards tracks in CEP. Then I find the corresponding peaks and do the resample procedure.
Rosey: I see. This is all new to me, very interesting.
Chris: Once your son has the four tracks on the computer he can pretty much do as many overdubs as he wants. At some point however the CPU will reach its limit, but then all you need to do is bounce tracks down. The loss of quality will not be noticeable even in 16 bits, at least I don't hear it. If you try to bounce tracks even once with a cassette recorder you hear a pronounced degradation.
Rosey: What does that mean, bounce them down?
Chris: Means like if you have, say, 5 background vocals, you can combine them to two tracks in a sub-mix.
Rosey: Oh, two tracks instead of five? A compression, of sorts?
Chris: Well, I suppose you could call it that. It's only necessary if the program starts to perform badly, since that will most likely happen before the program runs out of available tracks. But as I said, I have had 13 tracks playing in CEP while recording 2 tracks in Cool Edit 96, sending 8 outputs to a mixing console and two back into the sound card, and the PC did not struggle.
Rosey: I told him that...he was impressed.
Chris: The disadvantage of doing that is that everything is again converted twice. The eight tracks go through d/a conversion on the way out, and the stereo master goes through a/d conversion on the way back in.
Rosey: What is that, a/d and d/a, please?
Chris: Analog/digital digital/analog.
Rosey: Got it, thanks.
Chris: Another advantage of bouncing tracks on PC is that you can retain the original isolated tracks. With tape you usually have to decide at some point to erase those tracks.
Rosey: I see.
Chris: It is also best to undo transforms that are unsatisfactory and do them again differently, than to do compound transforms, especially in 16 bit audio. CEP gives you the option to convert to 32 bit for doing multiple transforms, but eventually they have to be dithered back to 16 bit to be used for anything practical, and this either adds some noise or distortion. But doing many transforms in 16 bit is worse, because the dithering is more severe. Like, say you want to reduce the amplitude by x percent. All the samples in that passage will then be dithered to a close approximation of x percent, because in 16 bit there are only a few thousand possible values. When you do this over and over, the whole thing starts to sound like shit.
Rosey: It loses integrity, is what you are saying, then.
Chris: Yes. The basic idea is to do the transform, and if its not right, undo it and do it again differently.
Rosey: That makes sense.
Chris: That one idea is very important to most everything you could do in digital audio.
Rosey: I found out a small bit about this when I was increasing amplitude in wavs, so I think I am following you here.
Chris: Also, unfortunately there is no way to judge the overall level in rendering a mix in CEP. You have to render it first, and almost invariably the first trial is way too loud. In 1.0 you have to go and reduce the volume of each track individually by the same amount. In cep 1.1, though, they added a master volume control.
Rosey: That reduces all at once, then.
Chris: I would imagine so, yes.
Rosey: Seems reasonable.
Chris: A lot of times if it is something simple, I dont even use the multitrack mode. I have done entire pieces just using precision mix pasting in CEP. It It is a feature in Cool Edits that allows you to add a sound to another without affecting what is underneath.
Rosey: OK, I have done that with wavs, understood.
Chris: using the Sample Display Time Format and the Time Display Fields its possible to do it precisely. So, if you have a selection in one track and want to get to the same start position quickly in another track, you can copy the figure from the Time Display Field for Sel/Begin by clicking on it. Then select Viewing Range from the View menu, and paste the figure into the From field, and the window will jump there. Select what is displayed by clicking Ctrl + Shift + A, and hit F8. Now the start point of the current selection is the same sample as that of the other track.
Rosey: OK.
Chris: It's possible to do the same thing by moving the cursor and zooming repeatedly, but this is quicker.
Rosey: These are good tips, thank you.
Chris: I suppose you already know to go to the window menu if you have more than one file open.
Rosey: Yes :)
Chris: It's different in cool edit 96, you can only have one file open, but multiple instances of the program.
Rosey: Yes, I have seen that.
Chris: To make a good master, once you have rendered a mix in CEP, firstly, if there are too many clipped samples, or if it's way too low, do the mix over. If its basically OK but there are some really high transients, you can use an envelope to reduce them. In my presets they have names like 99% insert, 80% insert and so on. They are shaped so that the beginning and end values are 100% but in the middle they dip.
Rosey: Are you talking about volume here?.
Chris: Yes the envelope affects the volume.
Rosey: OK, good, I am following, then.
Chris: The quickest way to get to a peak is to select around it and use Analyze/Statistics. Among other information, it will show you Peak Amplitude. Click the corresponding arrow button and the cursor will jump to the peak. So then you zoom in on the peak using Ctrl + Home, and select just a few samples around it, preferably down to each surrounding zero amplitude sample, and apply one of those envelopes. This will reduce the peak and enable you later to amplify to a much higher level without clipping. The result is either unnoticeable during those transients, or it makes it smoother.
Rosey: This makes sense.
Chris: You can do the same sort of thing globally using the dynamics processor, but it took me a long time to do anything productive with that. It is an effective compressor, but I did not find it intuitive at first.
Rosey: The dynamics processor is a part of CEP?
Chris: Yes, and of Cool Edit 96.
Rosey: OK, I am unfamiliar with it.
Chris: You can find it under Amplitude in the Transform menu.
Chris: Oh, but it does have some very good uses besides broadband global compression. For instance, one of my presets - actually a few - are for frequency-specific limiting. Sometimes, to get presence in a vocal, you will boost some high frequencies. But doing this can emphasize sibilance.
Rosey: What is sibilance in this context, please?
Chris: Every time there is an "s" or a "t" consonant in the vocal, the high frequencies are too loud.
Rosey: Oh, OK, understood.
Chris: To control this, you apply a high frequency limiter to the vocal track. First, select a very brief portion of the vocal track just where this sound occurs. Then use the Frequency Spectrum Analyzer to determine the frequency content of the selection. You'll see it curve down gradually from the low end, then suddenly around 3000 Hz or so it will jump way up. Point the cursor at the two sides of the peak and it reads out the frequencies. Then close the analyzer and open the dynamics processor. Enter the frequency values for Low Cutoff and High Cutoff. Choose peak limiting with low values for Attack Time and Release Time, and a hard knee compression. When you apply it to the vocal track it will affect only those frequencies you entered, and the sibilance should be under control. The same thing is done in analog, but a key input is used with an outboard compressor. It's sometimes called a de-esser, but it's more accurate to call it high frequency limiting or hf limiting.
Rosey: That term I know.
Chris: In analog, the vocal track is split; one lead is sent to the compressor input and the other goes through a parametric equalizer or other filter. In the equalizer, the high end is boosted and the mid and lows are cut. Then the output of the equalizer is sent to the compressor's key input. The key input is not heard in the output, it just tells the compressor what frequency range it should act on. It is a little different though, since in analog the compressor is always affecting the whole signal. The key signal just tells it when and how much. In the CEP Dynamics Processor it is in fact only affecting those frequencies you enter as the cutoff frequencies.
Rosey: OK.
Chris: Anyway, back to the post-mix editing and mastering phase. After dealing with unwanted peaks, you might want to amplify or equalize the whole mix.
Rosey: Yes, I might.
Chris: Or you can trim the excess from the head and tail, the sequence doesnt matter.
Rosey: OK.
Chris: But you want to do those two things before applying envelopes for things like fade-outs.
Rosey: Why?
Chris: Because, well, say you apply a fade-out first.
Rosey: Yes.
Chris: The last sample has a value of -96.32 dB in 16 bit, or effectively zero amplitude. When you then amplify, the whole floor is raised. So you no longer have a complete and smooth fade. So, if there are to be fade-ins or fade-outs, they should be done after the waveform is normalized and trimmed. The last thing is to add a brief period of silence to the head and tail. This prevents the beginning and end of the sound getting cut off when played in software or in a CD player that ramps the sound. Ramping is done to prevent any loud clicks when the device initializes. So the output is muted briefly. A quarter or a half second of silence is an adequate margin to prevent the sound being partially cut off. In CEP you'll find silence under the generate menu. Of course, I have a button for it on the toolbar.
Rosey: Oh, is that what silence is for? I have seen that.
Chris: Well, it can be used for other things as well.
Rosey: Such as?
Chris: For instance, when you want to add reverb to a track. At the end of the track you can add a certain amount of silence so the final decay of the reverb can fade out. The amount of silence you add should be a bit more than the total length of the reverb.
Rosey: Understood. I saw something somewhere that did that automatically, I thought.
Chris: Oh, one of the echo transforms has an option to delay beyond the end of the selection.
Rosey: Yes, that was it...echo...right.
Chris: But I think that only applies to a selection. If it's applied to the entire waveform I don't think it will automatically extend it. I'm not sure offhand.
Rosey: OK.
Chris: But to do all this stuff to 8 or more tracks requires a lot of dedication, especially when the program crashes and you have to start over.
Rosey: Ack...no crashing allowed.
Chris: That's why Cubase is a lot less tedious, the DSP is real-time. All of it can be done without saving new sound files. You can save numerous mixes of the same sound files and the mix directions are put in a file that is usually less than 100 kb. Well, CEP has its own format for that also, but it only saves the track volume and pan and sound card assignments. And also track positions on the timeline.
Rosey: That's probably sufficient for our needs at the moment. Either way, this will be fun.
Chris: One feature that Cubase does not have is sample accuracy. Everything in CEP can be done with sample accuracy, positioning of tracks and so forth.
Rosey: I see.
Chris: You can even manually alter the amplitude of an individual sample with the mouse.
Rosey: Oh? A right click menu?
Chris: No, just zoom in on the sample and left click on it, then move the mouse.
Rosey: Easy :)
Chris: I sometimes use that if I have three or four clipped samples and the envelope somehow screws it up.
Chris: I can unclip the samples manually.
Rosey: sounds simple enough.
Chris: {S simple.
Rosey: hehe
Chris: There's a lot we haven't even scratched the surface of, such as all the filters and equalizers and so forth. The FFT Filter is effective for treating, say, a 60 cycle hum. You can use it as a notch filter or a bandpass filter, and the effect can be changed over the course of a selection. You have hiss reduction, which can often be effective but it takes a lot of experimentation to get good results. Same with the click and pop filter, which is best for treating transfers from vinyl records, but can also remove glitches from other sources. Eventually I guess we'll get around to covering more features in detail...