- Take Picking Chart[+]
Download the Take Picking Chart
Download These Instructions
In some browsers, you may need to right click on the link, and then select "Download" or "Save Link As..."
I wish I could take credit for this method, but the idea came from pianist, Allen Reiser. It's a solution to the problem of wanting to jot down comments while reviewing takes, but winding up with a score that is all but illegible because each comment must also include some indication as to which take it refers. An added issue is that while time is spent concentrating on writing a comment, focus on the music is lost, requiring more repetitions for thorough analysis.
The goal is two-fold: to provide a suitable receptacle for comments that makes it extremely clear which comment is attached to which take, and to establish a form of code - quick, easily made marks that can be jotted down without losing concentration on the music.
The first step is to set up your playback so that you can easily audition a number of takes, allowing you to quickly and easily compare the same passages in each. Recording sessions usually have a few takes that cover the whole piece or movement, and then several shorter takes covering problematic sections. It can be helpful to have some idea where and when the material in each take falls in work as a whole. My recommendation is that you use Apple's Quicktime - a free utility available for both the Mac and Windows operating systems. Quicktime allows you to have several files open on your desktop at the same time, and moreover, you can change the horizontal size of each player window. With this feature, you can stretch a complete run-through across your computer screen, while shorter takes can occupy the rough duration that they cover, and be positioned so that you have a vertical alignment of the music across all your takes. Once that's set up, you can start playing one file, then quickly jump to another one, or audition the same passage from several takes in quick succession. Quicktime only allows one player to be active at a time, so just clicking on the time line bar in each window will start playback from that point, while muting whatever was being played previously.
To begin making a record of your impressions of the takes, first print out the blank chart enough times so that you have a horizontal division for each measure of music. Each horizontal line then becomes dedicated to one individual take, with the bar numbers for all the takes lining up vertically. The decision to make one line per take and one box per measure is arbitrary. If you only have a few takes, you can save some paper and use the upper half of each page for sixteen measures, and the lower part for the following 16 measures, and so on. Also, don’t be tied to the “one box per measure” scheme - if you need finer resolution, by all means have each box represent a single beat in a measure. Make the chart work for you, not the other way around.
Find a way to notate your judgement of each part of each take in such a way that you don’t get so caught up with writing that you miss what happens in the next four or five measures. I suggest that you use little symbols that can be quickly notated, so that you can continue concentrating on listening. If you're familiar with Pitman Short- or Shorterhand, you may be way ahead of me. For those who are not, I’ve found that an arrow pointing up can indicate good things, one pointing down indicating bad things, and an “N” to indicate intrusive noises that you’d like to avoid. You can probably come up with some other shapes, but the idea is for them to be quick to write but informative, so that you don’t lose concentration on the listening. The alternative is to listen to each take over and over and over…
The best, most inaudible edits occur in the places where a listener will be least likely to hear them, so choosing transition points is key. Putting an edit right on the dividing line between musical phrases or sections is almost always audible -- it’s much better to hide the transition inside something busy. It's easiest to edit at a point where the music shifts abruptly from soft (or silent) to loud. The most difficult edits occur at points where the music goes from loud to soft, as it’s very difficult to maintain the same ambience and reverberation, and the shift in ambience is almost always audible. Similarly, “cold starts” are to be avoided, where you have a take which starts in the middle of the music, and you place the edit right on its very first note. Since this take contains no overhanging reverberation caused by performance prior to this point, but the take onto which this one is to be joined has the ambience is fully active, the edit will sound like all the air was suddenly sucked out of the room when the prevailing reverberation suddenly stops -- VERY audible.
It’s very easy to get caught up on notating all the problems, but if you do just this, you’ll wind up knowing where all the WRONG material is located, but not know where the good performances are - more to the point, the BEST performance. It’s no good knowing that you have three takes with a good Bar 47 in them -- which one is the best one? It’s probably more important in the long run to have notations to help you decide which takes contain the best parts to be stitched together in the final edit. You want the best of the best.￼
Once you’ve gone through the takes with the chart, you might want to listen again, this time to make sure that your preferred takes really are the best of the best. Only when you’re sure, take a highlighter pen and see if you can draw a line through the best parts, jumping vertically from take to take so as to create an unbroken line of good performance. If there’s a part where there is no “best”, well, then we missed, and we’ll have to either pick from the “least worst” or cobble something together somehow…
Once you’ve figured out where and when to edit, you need to transfer this information to a score, so I can edit while following the music. The notation I prefer is to draw a large “T” through the music, where the vertical stem of the “T” passes through the notated music at the point where the edit is supposed to happen. Then - and this is crucial - write the number of the take that we’re moving FROM in the area just under the horizontal crossbar of the “T” to the LEFT of the vertical stem. Then, write the number of the take that we’re moving TO in the area just under the crossbar in the area to the RIGHT of the stem.
When you’ve finished, you can check your work by making sure that all the “To”s become the following “From”s -- 86 to 24…, 24 to 43…, 43 to 19…, etc. etc.
A slight variation is to make the crossbar of the “T” somewhat wavy in shape when there is some latitude in where to place the edit. This reserves the straight crossbar for those instances where the edit absolutely HAS to be at the point notated.
- Test Tones for Compression/Limiting
[+]Test Signals are provided in standard WAV file format, single track (mono), 24 bit, 96 kHz., - at no charge. I and LockwoodARS are not liable for the consequences of mis-use, so keep playback levels low to avoid any possibility of damage to hearing, or playback equipment. Download the Test Signals, HERE.
In order to see the effects of a Compression or Limiting plug-in, mount one of the Test Signals to a single, mono track in your preferred Digital Audio Workstation, and then mix-down through the plug-in you’re investigating. Mount the resulting file to a second track in the DAW, preferably right below the first to allow easy A/B comparison at various zoom levels.
Zoom out to view the effects of different Attack and Release timings on the Envelope (amplitude through time) of the signal. Provided your DAW has the capability, it’s advantageous to View the zoomed-out signals in a “Half-wave” or rectified form as it makes the Envelopes easier to see.
Any of the intermittent signals such as the “60 Hz -3 dB with 1 sec decay”, are best for this application. As the name suggests, each tone (or noise) starts abruptly at a level of -3 dBFS (peak) which is sustained for one second, followed by a linear decay which also lasts one second. You should be able to see how a slow Attack time can change the initial transient, or a slow Release time can change the signal’s decay from linear to something more curved, as well as how these controls can interact unpredictably.
Zooming in to the point where you can see individual waves, you should be able to see any deviation from the pure sine-wave shape, resulting in varying degrees of audible harmonic distortion. If, for example, you are testing a Hard Clipping plug-in, you should expect to see the flattened tops and bottoms of the wave shapes clearly. A Spectrum Analyzer can provide more detailed information. Testing with the Sine Tone Sweep signal can reveal changing amounts of harmonic distortion across the audible spectrum. With a Clipper, expect to see the addition of odd-numbered harmonics (fundamental x 3, x 5, x 7, etc.). Spectral analysis can also assist when testing the effects of what I call “mystery” controls - parameters which have proprietary or uninformative names, or otherwise lack adequate explanation of their function.
A very good (and Free) real-time Spectral Analysis plugin is Voxengo’s SPAN. Apart from being a satisfied customer, I have no affiliation with Voxengo. Downloaded SPAN for free, HERE.
Of course, the final test is to determine whether a perceived improvement in the sound is worth the inevitable trade-off of sonic degradation. This is best done by ear, with typical program material such as a soloed vocal or instrumental track. Have fun.