Requests for Version DC8 Product Inputs

[sus4chord]

One thing I do a lot is to highlight a spike (that's a musical spike, not a tick or pop), and then do a (-2) decibel gain change. Going through an album after I record it and doing that to the 3 or 4 places where that happens lets gain normalize actually work well (to bring the sound up to CD player levels). It would be nice to have either a button or a hot key (sort of like the interpolate hot key) to quickly do a gain change based on the current setting. That way, I can get through the file a lot quicker. All it takes is one short spike anywhere in the file to prevent normalize from increasing the gain of the rest of the file.

[Craig Maier]

Hi John,

I am curious - - - what happens if you gain normalize to +2 or 3 dB first and then gain normalize to -2 dB? I have never tried it, but that may solve the problem.

Craig

[sus4chord]

Haven't tried that, but in viewing the wav form, it's pretty obvious where the spikes are. Sometimes I end up reducing a spike a couple of times, and once done, that usually makes it so the song normalizes well. And it's usually just less than ten or so to do.

I know I can apply a compressor to bring everything else, but I end up losing the dynamic range of the songs doing that. I prefer to eliminate the few points where the song spikes, then bring the rest up, and preserve the dynamic range of the music. But that may just be me.

Maybe I'm the only one that does something like that, but I've found it works well and I don't notice the difference in the music.

[Craig Maier]

Hi John,

Try my suggestion; I think that it will work nicely without having to do any manual work.

And, if there are only a few in a song, then your ear will not detect the clipped peaks. You need to have numerous clips close together in order for the human ear to "integrate up" and hear that as a distortion.

And, even if there were numerous clips produced by normalizing to +2 or +3 dB, you always have the ability to de-clip it after you normalize to -2 dB.

Try it, you may like it!

Craig

[Dan McDonald]

I've done this also, where you get very fast peaks that look like clicks but are just music. I'll try Craig's technique!

Dan

[quartet87]

I work almost exclusively with 33's and frequently encounter the same issues as John does. Heretofore I've been using another app which has a "hard limiting" (clipper) to reduce a couple or a few spikes, store that file, then open it with DC7 and gain normalize.

An undesirable side effect of that clipper is that it visually and audibly changes frequency levels which are well below the numerical clip level I've entered in the window i.e. if I set the clipper to -1.5dB the result is a slight and acceptable change in lower volume frequencies, but if I go to -3dB the changes are more radical...and the changes manifested is a gain in the lower frequencies and attenuation of upper frequencies, most notably in sibilance which were good in the un-clipped file...and those frequency gain changes are evident before I normalize the clipped file.

So I'll give Craig's method of clipping a try and let y'all know how it goes.

[Craig Maier]

Hi,

My recommended method is essentially a "Hard Clipper". That is the reason why we extened the gain normalize range of control to positive values of dB.

Craig

[Dan McDonald]

Here's a related one:

Sometimes with digital recording, people push the levels too high and you get a flat spike at 0dB. You can fix this somewhat with the declipper, but I found that if I try just the de-clipper without reducing the gain first, it doesn't really do anything much at the peaks; when I reduce gain 2-3 dB and use the de-clipper, I'm still left with a flat spot in the peaks, although that flat spot is smaller.

I get better results by reducing the gain by about 2-3 dB and then finding the spikes and interpolating by hand, which is fine if there are only a few spots, but a pain if there are a lot.

How about a filter that is designed to interpolate this kind of spike? I assume it can be done with adjustment of the de-clickers, but a specialized filter might be easier to figure out. It could be designed to be put in a multi-filter with a gain reduction.

Dan

[Craig Maier]

Hi Dan,

The De-Clipper is an interpolator of a special sort. You can adjust the degree of roundedness produced at the peak by raising the "Strength" Control. Also, I have seen times when it is necessary to reduce the gain to as much as -5 or -6 dB in order to have enough head-room to obtain a proper de-clipped interpolation.

Craig

[Dan McDonald]

OK - so the de-clipper should be able to take care of it as well as manual interpolation if the gain has been reduced enough?

Dan

[Craig Maier]

Dan,

I think so.

Craig

[quartet87]

Craig, your method of clipping works wonderfully. As most of my work is on selected cuts from 33's for airplay, I usually apply a light amount of clipping to maximize volume while preserving analog dynamics. Your method worked very well on a few files recently which happened to need little clipping. I normalize all my files at -0.5dB and use the Classic mode, so I first normalized 1 or 1.5dB, always clicking yes to save undo, then I normalized again back to my usual -0.5dB...again clicking yes to save undo. I'm glad to remembered to save undo because I did use it as the first clipping on any given file was either too much or not enough or I was unsure and wanted to try a different value. I'm sure I'll encounter files in the future which will require higher normalization values for clipping.

Now I don't need to outsource clipping to another app, which changed lower volume frequencies in process...the Craig DC7 method does not alter those lower level frequencies...it produces only what its supposed to and that is clipped audio.

[Craig Maier]

Danny,

That is good to hear; it is working as intended.

Craig

[ebrjvd]

Automatic speed/pitch adjustment to mains hum

Due to mechanical tape speed variations, hum components from tape transcriptions often deviate from what one would expect.

A feature that would be most welcome is an automatic speed adjustment to the 50 or 60 Hz mains frequency (present in recorded hum around 50/100 or 60/120 Hz).

Possible complications:
1) Tape speed (hence hum frequency) might slightly change from beginning to end of tape

2) When a tape has been copied from another tape with different recorders with slightly different speed, there might be two different hum components, one from each recorder.


Another suggestion: would it be feasible to add a default mains frequency setting in "Preferences", e.g. for dehumming in 50 Hz countries?


Keep up the good work!
Jos

[ebrjvd]

Separate EQ for sum and difference channels (M-S)

Another idea:

Could you provide separate EQ for the sum and difference channels (M-S)?
This would allow "shuffling" as described by Blumlein and Gerzon.

You replied:

That is an interesting idea. I am thinking that it probably should be a completely seperate item under the "Effect Menu" since some other controls are required. I am also thinking that it only requires the 10 band graphic EQ, not the others.

Another useful application of separate M-S processing is reducing hiss in FM radio stereo recordings, which is present in the S-channel only.

Jos