Restoring process

[cb831]

Hi

I want to restore some old 3 3/4 in/s tapes to 44.1 / 16bit.

My strategy is to record using 24bit and set gain to peak around -6dB to ensure no clipping is occuring without having to be very precise on the volume control. Post process (except for cleaning up stuff) I will normalize gain and convert to 16bit.

Is this a good approach or will it introduce errors due to calculation and new quantization ? Am I better off trying to hit 0dB and use 16bit directly ?

Regards
Claus

[TimG]

Restoring process

Whether it is worth the extra processing time to record at 24 bit and convert to 16 bit, depends on the quality of the source (tape) material (among other things). Whereas setting gain to peak around -6dB to ensure no clipping is a good practice, then normalize to the level you want in the end. This gives you plenty of 'head room' for filtering and effects.

[Dan McDonald]

I would recommend 24-bit even in very poor quality situations because you can get more aggressive in some filters, as long as you have the space and the processing power to handle it.

[cb831]

Restoring Process

Thanks guys for the answers

How much headroom in general will you recommend for filtering ?

Claus

[Audyossey]

How much headroom in general will you recommend for filtering ?

I watch the meters when I'm in preview mode. The 6 dB you mention is generally enough, but successive passes of equalization can erode that headroom. Occasionally it may be necessary to renormalize at something like -9 or -12 dB. I've never heard any artifacts in this connection, even after successive normalizations. As a practical matter, the dynamic range in digital recording is much more than needed in most home situations. If you live in a city as I do, the noise floor allows for something like 60 dB and there are concert halls that don't do much better than that. Car environments are worse. So if you wanted to make -12 dB your standard, you'd probably be fine.

The LPs themselves will probably have a dynamic range of 60 dB or less. Your processing will reduce the noise floor to nearly nothing by reducing surface noise and tape hiss, but the music won't actually use it unless you expand its dynamics. To be sure, the mastering engineers may have played a role in compressing the dynamics but, as you can see from the foregoing, there are valid reasons for doing so even with CDs. There are times when practical considerations should trump the ideal of total "realism." Most musicians are going to limit the dynamics of their own performances instinctively anyway. So, you have plenty of room to play with headroom without disturbing the music.

For no good reason at all, I do the final normalization at -.1 dB. I just can't give up the gut reaction that something in the chain could screw things up. It's silly, but I do it.

HB