Hello everyone, n00b here..

[Shaman]

I'm new to DC6 (it hasn't even arrived yet) but not to digital signal processing as an art. I'm an amateur astronomer who has been using Astroart to process my digital images for about a decade now. I've also used Photoshop, GIMP and Paint Shop Pro extensively for image processing.

I've been playing with the demo for a while now and I have a couple of questions for the crowd here.

I've been testing my system by recording silence with various pieces of equipment hooked up and then playing back the recording into the Spectrum Analyzer. This is a common technique in astronomical image processing and it seemed perfectly logical to do the same thing in audio.

At the moment I have a Turtle Beach Santa Cruz sound card installed in my rather aged and lethargic 1.8 gHz Celeron box, although I have an Audiophile 2496 card set aside for the new dual 2.66 gHz Pentium D box I'm trying to debug right now (it looks like the motherboard is borked).

By watching the VU meters during playback, it would seem that my noise level is about 72 to 74 dB below clipping when nothing is plugged into the sound card. Is this a reasonable figure, or do I have a defective sound card?

There is a noticeable rise in the noise level below about six hundred Hz or so, is this normal?

It was quite illuminating to find out that my preamp (a Yamaha RX 500U receiver) is about 6 dB noisier without anything plugged into the phono input than when the turntable/cartridge combination is plugged in.

My overall noise level with everything plugged up is about 68 dB or so below clipping. Is this an adequate figure?

By generating waves and then running them through various filters I have noted that the brick wall filter seems to be less effective than the paragraphic equalizer set with a severe turn down at the selected frequency. Not to mention that you can keep the pass band flat with the paragraphic equalizer and there is a distinct hump in the passband of the brick wall filter right above the turnover frequency when the filter is set aggressively. My turntable has a decently low rumble figure but it would seem that a lot of the records I'm ripping have considerable low frequency noise that I'm trying to block out and apparently the paragraphic equalizer is a better tool for that than the brick wall filter.

Is this observation correct or is there something blatantly obvious that I'm missing?

Oh, and one more thing, the paragraphic equalizer runs a lot faster than the brick wall filter when it is set aggressively.

That's about it for now, I'm sure that when my CTP, test record and DC6 arrive I'll have many more questions.

Cheers,

Jon

[Craig Maier]

Hi Jon,

I will try to answer some of your questions.

--------------------------

By watching the VU meters during playback, it would seem that my noise level is about 72 to 74 dB below clipping when nothing is plugged into the sound card. Is this a reasonable figure, or do I have a defective sound card?

Answer: That is about right. In general, soundcards do not perform as well as they claim. The noise level will decrease when you connect it to the correct input terminating impedance. When you get your users manual, you can find an equation for calculating noise performance based on Boltzmans constant, terminating resistance and termperature.
--------------------------
There is a noticeable rise in the noise level below about six hundred Hz or so, is this normal?

Answer: That is normal because your RIAA preamplifier has a rise in gain starting at 500 Hz (referred to as the turnover frequency of the RIAA curve).

------------------------

It was quite illuminating to find out that my preamp (a Yamaha RX 500U receiver) is about 6 dB noisier without anything plugged into the phono input than when the turntable/cartridge combination is plugged in.

Answer: Any amplifier will exhibit a higher noise floor when its input in unterminated. That goes back to noise floor as a function of temperature, terminating resistance, Boltzmans constant and bandwidth. Again, the forumulae for calculating noise floors can be found in your users manual.

---------------------


My overall noise level with everything plugged up is about 68 dB or so below clipping. Is this an adequate figure?

Answer: Yes, considering that the best mechanical records only have a S/N of around 55 dB. The Diamond Cut software will improve on that situation.

-----------------------

By generating waves and then running them through various filters I have noted that the brick wall filter seems to be less effective than the paragraphic equalizer set with a severe turn down at the selected frequency. Not to mention that you can keep the pass band flat with the paragraphic equalizer and there is a distinct hump in the passband of the brick wall filter right above the turnover frequency when the filter is set aggressively. My turntable has a decently low rumble figure but it would seem that a lot of the records I'm ripping have considerable low frequency noise that I'm trying to block out and apparently the paragraphic equalizer is a better tool for that than the brick wall filter.

--------------------------

The brick wall filter is of the FIR type whereas the paragraphic, lo pass, high pass and bandpass are of the IIR type (analog simulations). You can effect the steepness of the Brick Wall filter by increasing the filter length. As for creating a rumble filter, consider using the Highpass at 24 dB/Oct, Butterworth response. That works quite well. The freq would depend on your turntable rumble freq. My turntable's rumble freq and resonance is 12 Hz. I set my HPF to 20 Hz at the mentioned slope which gives me roughly 20 dB of attenuation at the rumble freq. You can use the spectrum analyzer in hi res mode to measure the rumble freq. Butterworths produce a maximally flat response up to the corner freq with little or no bump. Greater attenuation can be had using the Chebyshev setting, but it does exhibit ripple in the passband as well as a bump at the corner freq.