Syntrillium Backup Forums :: View topic

Havoc · Posted - Wed Aug 21, 2002 11:17 pm

Last night, I was thinking of adding a non-RIAA output to my phono preamp for work with 78rpm. Just then I thought: why not always record flat and apply RIAA in the digital domain?

Not that it would be very hard to switch between flat-RIAA, but would there be advantages of doing the RIAA in the digital domain?

post78 · Location: USA

I'm not a vinyl man, but sinse the general agreement around here is "fix it at the source", why not just let the appropriate hardware handle the filtering? Is the preamp not very good?

jonrose · Location: USA

Hi Havoc,

If you really want to try it, you might check out the criteria that Detlev posted some months ago, if you haven't already found it:

'RIAA Curves'

Best... -Jon

Havoc · Posted - Thu Aug 22, 2002 3:32 am

Post: mostly I agree to fix it at the source and the preamp is good (should not deviate more then 1dB). The thing is, I do not know enough of vinyl myself to answer my questions.

Jon: just because the Detlev EQ is so easy to do, the question follows: is it an advantage to do so? It certainly is if you are dealing with 78rpm material that does not follow standard riaa. But why no always do it?

Now my reasoning is as follows: IF we can record at a higher analog input level without riaa (then we would record the same material with riaa), and then apply the riaa when the data is at 32 bit float followed by a normalise later on, would that not be equivalent of recording at an even higher input level? Notice the big IF. So the question is, would this improve the S/N or not?

AndyH · Posted - Thu Aug 22, 2002 2:15 pm

two cents from someone even more ignorant.

Doesn't the output from a cartridge rise significant with frequency? The phono amp equalization is to make the final result reasonably flat, i.e. the final output matches the original input, in spite of the cartridge's wierd behavior.

Without that equilization, which, I think, can be as much as 60 dB over the audio spectrum, might you not have to record at so low a level, to avoid clipping the higher frequencies, that the lower frequencies would be making poor, inefficient use of your available bit depth?

jonrose · Location: USA

I am sure someone who is better with this stuff will answer you Andy, but I'd venture a guess that an adequate analog signal path and a good, clean 24-bit card would at least get everything in there - as it appears off the cartridge.

Whether or not digital refinements will produce similar results is another question, which is what Havoc is entertaining with this thread.

I don't do these types of restorations, so I'm going to butt-out and lurk, but I must admit this is interesting, and I do have a hard time believing that the difference between an analog curve-correction vs. a digital one wouldn't be (at least, slightly) audible in a side-by-side test.... But this may just be another one of those subjective things!

:???:
Smile

All the best.... -Jon

AndyH · Posted - Thu Aug 22, 2002 4:20 pm

Yes, it probably will all be captured, and I realize it isn't exactly the same thing, but if the lower frequencies are down 60dB, isn't this somewhat like recording (in general) at -60dB rather than a near 0dB? It doesn't seem to be making optimal use of the available noise floor of the soundcard.

Some phono preamps in general use probably don't have more than 60 dB signal to noise ratio, although Havoc's may be much better. Also, the background LP noise on anything I record is no where near 60dB down. It, of course, is also effected by the equilization, so perhaps that does not matter (?). Some experiments might provide interesting data (I don't see any way to switch out the equilization in my preamp, so I can't do any such experiments).

I also realize that the 60dB difference only applies to signals that were, as captured in the original recording, all at the same level to begin with, and that is not a description of real music, but the general principal?

jonrose · Location: USA

Cue Detlev.....
Unfortunately, I haven't seen him around here lately, but I'm sure he'd have something interesting to say on this.
Smile

(BTW, whatever happened to uru?)

Graeme? Have you any thoughts on this one?

Best... -Jon

post78 · Location: USA

SteveG · Location: United Kingdom

We did actually cover this ages ago - but I can't find the thread. The basic problem you have to overcome is to make sure that however you do the EQ, that the cartridge is correctly terminated into 47,000 ohms, (47k) otherwise all bets about the eventual output flatness are off! Most mic preamps, whilst they may well have the gain to amplify a cartridge, are designed to work at much lower impedances than 47k, and will be rather noisier with this high an input impedance.

So the real answer to Havoc is that if you manage to remove the EQ from your correctly designed preamp without destabilising it (which is usually possible if you know exactly what's in it - ie, you have a schematic), the yes, you can apply the correction digitally in CEP, and yes, in some circumstances, there might be an advantage. But you really do have to get the electrical conditions correct, otherwise the results will probably be worse.

Steve

Havoc · Posted - Sun Aug 25, 2002 12:36 pm

I wanted to do some tests this weekend, but it did not happen. As for the preamp, yes, I have the schematics, and the board layout. Did both myself, so it will not be that difficult. It consistes of a differential pre-preamp, followed by the riaa stage. So bypassing the riaa will not influence the loading of the cartridge. Now that I know it can be beneficial, I'll test it for sure sometime this week, whenever I find time (this is in fact the hardest part of it all......).

Skinah · Location: Australia

Have a look at this product.

http://www.audiosupermarket.com/Digital$20Sound.html

It should be about US$90 ! Some music stores in the US carry them in stock.

Edited by - Skinah on 08/25/2002 8:22:40 PM