[AmpDesigner333]

The "super simple" version for preamps:
    - Lower output impedance is generally better
    - Higher input impedance is sometimes preferred, but not as important as output impedance
    - Most important spec: SNR
    - Hard to find: DC coupling throughout

The rest is about your preferences, like single-ended/balanced, number of inputs, etc.

Personally, I like to bypass the preamp altogether (Digital Source --> D/A --> Amps).  There's nothing quite like DC coupled, balanced, and short signal path combined.  Remember, anything the signal goes through will add noise and distortion, however small.  Therefore, digital domain volume control has an advantage as well.

Many of our Classic Cherry customers like the synergy with tube preamps, but Maraschino customers are mostly using SS preamps or no preamp.  A tube pre might shave off bandwidth (so check freq resp) and erase the benefit of super high bandwidth amps.  Thanks for reading my post (:

[Steve]

http://www.decware.com/paper55.htm

Decware's paper leaves some to be desired. First, as mentioned earlier, a 10 to one ratio is nice and what I recommend.
With almost all situations, except extended ICs, an output impedance (Z) of the preamplifier can be
2k to 3k ohms. I have a white paper still up on my website explaining such listed below.
 
One has to weigh the advantages of adding a buffer stage to the output as it will add artificial flavors to the music. However, it
does minimize the complex consideration of using just a passive control (even transformer type).

The IC capacitance should always be as low as possible. For instance, I found that an expensive MIT IC, which had three
positions switches, had a capacitance of 500pf, 750pf, and 1100pf (.0011uf) for 1.5 meter length. That is very, very high.
My own personals ones are 40pf for one meter.

A major aim is to keep the number of stages to a minimum, as each stage adds artificial flavors, degrading the performance.

Concerning passive resistance volume controls (also called passive preamplifiers), both the input Z and output Z are very sensitive, easily causing frequency response problems. One has to have high enough passive input Z so the bass response is flat. On the other hand,
the output Z and frequency response will depend upon the sensitivity of the amplifier input. The more sensitive the amplifier input is, the lower the output Z of the passive control, allowing for higher frequency response.

It can be a real mess as one has to consider the source output design itself, amplifier input Z, IC cable capacitance, and passive control input/output Z.

Transformer types have issues as well, such as ringing problems, and frequency response limitations.

Finally, a large problem is the use of a low gain stage in the source itself (let's say cd player or dac). This final analog stage usually has a gain of only 2 (6db) and is usually of cheap parts, so is a bottle neck in the audio chain. Without this cheap stage, the output is still approximately 1 vrms, and plenty for most applications. The only problem would be a passive control into an amp requiring more than 1 volt rms.

Integrates also have their weaknesses, although less complex, and probably more expensive.

http://www.sasaudiolabs.com/sasaudio.htm

Cheers and hope this helps.
Steve