OK...this one's a bit bizarre, but I wanted to know if other 990 users out there are experiencing a similar issue. I ahve a ticket in for this, but it sounds like the Outlaws are pretty busy with new product development, and as such, I'm trying to determine if the fault is in my unit (specifically) or if this is a global issue that was somehow missed.

I should mention that I first noticed this when the unit was new, but at the time, only on the FM tuner - I had no analog sources connected at the time - and I don't listen to FM much, so this kind of fell off of my radar.

Note: what I am about to describe does NOT happen with digital sources - only analog ones.

OK, so one day while playing a recording of waves on a beach (which, at times, approximates white noise) from my Turtle Beach Audiotron, when I switched to the "CD" input (which gets the analog signals from the Audiotron), the character immediately changed. I have the analog inputs connected so that this signal can be routed to Zone 2, which then feeds the rest of my house.

Anyway...with the analog input selected, I noticed a distinct comb-filtering effect. Before anyone out there attributes this to low bit-rate mp3 artifacts, keep in mind that my music library is in .wav format (uncompressed), not any lossy format. Also, that this was not present when listening to the digital inputs told me that it was not the source file.

Back to the issue...when I switched to the optical input, the comb filtering stopped. I switched back to the analog input, and the comb-filtering returned. Mind you...this is MUCH harder to hear with music, but the 'ocean waves' file, due to its nature, has properties that approximate white noise (at times); white noise is one of the most useful stimuli out there to use for diagnostics.

While still listening to the analog signal, I started to switch between stereo, bypass, and upsample modes. Though the character (timbre) of the comb filtering changed as I cycled through the modes, it was always there. The one exception was one of the Dolbly modes - in this mode (or maybe it was two of the Dolby modes...I have to check my notes) it sounded normal. However, when using the digital input, cycling through the modes did not produce the comb-filtering effect. I even went as far as to remove the Audiotron's optical 'send' (leaving only the analog connections in place) to see if there was some odd crosstalk between the CD input (analog) and the optical input. Regardless of whether the optical feed was present, this problem continued.

I then listened to the Audiotron's analog outputs using headphones (the headphone jack on that unit is normaled to the analog outputs) knowing that if the problem was in the Audiotron, then I would hear it over the headphones. Result? No comb-filtering whatsoever.

Hmmm...

So to help me diagnose this, I cooked up a white noise wav file and loaded it onto my iRiver portable player (which supports mp3, wav, FLAC, etc) and connected its analog outputs to the CD inputs. Sure enough, the same comb filtering that I hear while playing back the wav file from the Audiotron was present.

I then tried the other analog inputs (AUX, TAPE) and the very same thing resulted.

To document what was going on, I then connected a four-channel R-44 recorder to the outputs of the 990 - the first pair were the "front" signals and the 2nd pair were the "Zone 2" signals. I connected the analog outputs of the mp3 player (playing the white noise file) to the CD inputs and recorded about 30s of each mode (stereo, bypass, upsample etc). In this way, I would have a recording of each mode and what was going in in Zone 2 at the same time.

Result: All of the Zone 2 recordings are white noise and nothing but white noise - absolutely no comb-filtering effect was noted or showed up in the power spectra. However...what I had heard (from my speakers, which get the FRONT feeds) in all of the modes that I originally listened to while noticing this were (not surprisingly) present in the recordings.

Again...Zone 2 (which had been assigned to CD) was always absent this comb filtering.

The spectra reveal, conclusively that comb-filtering is taking place. Again, it's not all that perceptible with music, and you have to quickly switch between analog and digital feeds to have a good comparison, but the fact is, it is there - white noise is a great diagnostic tool.

How do you get comb-filtering? Well, it comes from 'delay and add' in a digital system. That is, if you look at the frequency spacing of the peaks in a classic comb-filter, the delay is proportional to that frequency's inverse. That is, a 1 ms delay and add, will produce maxima every 1000 Hz; a 2ms delay will produce maxima every 500 Hz...and a 10 ms delay will produce maxima every 100 Hz.

Incidentally, I was able to recreate this comb-filtering effect as a simulation (using the amount of delay noted in bypass, stereo, and upsample modes), and the resulting files have almost identical spectra as the files created from the 990's front outputs. Moreover, if you listen to the simulation as compared to what is coming out of the 990, you really cannot tell a difference. based on this, I am sure there is a 'delay and add' issue going on in my unit.

So...here are my questions:

a) when listening to FM (especially if off-station when the level of hiss is high), if you cycle the modes (stereo, bypass, upsample...etc) does the quality (the timbre) change?

b) has anyone tried sending white noise to their 990 via the analog inputs and cycling through the modes to see if they notice it?

Again, it's harder to hear with program material, but unmistakable when using white noise as a stimulus.

Comments?