Low analogue .1 output from DV29

[mozart]

In my DV29/Primare SPA20/Primare A30.2/5xSonus Faber Concertino/B&W PV1 setup, the sub level is extremely weak when fed analogue 5.1 signal from DV29. This connection is of superior quality (especially with Nordost Red Dawn interconnects) compared to feeding SPA20 a digital signal, but I have to boost the .1 signal in amp by about 20 dB (happily this is done in the analogue domain in SPA20, with no AD/DA nastiness in EXTERNAL mode), as well as turning the volume on PV1 to two o'clock.

When the amp is fed a digital signal from either DV29 or Xbox, this is not the case, nor did I have this problem with SONY DVP999ES.

I can set levels separately for all modes in SPA20 (DTS/DD/Other - 2channel and analogue external prosessing) so there's no major problems - but I am concerned about what boosting the sub level does to the s/n ratio.

Any thoughts? Something wrong with the DV29 analogue output? (The bass quality is good, by the way.)

[Ettepet]

Isn't this the same problem as mentioned previously for the dv27a? A drop of 18dB for 5.1 (or 6dB for stereo) when using bass management on the dvd/cd-player, by setting speakers to "small"? If so, this can be partly corrected by setting the "stereo + sub" signal to -10dB on the player. The original LFE-signal for 5.1 remains at 0dB however, so any boost to correct the bass signal on your amp cannot really fix things, because that original LFE-signal will appear unproportionally loud after the boost.

[mozart]

Thanks Ettepet. Why on earth does Arcam do this? This makes the otherwise excellent analogue stage utterly useless (unless speakers are set to large, that is)! John Dawson, please?

[Ettepet]

I would love to see an update that fixes both bass management problems as well.

[Jeff]

When I discussed this problem with John Dawson he said the lower sub level is a statutory limitation imposed on them by Dolby Labs because the DV27A is a source product. In a source each channel must not exceed a certain output, since with bass management the sub output is a sum of all the channels they have to lower the output to ensure the maximum output isn't exceeded. The main problem is that they don't reduce LFE the same way so there is a level missmatch between LFE and crossed over output.

I much better solution (IMO) would have been to lower the output of all the channels by 3db.

[mozart]

Thanks, Jeff. Sounds like a better solution. Let's hope for this to be changed.

[mozart]

Ok. This is killing me. After Ettepets kind reply posted above, I've set the stereo + sub to -10dB on DV29, boosted .1 channel by about 16dB in SPA20, and I've finally managed to get a relatively even frequenzy sweep for two channel material (I have a nasty 10dB peak at around 31 hz but that's room related, and doesn't come into play on the majority of 2 channel recordings), without loosing all of the 5.1 subwoofer output, but with these settings I still come up ca. 5dB short (THX optimizer) on LFE channel. This channel dropped by 18dB when all speakers are set to small. Ettepet and Jeff, are you shure the LFE signal isn't cut, as well? Maybe this has been changed on the DV29...
Arcam, please do something! Either drop signal for all channels by 3dB (as suggested by Jeff above), or even better, ignore the Dolby limitation. I'm shure they won't go to court. :-)

[mozart]

This was the reply I got from Arcam:
Dear Tor,

"I have discussed this with our DVD Development team, and as this is not a fault or a problem it will not be changed as it is the specification of the player and it meets the requirements of Dolby Labs. The actual drop when the speakers are set to small is -15 db due to the low pass filters needed to filter the signal to the LFE. If this was set any higher then it would start to clip, the team are not aware of any other player that does this differently as most where tested while the DV29 was being developed. Unfortunately as the vast majority of customers are satisfied with this there will not be a fix for it as they feel there is nothing to fix.

Thank you very much for your input on this subject."

I guess that's that. A monster drop of sub channel - Oh yes! We like! A small drop of all channels - don't even think about it. Let the subwoofer work like hell, s/n ratio singing in the rain. The subwoofer probably deserved it anyway.
Rich, well defined bass has left the building.

[bjoern]

Mozart, I must agree, this is very annoying. I am not part of "the vast majority".
Especially since this (and DV27A in my case) is a DVD-A player... I mean, you are supposed to play multi-channel material and manage the bass directly in the player. The result is a whisper from the sub when you are playing DVD-A's.
The Dolby Labs specifications can't really be meant to apply to DVD-A players. OK, it is still a source, but you are forced to handle bass management in the player since it is not possible to do anything with the signal in the processor when using analog 5.1 input.

[Stevesky]

I think Dolby are getting a bit of bad press here, there are reasons for this madness:

When audio from one channel is mixed to another, as in the case of redirecting bass to the sub channel the final sumation can result in a signal more than 0dB. This will result in clipping and will sound pretty horrible. There are several solutions to this:

1. Scale the level down in the digital domain so when you sum the redirected bass together it doesn't clip. Downside is that the audio output will be low compared to all other channels by the amount of headroom allocated. This is what Arcam have done in the DVD player.

2. Scale the level down in the digital domain so when you sum the redirected bass together it doesn't clip, BUT scale it back up in the analog domain so reference levels are met and the summed signal will go over 0dB. To do this properly you need variable gain in the analog domain so you can set the output gain based on the number of channels being redirected. This is done in higher end AV players but not normally done in DVD players. Some companies have a fixed (high) gain on the sub channel to cope with the worst case situation but this then increases the noise floor(hiss), hence lowers performance. Other compromise tricks are to have a few fixed gains that can be switched in using an analog mux based on number of channels redirected, one has to be careful about switching noise (click!) as output gains are changed.

Some AV processors have facilities to adjust the levels of the 5.1 bypass input so you can compensate for the level difference of the DVD player sub output vs all other channels.

[mozart]

But wouldn't a comparatively small drop (-3dB) of all channels (before they are redirected) be enough to prevent clipping, and a better solution than killing the sub output? In this way the redirected signals in sum would be -15dB. If the LFE signal also is lowered 3dB, as it should, the result is -18dB, enough for preventing clipping?

My Primare SPA20 does have level adjustments in bypass 5.1 mode, but I believe dropping by 15db and then boosting the signal again isn't the best bet for quality of the signal.

Thanks for your inputs bjoern and Stevesky. I was starting to feel a little lonely!

[Stevesky]

Apart from that decibel summation doesn't work like that, you could lower the outputs of all the other channels, but it will need to be more than 3dB and then you won't achieve Dolby reference on the main outputs. Also when playing CD's you don't really want the left, right outputs attenuated as you won't get the best performance from the DAC.

[mozart]

Thank you Stevesky.
Why isn't a total of 18dB enough, when 15 dB is? If the point is that if there's a high level of bass information in the main speakers, this could exceed the 3 dB drop, this will be the same with the 15 db drop in sub signal?
The 3dB drop of main channels in stereo isn't optimal of course, but a comparatively small price to pay, I guess.
Dolby reference is not a great concern for me.

[Stevesky]

You will find that there is already more than 3dB of attenuation before they are summed into the sub channel. For example to redirect just two channels into the sub channel requires 6dB of attenuation in the summation to avoid clipping.

The other problem with having a huge amount of output headroom on a source component sub channel is that it will probably clip the analog input of your AV processor. That's why Dolby prefer to keep the sub channel scaled down and then assume that it will be compensated in either AV processor or the subwoofer.

[mozart]

I understand that two channels needs 6 dB attenuation (L -3dB, R -3dB). So 5 channels need -15 dB, but still only -3dB per channel (L -3dB, C -3dB, R -3dB, SR -3dB, SL -3dB). That means the LFE (in 5.1 tracks, not the filtered sub signal) output needs to be attenuated by 3dB as well, just to stay at the same level as the other speakers. Which means a total of 18 dB attenuation for movie tracks, but a much more "democratic" attenuation, because each channel is only dropped 3 dB, before filtering the below-crossover signal. This way levels are all matched and with no need for boosting sub, limited dropping of s/n ratio for subwoofer signal, and limited extra work for subwoofer.
Would this signal be too loud? I still don't get it. Sorry Stevesky, I must be getting on your nerve.