My preference has always been SQ!QS's main downside for me is the stereo separation. Quite a narrow image. I do think I prefer QS to DPLII though. The Hotto QS Encoder is fun to play around with
Best method to downmix surround to stereo? Are LoRo fold downs outdated?
- Thread starter nickleb474
- Start date
Help Support QuadraphonicQuad:
One question - what is the surround sound to stereo downmix for (is stereo the final form or is it a carrier for later matrix decoding)?
Kirk Bayne
Kirk Bayne
I'll vouch for that. I've made encoded CD's from SACD's and DVD-A's to play in stereo in the car, and they decode just as Chucky says when played on the home system. I'm a believer!
CD-4 didn't have a center channel, Kirk. All of the center information was included in the front channels.
Kid, you just made a case for using the Involve encoder for (I'll use the term) RM-encoding from discrete sources. This encoder actually does have a center channel input when encoding from a 5.1 source, which won't be on vintage encoders. Using their "Intelligent Encode" mode, you retain the front channel separation. It's not reduced to 12dB, as vintage QS decoders did. As for it being "fun to play with", I'd put it right up there with the Hotto.
In my case, it's for both.
I think that the original intent of this thread was that the stereo downmix was for stereo use. If the need was for matrix decoding the LtRt downmix would have to be used. nickleb474 has stated that he doesn't like the phasey sound of such downmixes. The LoRo option is not ideal either.
nickleb474
Member
I may not be a big fan of matrix downmixes, however I'm still interested in hearing everyone's favorite matrix techniques! Maybe I'll pick up an Involve Encoder some day soon 
Sounds very cool!
Sounds very cool!
https://en.wikipedia.org/wiki/Haeco-CSG
The downmix issue has been around for a while, this CSG tech could be applied to summing L&LS & R&RS in a 5.1 mix to create stereo.
Kirk Bayne
The downmix issue has been around for a while, this CSG tech could be applied to summing L&LS & R&RS in a 5.1 mix to create stereo.
Kirk Bayne
I wouldn't do that it would just create a phasey mush. Neil Young hates it!
Basically, the downmix situation is hopeless, some content will be increased in volume, some reduced in volume and some lost.
A story about how SQ matrix encoding causes losses (of course, the CD-4 type downmix causes losses too):
https://www.wendycarlos.com/surround/surround3.html
^^^
(A Saber-cut to the Heart of the Matter)
It seems to me that the CD-4 type downmix would be best for your need (5.1 L&LS->stereo L & R&RS->stereo R [5.1 CF&LFE->L & R equally])
Kirk Bayne
A story about how SQ matrix encoding causes losses (of course, the CD-4 type downmix causes losses too):
https://www.wendycarlos.com/surround/surround3.html
^^^
(A Saber-cut to the Heart of the Matter)
It seems to me that the CD-4 type downmix would be best for your need (5.1 L&LS->stereo L & R&RS->stereo R [5.1 CF&LFE->L & R equally])
Kirk Bayne
nickleb474
Member
What you're describing sounds to me like LoRo without the rears attenuated by 3db. This technique would suffer from issue #2 as I described here:
Though my own technique that I describe a bit later in that post is essentially what you're suggesting and it avoids this issue.
Strongly disagree, hardly hopeless. That type of downmix, CD-4 type (LoRo) is not optimum as it tends to place everything full left, full right plus centre. I would prefer an SQ encode but that introduces phasyness. The alternative is to pan the rear channels toward the centre. That will image like an SQ mixdown but without phaseyness. Alternately you could narrow the front separation and place the rears full left and right. LFE can almost always be dropped, full bass should already be in (at least) the front channels. Adding too much bass will muddy the sound.
Last edited:
That could be a problem for sure. Side images aren't that common but could be troublesome especially with pans. Mixing as I described would not eliminate the problem but would certainly reduce it.
One puzzling thing is that the official Dolby Digital 5.1 -> Dolby Surround downmix process (which matrix encodes DS into stereo) throws away the LFE "channel", I've never understood why Dolby Labs decided to do this, it seems like LFE should be added to L&R.
I only recall one case where the quad encoded stereo (SQ matrix, IIRC, the Boulez conducts Bartok album) was taken into consideration when preparing the quad mix.
Kirk Bayne
I only recall one case where the quad encoded stereo (SQ matrix, IIRC, the Boulez conducts Bartok album) was taken into consideration when preparing the quad mix.
Kirk Bayne
That is because of what I just said in my previous post. The bass is already contained in the main channels there is no need to add more the lfe should simply be discarded. Don't forget that if the bass is in the rear channels as well it will add in the final mixdown as well. Too much bass will make the mixdown sound muddy.
https://www.quadraphonicquad.com/fo...d-it-ever-sound-good.32306/page-4#post-625723
If the SQ matrix system is used to create the 5.1->2.0 downmix, I recommend swapping the way LB and RB are encoded into Lt and Rt (see corrected image attachments).
Kirk Bayne
If the SQ matrix system is used to create the 5.1->2.0 downmix, I recommend swapping the way LB and RB are encoded into Lt and Rt (see corrected image attachments).
Kirk Bayne
As long as the mix is not dynamically compressed, I would say 90% of the time, just a simple down-mix without messing with individual channel levels before down-mixing works good enough for me, and provides an improvement over the CD/streaming version, which is what i'm looking for with car listening. On albums that have a lot of vocals coming from the center, I will often discard the center completely. I check the results in Audacity and normalize to -.1 db. If the waveform shrinks during that process, then something is clipping, and I will go back to the multi-channel and lower some channel volumes, then try again.
national-kid
Active Member
Regarding listening to QS and SQ sound sources in stereo,
wherever you listen with QS, rear sounds are localized outside the left and right speakers, respectively, and the position of the sound source can be roughly pinpointed.
However, with SQ, rear sounds are localized inside the left and right speakers, but depending on your listening position, they may be localized in the opposite direction, with the right rear being left of center and the left rear being right of center.
So, while there are no problems with front sounds with SQ, it is very difficult to pinpoint the position of rear sounds, and I don't think it can be said that its compatibility with surround and stereo is higher than that of QS. (Even before that, there is a fatal flaw in that it cannot transmit dead centers...)
wherever you listen with QS, rear sounds are localized outside the left and right speakers, respectively, and the position of the sound source can be roughly pinpointed.
However, with SQ, rear sounds are localized inside the left and right speakers, but depending on your listening position, they may be localized in the opposite direction, with the right rear being left of center and the left rear being right of center.
So, while there are no problems with front sounds with SQ, it is very difficult to pinpoint the position of rear sounds, and I don't think it can be said that its compatibility with surround and stereo is higher than that of QS. (Even before that, there is a fatal flaw in that it cannot transmit dead centers...)
Similar threads
