SM2 vs SPECWeb vs PENTEO 16 Pro

[GOS]

Well, this might not be an audio title that folks here are clamoring for, that's for sure. But today I was doing the broadway cast album of "The Producers". I am not a big broadway musical guy, far from it, but I am a big Mel Brooks fan, and loved the original movie. I took my wife to see the play when it was on Broadway and at the time thought is was great as was most of the music, which Mel Brooks wrote himself. So we ended up seeing this play a few times in other places and I got the CD way back when so I know it well.

Anyway, my wife asked why I didn't have this for the car, and of course I don't like to take stereo in the car, so today I decided to finally take a crack at it and I ran it through Penteo just to have something for the car. But when I listened to it on the PC, it was pretty much the same info in all of the main channels. Very mudane and boring. Nothing was highlighted or low lighted. The lead vocal was in all the channels, as well as all of the other vocals. It was pretty much wide stereo. Not impressed. (Remember though, I was using defauts and DEFAUT.

A few days ago, Glenn asked me to try a SPECWeb parameter to see how it worked. He told me to enter -m4 when the program window says "Press Enter to use the defaults", which is what I always have done.

So today, I decided to try -m4

What a difference! The SPECWeb pulled out the orchestra, the background vocals, the sound effects, and kept the lead in the center. Very wide distribution. And hardly any artifacts, if any. (Remember, my ears aren't what they used to be, no freq's over 11K)

Anyway, it's totally different with SPECWeb than the Penteo on this one. Here's a look at the same exact song done with both programs. You'll have to take my word on the sound, but you can CLEARLY see that SPECWeb pulled stuff out of the stereo and put it into separate channels where Penteo just spread things around.

One of the amazing things on this tune is that at the end, Nathan Lane and Matthew Broderick are singing at the same time, and somehow the SPECWeb pulls Matthew out of the center into the Front Right, so the duet at the end is very discrete.

I really was blown away by this and will try the -m4 on tunes more likely to be listened to by the QQ Hordes!

Here's a look at the Wav files. Guess which is which...........


View attachment 44416

Hmm, cool Jon. Makes me wonder then....if punching in -M4 has that much of a positive impact, why not just make it default? OK, I'm sure I know the answer.....it simply might not work positively for all music. But yeah....cool.

 

[Plan9]

I wonder if Penteo's extraction isn't in some cases limited by the fact that it has to fold-down back to stereo perfectly, unlike -let me know if I'm mistaken here- SPECWeb.

 

[JonUrban]

I wonder if Penteo's extraction isn't in some cases limited by the fact that it has to fold-down back to stereo perfectly, unlike -let me know if I'm mistaken here- SPECWeb.

Good point

 

[J. PUPSTER]

Not sure I understand this correctly; but it seems like you just have minimal gain control for individual channels with Penteo, rather than how the program discretely mixes various sounds/instruments (you just lose some percentage of the quality of the stereo down mix.) You could do this easily it seems to me (amplitude adjustment via your favorite DAW afterwards for any channel adjustments.) And are we also talking about potential clipping here?
Does that seem right?

 

[zeerround]

Re: perfect downmix for SpecWeb; No, SpecWeb doesn't have the constraint that it must mix back down to the exact stereo mix, however we don't add anything (with the exception of a non-default setting that lets you add more extracted ambiance back in).

You could set things so it does that, but since you have control of widths to place certain sounds in certain speakers if you mixed that down to stereo it wouldn't 100% match the original.

I guess, however, you could argue that when a producer does a 5.1 re-mix of a classic, there is no expectation that it mix down 100% to the original stereo from back in the day. The reason Penteo 16 and some other have that constraint is for surround broadcasting, and for those that listen to that in stereo, which is not what we are doing here. It's more like the 5.1 re-mixer, we are making something new, a 5.1 version of a song that we enjoy hearing in 5.1.

However, looking at the waveforms Jon posted, it doesn't seem that channel level adjustments are what he sees as missing in the Penteo 16 results. All channels of the Penteo 16 wave forms look similar, except for volume. So even if you adjusted the channel volumes, it wouldn't result in waveforms that look like the SpecWeb output that show different shaped waveforms across the different channels.

Re: why -m4 (undocumented development test of "Zone 2", a different way of distributing sounds to speakers after ArcTan determines the perceived angle of the sound in the Stereo) is not the default in 2.0b1 (which I noticed yesterday actually makes short cuts on your desktop that say 2.0a2 ) That's one of two reasons why this version is a beta.

I need user feed back on if -m4 is "better" than the default (-m1 "Across" "Constant Power Panning") and if so is it consistently "better", so that it should be the new default going forward.

FYI the other reason is two of the width settings in this version.

From the ini file:

;centerwidth is the amount of the sound field to put in the center speaker
;0 would be no center at all and 119 (the maximum value) would be
;a big chunk of the common sound between left and right
;default 54 (SpecWeb 1.5 default was: 75)
centerwidth=54
;
;frontwidth is the amount of the sound field to put in the front speakers
;0 would be no fronts at all and 120 (the maximum value) would be
;a big chunk of the sound field goes to the fronts
;default 36 (SpecWeb 1.5 default was: 90)
;Note that there is no "rearwidth" because the rears get whatever is left
;over after you specify the center and front widths.
;adjusting the widths is one way to reduce artifacts
frontwidth=36

The new values, 54 and 36, were arrived at after a statistical analysis of 38 songs I have gathered, form across musical types and producers, to be used both as "torture"/regression tests in any changes to SpecWeb, and to be used as objective listening examples for changes.

However, the few alpha testers I have left gave feedback that they preferred the original settings (which derived from DKA suggested starting points for the original Spec in Plogue Bidule).

So, again, I need feedback from users on what the default centerwidth and frontwidth values should be. Values that work for most/typical songs.

Then I would release 2.0 (or 2.1 if there are bugs/features updates) with those defaults.

 

[HomerJAU]

Penteo‘s ‘Downmix to Stereo’ compatibility is only there to support broadcasting where 5.1 may be downmixed back to stereo on stereo only systems. It’s not a requirement or consideration for those here doing upmixes solely for our surround systems.

 

[J. PUPSTER]

Penteo‘s ‘Downmix to Stereo’ compatibility is only there to support broadcasting where 5.1 may be downmixed back to stereo on stereo only systems. It’s not a requirement or consideration for those here doing upmixes solely for our surround systems.

I guess that's kind of what I was trying to say, as in, it won't matter to us that keep it in 5.1 or in my case I prefer 4.0. Also just wanted another opinion on the Penteo performing "a hard set or required" program down mix to stereo and only affecting volume or amplitude of the channels not the mix offered up. I know all this stuff can get confusing fast amongst us non-technical folk here (I'm including myself in that.) I believe that's what Glenn was alluding to when he said:

"However, looking at the waveforms Jon posted, it doesn't seem that channel level adjustments are what he sees as missing in the Penteo 16 results. All channels of the Penteo 16 wave forms look similar, except for volume. So even if you adjusted the channel volumes, it wouldn't result in waveforms that look like the SpecWeb output that show different shaped waveforms across the different channels."

 

[JonUrban]

If you look at the above wav files (Top Penteo/Bottom SPECWeb 2.0 -m4), I can't see where you would think that the only difference is volume! Look at the definition and more discrete placement of audio between these two files. The SPEC pulled the orchestra into the rears with decent vocal separation between two vocalists. You can see this in the last section of the above files. The Penteo puts that vocal into all of the channels while the SPEC -m4 localizes it a lot better.

I need to do this on a tune everyone knows so I can post a snippet or two so you can hear what I'm talking about. Of course, this difference may be source specific with regard to this particular CD. But I am telling you it sounds better to me, more "surround-ish". I listened to the entire album in the car today and in some of the tunes there were like a "wow" moments. I did not experience this with the default Penteo files (although I did not bother to make them for the car)

 

[J. PUPSTER]

After seeing the wave forms above (which I figured Penteo was on the top part), I also thought we were all saying the same thing about Specweb doing a better job of separating the instruments/vocals and Penteo (perhaps due to it's stereo fold down requirements) was limited it's ability for certain songs like this one in creating a more discrete output. If I am wrong in my wording trying to explain my thoughts on this, please forgive me, just trying to get a handle on all the terminology as well.

But hearing the files some, would definitely be what I would go off of; because I'm not an expert at reading wave forms.
But again, bottom line for me is, I thought we all had agreed on this; just wish there were more folks here chiming in with comments

 

[sukothai]

With all the talk about SpecWeb and the new -m4 option, I had to try it out. It sounds good, but I'm getting very strong rears with the conversion. Here's a pic:

I'm applying just the -m4 option. As a novice at this, I'm wondering if I'm doing something wrong?

 

[beerking]

With all the talk about SpecWeb and the new -m4 option, I had to try it out. It sounds good, but I'm getting very strong rears with the conversion. Here's a pic:

View attachment 44446

I'm applying just the -m4 option. As a novice at this, I'm wondering if I'm doing something wrong?

I've not experienced anything like that on the albums I've converted today using the -m4 option.

Which SpecWeb version are you using?

How many albums have you converted?

 

[sukothai]

I'm on the 2.0 beta. I don't thing the -m4 option is available on previous versions. I've done 5 cd disks, all from the same artist and they all look similar. The files are 44.1; so maybe need to be upsampled? Or is this done automatically?

 

[J. PUPSTER]

I'm on the 2.0 beta. I don't thing the -m4 option is available on previous versions. I've done 5 cd disks, all from the same artist and they all look similar. The files are 44.1; so maybe need to be upsampled? Or is this done automatically?

I'd recommend asking Glenn at the source thread here:
https://www.quadraphonicquad.com/forums/index.php?threads/specweb-now-2-0-beta-1.21473/page-26

 

[beerking]

I'm on the 2.0 beta. I don't thing the -m4 option is available on previous versions. I've done 5 cd disks, all from the same artist and they all look similar. The files are 44.1; so maybe need to be upsampled? Or is this done automatically?
[/QUOTE

Perhaps try a different artist/Album

Have you tried without the -m4 option?

I presume you are using either WAV or FLAC files, for the final conversion

Is your source from a ripped CD?

I'm not the one to really answer this....Glenn is your man.

 

[zeerround]

-m4 is only in the 2.0 version.

2x up/down sample is performed by default (but has to do only with quality and artifact reduction, not separation or channel mapping) 4x and none are other possibilities.

The width controls are there to help you move things around. You could start with the old defaults for center (75 vs. 54) and front widths (90 vs. 36) from previous versions sof SpecWeb:

-m4 -c75 -f90

to get things moved towards the front.

Or by editing the ini or a copy of the ini file in the directory of you source or named the same as your source (editing the SpecWeb.ini in the \install\bin directory has a global affect, SpecWeb.ini in your song's source directory effects any songs in that directory, and "your song file title".ini in the song's source directory effects only the one song).

Command line options overide any ini file options.

Back to the width thing; higher numbers of degrees will capture a wider area of the stereo mix so -c75 captures 75 degrees around center and sends it to the center speaker. -f90 makes the front speakers capture 90 degrees, and whatever is left over after -c and -f becomes the rear width, and that goes to the rears.

of course, adjusting these live, during playback, is easier than non-interactive "try a settings and see", so you might want to try the Play, or Play remote (if your PC is in the sweet spot of your surround setup).

Hmm, since -m4 isn't fully deployed yet, to get it to work during playback I guess the best way would be to make a copy of your SpecWeb Play or Play Remote icon and edit the properties so it says -m4 -P1 at the end instead of just -P1.​

​

​

​

 

[HomerJAU]

I think I need some help with SW2. I tried playing using the -m4 option but I'm still hearing artefacts, swishing from about the 22 sec mark in the test file here:

https://drive.google.com/drive/folders/10YzED3h4YAT5JP2pGkSQe66ToGEg0lSD?usp=sharing
There are 3 files:

• Original stereo source file (first 30 seconds)
• Penteo UM (30s)
• SpecWeb2 UM (30s) - default 2.0 settings with -m4 (I think, not sure where to put that in the ini file but Play with -m4 has swishing)

Glenn, can you suggest settings to remove the swishes, drums and guitars from 22 sec mark?

Both upmixes sound great until 22 sec mark, then SpecWeb starts swishing like crazy.

THX

 

[sukothai]

-m4 is only in the 2.0 version.

2x up/down sample is performed by default (but has to do only with quality and artifact reduction, not separation or channel mapping) 4x and none are other possibilities.

The width controls are there to help you move things around. You could start with the old defaults for center (75 vs. 54) and front widths (90 vs. 36) from previous versions sof SpecWeb:

-m4 -c75 -f90

to get things moved towards the front.

Thanks Z! The settings: -m4 -c75 -f90 did the trick. I also tried:

-m4 -0-110 -3-110 -c0 -f90

to get quad with great results. Thanks for the help!

 

[J. PUPSTER]

Thanks Z! The settings: -m4 -c75 -f90 did the trick. I also tried:

-m4 -0-110 -3-110 -c0 -f90

to get quad with great results. Thanks for the help!

@sukothai thanks for the Quad settings recommendation; since I intend on converting my music to Quad, could you please let me know if those settings continue to provide good results for your other songs?

 

[sukothai]

@sukothai thanks for the Quad settings recommendation; since I intend on converting my music to Quad, could you please let me know if those settings continue to provide good results for your other songs?

So far the quad settings have worked great. My main goal here was to convert my families favorite set of Christmas CDs by Sufjan Stevens. He has an eclectic style that ranges from folk to rock and all of the discs (10 in all) sound fantastic. I've got a quad system up where the tree is; so needed a good quad conversion to set the mood. I'll try some others and let you know, but it is certainly easy to try out.

 

[GOS]

So far the quad settings have worked great. My main goal here was to convert my families favorite set of Christmas CDs by Sufjan Stevens. He has an eclectic style that ranges from folk to rock and all of the discs (10 in all) sound fantastic. I've got a quad system up where the tree is; so needed a good quad conversion to set the mood. I'll try some others and let you know, but it is certainly easy to try out.

I’m a Sufjan Stevens fan!! He’s amazing!


Sent from my iPhone using Tapatalk