Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion. Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules. |
19th June 2004, 17:18 | #581 | Link | ||
Registered User
Join Date: May 2004
Location: Montreal
Posts: 729
|
Re: Re: Re: Re: Normalization
Quote:
Quote:
We all tend to think of stereo as two completely different channels of sound. However, if you change the two channels to M (for "Mid" or L+R) and S (for "Side" or L-R) then things can get quite interesting. Adding the two signals together without changing levels gives one composite stereo channel back, while inverting the phase of the S signal before combining the two will give the opposite composite stereo channel. If you start changing something about either of these before recombining, you get different results. "Stereo wide" effects basically increase the ratio of the S signal to the M before recombining (the better ones do frequency filtering because without it, this effect can kill bass response). To relate this to the bidules, when I wrote that the same signal is going to both fronts and the same to both rears, I meant that the same "difference" or S signal is going to each front and to each rear. If you then add an M signal to one of these, you can recover some sort of stereo channel. What gives the sense of difference is that the S signal is manipulated by the Stereo Touch plugin to produce two signals that are somewhat different from each other. If you then add the same sum signal to each, you get composite channels that are also somewhat different from each other. However, in the case of the vocal-center bidule for example, no composite channel is created. Instead the M signal is fed to the center channel. Thus, any recombining takes places acoustically through soundwave phase cancellation and addition. This definitely gives interesting surround effects, but it can vary depending on where the listener sits, and especially on the type of music. The research Andreas translated into his original LCR bidules is quite right in saying that these can provide a wider sweet spot for a stereo image, but if this image is transformed or modified, so is the sweet spot and the results. Since Stereo Touch does much of its work using delays, any music with percussive, rhythmic sounds could get messy. In fact, if you read the description of Stereo Touch on the Voxengo free VSTs page, (http://www.voxengo.com/freevst/), it says that the plug-in is "most effective on mono pad sounds and alike." I don't know if you're familiar with this term, but a "pad" is usually a synthesizer patch or a real recorded sound that is smooth and continuous in nature, without any sharp percussion or rhythm. So running a complete recording through Stereo Touch can sound bad (especially one with a rhythm section). Shayne, to get back to your normalizing question, given that the same difference signal is fed through the same effects plugins with the same settings, there should not be any problems at all. If you're in doubt, the best thing for you to do is to find a way to measure the "average RMS level" of each channel. Programs such as Sound Forge or Wavelab allow you to measure this. If the numbers for each channel are close, normalizing should not upset the balance. If, however, one channel has a sound somewhere that's louder than any others then it might end up sounding louder than the others. But, as I wrote, that's unlikely given the construction of these bidules. Regards, Steve. Last edited by ursamtl; 19th June 2004 at 19:10. |
||
20th June 2004, 03:56 | #582 | Link |
Registered User
Join Date: Oct 2003
Location: Dixie USA
Posts: 23
|
I finally tested out the new bidules (the voice-non center_or_instrumental one) and unless I am doing something wrong, it has way to much reverb for my taste. I felt like the music was "bouncing" all over the place. Sorry...
I still think my Ambisonics bidule (basicly the one from the guide with a little extra reverb added to the rear channels) sounds much better. I tried this on two seperate sound samples tonight. On a separate note, in a standard 6 speaker setup, are there any benefits going to the 3rd level in ambisonics? |
20th June 2004, 08:58 | #583 | Link | |
Banned
Join Date: Jun 2002
Location: The Netherlands
Posts: 326
|
Quote:
(The default setting we used in Voxengo was "pretty wide" , because that one gave good results on our testmaterial.) kind regards, Eye of Horus |
|
20th June 2004, 09:10 | #584 | Link | |
Banned
Join Date: Jun 2002
Location: The Netherlands
Posts: 326
|
Re: What's about the voices in new method?
Quote:
There is also another goodie we found and are using, but you have to contact me or kempfand by PM ! We cannot publish all our secrets here And a bit of a warning : I wouldn't suggest you to mess in the groups (as someone "advices"). For more gain use the gain settings. They are in the bidule with a purpose, not just as a gimmick ! kind reagrds, EoH |
|
20th June 2004, 17:03 | #585 | Link | |
Registered User
Join Date: May 2004
Location: Montreal
Posts: 729
|
Re: Re: What's about the voices in new method?
Quote:
Changing to the "Stage" setting didn't do anything for the percussive doubling problem I've mentioned. Both the voice-center and intrumental bidules still smear rhythmic or percussive sounds. For anyone who wants to hear this, do the following: [list=1][*]Take a song with some well-definied repetitive rhythmic sounds. For example, I just used Pink Floyd's "Comfortably Numb." If you don't have it, any song with well-defined drums (especially cymbals), strummed acoustic guitar, sharp rhythm guitar strums, slow harp plucks, etc.[*]Load in the Voice-Center bidule and load your song in the Audio File Player.[*]Replace the Audio File Recorder in bidule with Microsoft Sound Mapper output device. To do this Right-click the Audio File Player, and choose Replace > Audio Devices > Output > MME > Microsoft Sound Mapper.[*]Load your file into Audio File Player and listen carefully.[/list=1] What you hear is the sound that will be in the front speakers of your surrounds after you burn your surround CD with this bidule. Yes I know the two channels still aren't normalized as suggested in the method, but all normalizing will do is vary the volume level of the channel. It won't change the nature of the sounds in the channel. If there's a problem with the sound information contained in a channel, normalization will not erase it, but only change the volume level of the channel. Back to Comfortably Numb. Listen to the door knocks and the guy saying "Time to go...." at the beginning. In the original stereo recording, both are panned to the left as the artists originally intended. Through the bidule, however, both come from the center. Ok so that's not the end of the world, this is surround sound, right? Things may get moved. However, listen carefully to those door knocks. In the original recording they were so sharp and well defined, if you closed your eyes, you would almost jump up to go and answer your own door! Yet through the bidule, the realism is diminished. That's because the original audio waveform's transients have been mangled and delays echo the sound in both speakers. It still doesn't sound bad, but the sharp focus of the original recording is lost. Surround sound shouldn't do that. Listen to the cymbals in the verses and the strummed acoustic guitar in the chorus. They no longer sound real and well-definied. Surround sound shouldn't do that. The tight locked-in bass and kick drum notes are also smeared. It's also harder to hear these because most of their sound energy is contained in the sum of the original stereo signals that is sent to the center channel. This brings up another point I've mentioned in recent posts. If you've got a monster center speaker with great bass and treble response, you'll be ok. Most 5.1 systems have a center speaker that's optimized for dialog. Bass response in these is rolled off. Since there's sound going to your LFE in this bidule, the only deep, solid bass you'll get is if your center speaker can produce it. As for increasing the front center gain to make the vocals more balanced and less reverberated, to do that, you have to bring the center gain up to +10dB. This is if one has a monitoring system hooked up to listen to the adjustments. On most songs, this causes digital clipping and distortion. Yes, you're going to normalize the files afterwards, but once the distortion is introduced, it can't be removed. Look both EoH and Andreas, I've got nothing against either of you personally, but the more I explore these "exciting new bidules", the more disappointed I get. The logic of spreading the sounds out among the four speakers is good, but the execution mangles the original soundstage too much. If you can find a way to restore the soundstage and position of the instruments as the original artists intended, then these might be a better tool for upmixing. You'd also have to ensure the bass frequencies at least go to the front left and right speakers and/or the LFE. I know, I know, many experts don't believe in using the LFE channel, but others do. Any 5.1 receiver with bass management will direct the bass energy where it belongs anyway, so that point is moot. Finally, as for "warning" people not to mess with the groups, why not? If the result is better, it makes sense to change the groups. After all the "improvement groups" in these bidules take Gerzon's method for spreading a stereo L and R signal among LCR and apply it to only the stereo difference signal, and route the signals to front and back speakers. That's "messing" with an expert's original intent. If you can do it, why can't we do it with yours? Now if you choose to take this as a personal attack, I'm sorry but it is not intended that way. If, instead, you take it as it's intended--a critique of the results of your method as it has been presented so far--then let's discuss it. I've already said to Andreas that we should all be working together and contributing here as is the general intent of a forum. Instead of doing this mysterious "we can't reveal all our secrets" routine and telling people to private message you for more, why not share these secrets with the world in a spirit of cooperation? Regards, Steve. Last edited by ursamtl; 20th June 2004 at 17:07. |
|
20th June 2004, 23:18 | #586 | Link |
Registered User
Join Date: Oct 2001
Location: Switzerland
Posts: 225
|
Steve,
No problem for me if the 3 new bidules don't match your expectation. In fact, I see this positive, as it will allow for improvement. Some people are quite happy with them, some are disappointed. I imply that we all care for music and howto get the best out of it, and improving current methods is not easy or else ther wouldn't be a sometimes heated discussion on this or there would be a straight-forward solution. The "original artist's intend" is of a big theme by itself, and highly disputed even amongst the best know names in academia and industry. We probably could open a separate threat on this, which would become hugegly complex due to all the variables involved (performance, recording, distribution media, reproduction at end users room and equipemt used, and so one). Adding to that is the change over time (20 years back: most performances aimed for live delivers, today: many performances aimed at CD-distribution). IMHO a very interesting topic, and one I've not come to a clear opinion yet. Back to the bidules: As you correctly noted, we used Gerzon's method for spreading the signal not only to the front, but also to the rears. I would be very keen on your opinion for the 2->5 groups which I published (Gerzon 1997 Groups & Bidules ). More concrete on 2 setups: a) Using the "Improvement_5_x_2" Group, and feeding output pins L7/R7 to L/R, feeding L6/R6 to SL/SR, and C5 to C. Or b) Using groups "Preservation_3_x_2" -> "Preservation_4_x_3" -> "Preservation_5_x_4", and feeding as with (a) the five output pins to the ASIO audio device. Reason I'm asking is to get feedback on on this modified original intent (of Gerzon), and keep out other complications such as voice removal and mono_2_stereo alternatives. Regards, Andreas |
21st June 2004, 00:28 | #587 | Link | |
Registered User
Join Date: May 2004
Location: Montreal
Posts: 729
|
Quote:
I agree that the "original artist's intent" is a very complex subject. All I was getting at by mentioning that was that Pink Floyd's music by nature was quite detailed and meticulous, so if they mixed a door knock over to the left, it was probably intentional and not due to track limitations. By 1979, multitrack studio technology had advanced enough that it wasn't like old mid-60s recordings where the tracks were limited to 4 or 8 tracks. I'd be glad to give the groups you mentioned a closer look. Unfortunately, the ftp server is down right now and I can't find them on the needfulthings server. I thought they were included in the original download with the 3 bidules but I can't find the actual group files anywhere. I've checked the bidules and I don't see pins L7/R7, etc. so, I'll give them a look this week. I'm also working on another bidule that spreads out the front soundfield quite well but still preserves the imaging and seems complemented and reinforced by Ambisonic surrounds. I stumbled upon it by accident when playing around with some hookups. Once I finish testing and implementing some ideas, I'll write a guide and post it. Have a good week! Steve. |
|
21st June 2004, 11:41 | #588 | Link | |||||||||
Banned
Join Date: Jun 2002
Location: The Netherlands
Posts: 326
|
Re: Re: Re: What's about the voices in new method?
Quote:
Although I said not to reply anymore. I will, because there are again a few lines in your post I am not pleased with. And no, this has nothing to do with a personal attack...... How many times did we say you have to chose a song that suits the bidule ? If one song doesn't give you the result you want, you are immediately disappointed with the method ! You should criticize your choice of music ! What gives good results on one song, doesn't necessarily give good results on another ! Quote:
Did you ever listen to DSOTM in DTS (the Alan Parsons mix from quad) ? or the latest on SACD ? Did they intend this in their original stereo mix ??? Changing the nature of the sound is indeed not the intention but again : it also has to do with the choice of music. Some will do better on this bidule while other do better with SAD5.1inBidule or Ambisonics. Quote:
Quote:
yes, there is too much in the center, but NOTHING is steered to the LFE by the bidule ! The LFE is steered by your receiver only ! Quote:
In the bidule you need to find a good balance between L+R and C with the gains. And then normalize ! If you don't do that, yes, you can get some clipping ! So : increase the gain on the C, decrese the gain on the L+R and then normalize ! THIS WILL GET RID OF THE REVERB ! Quote:
Quote:
Not all people have those skills ! And answering between 10 and 20 emails every day about the (in your and our eyes) most simple things, is enough for me. Don't need more ! Quote:
I also don't want to sound negative only : try the bidule with for instance "Breakdown" from the Alan Parsons Project. A song with a singer and a huge choir with men and women. With the voice-center bidule the singer comes from the Center only and the choir falls in from all speakers (except the center) with even some separation between the male and female singers (front, rears). An overwhelming effect ! There is also almost no percussion in this song, so that problem is solved too Second : about the private message : we try to keep up with the spirit of this forum and don't publish links to or cracks for commercial software on the forum. But sometimes you find a program that does exactly what you need (for the voice-center bidule). If any people will give that a shot too, I won't publish it here in public ! But give the link in a PM ! (Is that so hard to understand ? Why that immediate accusation about "not sharing" ?) Oh, also a third point : "why not share these secrets with the world in a spirit of cooperation" > personally I am insulted by this ! Our work is always shared ! This whole thread and previous ones is sharing of our work. A ridiculous statement from you IMHO ! As long as you make this kind of remarks, I will go back into silent mode again and find other ways to publish our work . (as we already did, BTW) We shared everything and then someone comes up with such a remark ? What we don't share is in the spirit of this forum ! You see, it's still not about what you say, but how you say it ! Every invention or improvement came from positive criticism and suggestions to improve things. Negative ones lead to nowhere ! And to me you're still sounding too negative for my taste. So I wonder : if you cannot find positive things in what we produce : why the heck do you want to work with us ?? Quote:
Last edited by Eye of Horus; 21st June 2004 at 12:02. |
|||||||||
21st June 2004, 21:57 | #589 | Link |
Guest
Posts: n/a
|
from EOH (great) explanations:
"What gives good results on one song, doesn't necessarily give good results on another !" this is the resume. and i'm a witness of his words: ..."I am more interested in the people that are satisfied" cos he is posting "the road" to this results! thanks EOH and Andreas for this magical thread! (EOH my friend, take it easy! ) |
21st June 2004, 22:16 | #590 | Link | |
Registered User
Join Date: Oct 2001
Location: Switzerland
Posts: 225
|
Quote:
No need to hurry with feedback, [off-topic mode on] as things are pretty busy here with the European championship Adding only 2 more teams would make it a real world-championship: Argentina and Brazil [off-topic mode off] Andreas |
|
21st June 2004, 23:33 | #591 | Link | |
Banned
Join Date: Jun 2002
Location: The Netherlands
Posts: 326
|
Quote:
Thanks for the kind words !! Yes, I take it easy BTW I send you a PM ! EoH |
|
21st June 2004, 23:43 | #592 | Link | |
Registered User
Join Date: May 2004
Location: Montreal
Posts: 729
|
Re: Re: Re: Re: What's about the voices in new method?
Quote:
Ironically, I could accuse you of something similar when you stated that I chose one song and then immediately expressed disappointment. I never said or implied that I used just one song to test your bidules. I gave one example with Pink Floyd’s Comfortably Numb because it provides both percussive and non-percussive sound effects at the beginning plus both percussive and non-percussive music throughout. For testing surround bidules, I actually use about 20 musical recordings in various styles. I will make one correction to my previous message to you. I apologize because I forgot to type the word “no” in the following: “Since there's no sound going to your LFE in this bidule, the only deep, solid bass you'll get is if your center speaker can produce it.” However, you completely ignored the essential point that all deep bass energy is going to the center speaker instead of the left and rights. Since most center speakers roll off the lows, either someone is going to need a good subwoofer and bass management on the amp to redirect the sound, or else he or she is out of luck! Even with these, the separation of the bass frequencies in the front speakers will be compromised because the center speaker is by nature mono. Anyway, I could go on, but I don’t see the point--my other posts have adequately expressed my technical answers on these issues for anyone who reads them with attention. Before I joined this forum, I spent some time reading the threads and I’m sorry to say I developed a negative impression of you and especially how you handle criticism. This is one reason I avoided joining for awhile. I thought perhaps that would change once I got into discussing with you the surround technology covered here. However, every comment I made, you seemed to take as a personal attack of you. This has degenerated to the point where I have zero respect for anything you have to say, and I get the impression the feeling is likewise. Therefore, if I have comments or questions concerning your bidules, etc., I will address them to Andreas. If I see something you write about the bidules and I disagree with it, I’ll ask him about it. I will continue to contribute to this thread because it is still presented by Doom9 at a public Sticky thread on ways to upmix from stereo to surround and not just your personal thread. I hope Doom9 eventually reorganizes the threads so that everyone who creates an upmix guide in a separate thread is given equal treatment (surroundboy, kpex, bleo, et al) because it seems unfair to newbies coming along that these other methods—all every bit as valid in their own right as yours—scroll off the front pages of the forum. These people are also posting 'roads' to results and contributing to the community. For now, however, it seems that this thread has evolved into a general forum for discussion of upmixing to surround. Although you may have started it--and I do congratulate you for doing so--it is now a forum for discussion of various methods by various people. As such, it is a valuable resource and everyone who contributes to it deserves praise, just as the people I mentioned above and others deserve praise for taking the time to contribute their ideas. Last edited by ursamtl; 22nd June 2004 at 00:38. |
|
22nd June 2004, 00:06 | #593 | Link | |
Registered User
Join Date: May 2004
Location: Montreal
Posts: 729
|
Quote:
So, although I have some testing to do with the others and exploration of the guts, I'd say this is a winner. I ran about a dozen songs through it quickly on the PC ASIO out. I've been doing this so much now that I'm able to anticipate how things will sound through the big 5.1 system and I have a good feeling about this. Regards, Steve. |
|
22nd June 2004, 08:08 | #594 | Link |
Banned
Join Date: Jun 2002
Location: The Netherlands
Posts: 326
|
Re: Re: Re: Re: Re: What's about the voices in new method?
[QUOTE]Originally posted by ursamtl
I don’t see the point in quoting your messages back in my response because you’ll only accuse me of twisting your words. {/quote] Of course, why should you react to what I say ?? No, you don't, you only post another tirade about how bad I am. Here in Holland we call someone like you a guy with an incredible plate before his head ! This was a nice forum...... EoH |
22nd June 2004, 08:20 | #595 | Link |
Doom9 Member
Join Date: Sep 2003
Location: Australia
Posts: 210
|
@ ursamtl,
I can't find Preservation_5_Speaker_v2_REC_.bidule on the needfull things server, can you give me link for it. Personally, I follow this thread every day and to be honest there is a lot of improvement since the original CoolEdit-Aurora method. However there is some "fuss" in the air that I don't like and turns this thread in "unneeded" discussion. Keep up the good work ursamtl,andreas.... DSP8000 |
22nd June 2004, 11:27 | #596 | Link | ||
Registered User
Join Date: Oct 2001
Location: Switzerland
Posts: 225
|
@ Steve: Thanks for your feedback (BTW having the same experience here that running things through the ASIO gives a good anticipation on how it will come out on the big 5.1 system). I also was impressed when I first listened to the Groups and bidules using the Gerzon 1997 approach, and my mental note was a simple statement saying: "Prooven Gerzon Technology". Changing/adding a gain element for the Rears (as you did), or even a gentle reverb (using one instance of SIR) gives so much space for creativity and personal preference.
Quote:
When you use them, just make sure you feed the output-pins the right way (the one in the center goes to the C-speaker, the ones at the outmost left & right will go to the Surround, and the ones in the 'middle' go to the front L & R). I've added bidules in the RAR for illustrative purpose. Quote:
But I also would like to point out that EoH started it all with the original CoolEdit threat here (the Ambisonics method), and that his initial threat was my personal starting point for what I call a wonderful hobby that I wouldn't want to miss. Regards, Andreas Last edited by kempfand; 22nd June 2004 at 11:29. |
||
22nd June 2004, 12:48 | #597 | Link |
Doom9 Member
Join Date: Sep 2003
Location: Australia
Posts: 210
|
Thanks Andreas, found it.
Exactly, easy of use, speed... The original idea it is probably the same(you guys should know), but there is always room for improvement. With instr.music maybe it's better to experiment, as voice based tracks are a bit different to handle. All I'm aiming to achieve is complete separation of diff. channels,frequencies,to preserve the original recording sound across 5.1 encoded file without using too much reverberation. My Creative ZS Pro does nice job with the DTS NEO update driver but, for my taste the rears sounded "plastic" Thanks again & enjoy the sounds DSP8000 |
22nd June 2004, 13:14 | #598 | Link | |
Registered User
Join Date: May 2004
Location: Montreal
Posts: 729
|
Quote:
Another important point that's been made here several times is to convert the input wave files to 32-bit floating point format before the input. The difference may be subtle, but details on high frequencies, the overall sense of transparency in the music, etc., is improved. If you don't have a commercial program to do this, I've found an excellent freeware DOS utility that converts a file quickly. It's available from the Telecommunications & Signal Processing Laboratory at McGill University here in Montreal and is called CopyAudio. You can download it at ftp://ftp.tsp.ece.mcgill.ca/TSP/AFsp/ and read more about the whole package at http://www.tsp.ece.mcgill.ca/MMSP/Do...Fsp/AFsp.html. The actualy CopyAudio.exe is in the MSVC\bin folder of the extracted files. Simply add -D "float32" plus the input and output filenames to the command line and it converts your file very rapidly. It doesn't provide any onscreen feedback while working. Just wait until the command prompt returns and you'll find your finished file is converted. There are some other interesting utilities in the AFsp package for comparing audio files, resampling, processing multichannel files, etc. Too bad these aren't VST plugins! By the way, I noticed you're using a ZS Pro. How do you like it? I'm planning to upgrade my old Live 5.1 later this year and the ZS Pro looks like a good target. Have you tried the kX Project drivers on it? Apparently, they work fine on the ZS Pro and offer improved performance when compared to the Creative drivers. You can check the forums at www.kxproject.com for more. One other question, are you able in Plogue to add a multichannel ASIO out using the Creative drivers? Since Creative never supported ASIO in their official Live drivers, I've completely ditched anything from Creative and use kX exclusively. (Sorry to everyone else for bringing this questions into this thread, but since DSP doesn't accept PMs I have no choice Regards, Steve. Last edited by ursamtl; 22nd June 2004 at 14:44. |
|
22nd June 2004, 13:23 | #599 | Link | |
Registered User
Join Date: May 2004
Location: Montreal
Posts: 729
|
Quote:
One thing I did like about the 3 previous bidules you posted was that you added dithering to the outputs. I never felt we resolved this issue when we discussed it before. I know Surcode (for example) accepts 32-bit input files, but I was never able to find out whether it 1. dithered to 16-bit output, 2. truncated to 16-bit output, or 3. left the output at 32 bits. I seem to recall reading something that led me to believe that the digital DTS wave file is seen by a DVD player as a 16-bit file. I know from experience that the input file has to be at 44.1K for the DVD player to recognize the DTS wave file as a music track, so I assume it recognizes it as 16 bit as well. Therefore, I think dithering the outputs makes sense as one is sure that proper dithering will be done instead of hoping it's done by Surcode or whatever AC3 encoder one uses. What are your thoughts on this? Regards, Steve. Last edited by ursamtl; 22nd June 2004 at 14:08. |
|
22nd June 2004, 14:11 | #600 | Link |
Doom9 Member
Join Date: Sep 2003
Location: Australia
Posts: 210
|
@ursamtl,
From the kxProject FAQ "Audigy2 ZS series aren't fully supported at the moment, due to the fact that we do not have an Audigy2 ZS card for testing. However, as soon as we have enough money to buy one, support will be implemented." I'm happy with the Creative drivers and my PC setup so there's no need to change anything, maybe if they have enough good features in their drivers i'll give it a go. About CopyAudio, I'll definetely have a look(on the weekend) at the moment I'm using Audition. Yes, Creative ZS PRO is my choice, I used to have M-Audio Audiophile 24/96 and that is another very good sound card, but not suitable for my needs. Actualy this is one of the very first semi-pro multimedia/multichannel based cards that performs a bit better at 44.1 kHz. Native 48 kHz is excellent!!! All in one it is very good package. I do accept PM's just my bigpond email is bugged so I have to use this one ivancosax@hotmail.com ,geez you can tell that I play sax from the email. Regards, Ivan |
Thread Tools | Search this Thread |
Display Modes | |
|
|