#1002929 - 05/24/05 06:51 PM
Music DVD questions
|
Graham B
Member
Registered: 05/24/05
Posts: 4
Loc: East Anglia, England
|
Offline
|
|
I am mixing a live concert video DVD for a band and I have several questions. I hope you guys may be able to help me, despite not all of them being surround related!
My first problem is where to head with the overall level. After years of mixing and mastering for CD, I am used to peaking close to 0dB Fs. I would say that my 0VU (PPM4) level for the mix as it stands, is at about -8dB Fs. After investigating levels on commercial DVDs, I find that they often are closer to broadcast levels, with peaks to around -12dB. Some music DVDs however, have really hot CD kind of levels. The Led Zeppelin live DVD for instance peaks to -.5dB on both the PCM stereo and the DSD tracks, but is about 8dB lower on the AC3 track..
I believe there is some reason for keeping levels relatively low before AC3 encoding, is this simply for compatibility with feature films? Do you think it would be rather irresponsible to make my stereo mix that hot?
Another issue is should my stereo mix be data compressed? I would love to simply provide a 16 bit 48k file for authoring, but the programme is around 100mins and it seems very greedy to demand over 1Gb out of 4.7Gb (can’t afford dual layer), just for the stereo mix! How about MPEG encoding? Or AC3 stereo? Then I’m back to that levels question. Aaagh! 
I am intending to make a 4.0 surround mix. No centre, no LFE.
Any idea how big will the AC3 encoded file will be?
Sorry my first post here was a bit long and little to do with surround!
Thanks in advance for your help.
_________________________
Graham B
|
|
Top
|
|
|
|
|
#1002930 - 05/25/05 03:38 AM
Re: Music DVD questions
|
alfonso
Platinum Member
Registered: 11/06/00
Posts: 1080
Loc: Fregene, Italy.
|
|
Offline
|
|
I'm not a surround expert, I was just reading to learn, but I think that when you go out digital to consumer level converters you have to be very prudent with levels because you don't know what accuracy can be found and a too hot signal could make a mess, and most commercial converters are really crappy.
|
|
Top
|
|
|
|
|
#1002931 - 05/25/05 06:26 AM
Re: Music DVD questions
|
Neil Wilkes
Gold Member
Registered: 07/06/02
Posts: 514
Loc: London, UK
|
|
Offline
|
|
A 2/2 Dolby Digital file should be encoded with a bit rate of either 0.384 Megabits/second, or for maximum quality without running into compatibility issues, 0.448 therefore, at 100 minutes, your file size will be 0.384 * 60 * 100 = 2304 Megabytes or thereabouts. At the higher bitrate it will be around 2688, so personally I would use the higher rate. Unless you are happy with using the downmixing, then I would also include a DD stereo stream too, which is at 0.192 Mb/sec, giving you an extra 1152 Mb in your project.
Not sure what you mean about DSD tracks on the Led Zep DVD - there ain't any. DSD is SACD, not DVD. I think you mean DTS, which is the same bit rate & file sizes (approximately) as LPCM stereo is.
Volume levels.
This one has been done to death over the years.
DO NOT overcompress
DO NOT hit 0dB peak, try to peak at no more than -3 when mixing.
Use the K-system, and set up the system correctly using filtered pink noise.
see http://www.digido.com for details, under the articles section called "Level Practises".
DO NOT use tools such as L2, or maximizers to raise the average levels. Believe me, 5 channels (or 4 in your case) at 85dB SPL is plenty loud enough. Using brickwall limiting to bump up the levels will destroy the DTS and Dolby Digital encoding processes.
As for actual content, in NTSC you must have either Dolby Digital or LPCM. DTS is optional.
As this is a music video, I personally would rather drop the video bitrate as opposed to compromising on the Audio. What is more important to you - the pictures or the music?
|
|
Top
|
|
|
|
|
#1002932 - 05/25/05 08:45 AM
Re: Music DVD questions
|
mudsmith@earthlink.net
Gold Member
Registered: 10/30/00
Posts: 750
Loc: Martinsburg,WV,UNITED STATES
|
|
Offline
|
|
....Ditto most of what Neil says....If you think about spreading your stereo mix over 6 channels, you will realize two things: 1)The front levesls are bound to be quite a bit higher than the surrounds, center and LFE, 2)By spreading the material over 6 channels (well, 5 depending on how you do this), the peaks of the individual channels will tend to be reduced, but not necessarily always because of the changes in panning within the mix and/or pan law within your software or mixer.
If you have mixed your stereo mix without overall stereo compression with max peaks that hit right at or just below digital 0, when you convert this stereo mix to surround you can do the same thing for the front channels on the surround mix- i.e., raise or lower all channels to achieve this without adding any overall compression.
I think it is then okay to add some light compression if you have a good, true surround compression algorithm or hardware box, but this is a hell of a lot trickier than it is for stereo. Very, very tricky, and I have heard it botched badly by some of the big mastering houses. I would heartily agree that slamming a surround mix with heavy limiting is grotesque and should be avoided at all costs.....In other words, no compression at all is much better than going down any of these other deadly paths.
There are DVD manufacturers and distributors that routinely include a straight PCM stereo track that plays when you choose stereo on the main audio menu, but this is often defeated when you are using the digital out of your player, which will default to the Dolby Digital downmix. Unfortunately, the Dolby Digital downmix is not necessarily a great way to derive stereo.....The analog outs usually seem to give you the PCM stereo mix in this situation, however....This PCM mix has been included on everything I have had come out on Image entertainment, including some pretty lengthy titles (over 100 minutes for at least one, plus some interviews), with no obvious need to degrade the video. All these DVDs have also had both Dolby Digital and DTS surround mixes.
A lot of this kind of data usage depends heavily on the strength of the authoring tools and the authoring operators. The folks at Image seem to be pretty good at juggling the bit rates.
I would agree that compressing the video more heavily is a way out of the data limits. When you start getting into a lot of extra features (alternate views and mixes), this becomes absolutely necessary. The newer tools and double-sided and double-layered discs give you a lot more latitude than just a few years ago, however.
|
|
Top
|
|
|
|
|
#1002933 - 05/25/05 10:29 PM
Re: Music DVD questions
|
Graham B
Member
Registered: 05/24/05
Posts: 4
Loc: East Anglia, England
|
|
Offline
|
|
Guys;
Thanks a lot for your comments. Most helpful.. Got me thinking... You guys have pursuaded me to use no mix compression. I'm only chopping 4 dB max with an L2 in any case. I'm enjoying the mix dynamics without limiting, it follows the on the night mix dynamic better (I also mixed FOH). Hopefully the punters will turn it up if they want it louder! And it will better match a feature film.
I'm also wondering if I can get away without a dedicated stereo mix.. In a few days i'm going to set up my pair of Genelec 1031s as surrounds with my friend's 1031s as L/R at his studio (and calibrate the levels). I'll experiment then with how well a downmix would work. I do like the idea of dispensing with menu selection of stereo, and this digital out confusion that mudsmith spoke of.. And the reduced amount of data; the band cant afford double layer discs..
In the stereo mix I'm working on at the moment, I have balanced my room mic (a Calrec Soundfield near the back of the room) quite high in the mix. I have pre-delayed it to align the PA sound with the multitrack sources, and it sounds great! A lot like being there in fact.
For the surround mix, I was thinking of simply panning my (seperate) stereo audience mic to the surrounds. This mic was actually on a lighting rail, pointing down and back. I'll probably use an extra stereo reverb sent to the back pair too. I think that ought to sound good and fold down well. Not using a centre may help too. I hope downmixing won't be a compromise at all, and of course the encoder will let me choose the balance.
Looking forward to trying it! I'll let you know how it sounded if you're interested..
I intend to set Dialog Normalization to -31dBFS. Good plan you think?
_________________________
Graham B
|
|
Top
|
|
|
|
|
#1002934 - 05/26/05 09:16 AM
Re: Music DVD questions
|
mudsmith@earthlink.net
Gold Member
Registered: 10/30/00
Posts: 750
Loc: Martinsburg,WV,UNITED STATES
|
|
Offline
|
|
Although it galls me, you have to realize that the majority of DVD listener/watchers will not hear your surround mix. They will only hear it in stereo. Additionally, the fold down of a Dolby Digital mix, to my ears, sounds greatly more degraded than a non-encoded fold down. Everything Dolby does to use lower data rates for surround ends up hurting the recombination.....I have also found that I tend to add more reverb overall to the surround mix than I did to the stereo mix which served as the basis for it. The clarity of the extra speakers seems to shout out for it, in my opinion. This ends up creating a more mushy downmix.
In short, I think a dedicated stereo mix is important. I'll bet you can find the bandwidth.
Creating a dedicated stereo mix is not necessarily a huge extra time function. I tend to do the stereo automation first, which settles more of the real mix issues for me. The automation pass with surround ends up being somewhat quicker after that. The reverse should also work more or less the same way....One pass requires the big, creative problem solving decisions. The next requires tweaking and redirection to the other format.
It certainly requires some time, but surround demands a huge increase in housekeeping and thought time on its own.
|
|
Top
|
|
|
|
|
#1002935 - 05/26/05 02:36 PM
Re: Music DVD questions
|
Neil Wilkes
Gold Member
Registered: 07/06/02
Posts: 514
Loc: London, UK
|
|
Offline
|
|
I am in full agreement with mudsmith about using a dedicated stereo mix. Trouble with the DD downmixes is you are seriously limited in your options. For the rears, you can drop the levels of what gets recombined by 0.707 or 0.5 of full scale, or else leave it at 0dB, all of which are equally grotty. A dedicated stereo version takes up so little space it is definitely worth losing video bandwidth in favour of the extra audio stream. The loss incurred to make the space is truly insignificant. If you use a decent MPEG encoder, then you will find that for an investment in time you can get some really good results at surprisingly low bitrates.
Again, if you are planning to make DVD-R as opposed to having the discs stamped, then you don't really want to go any higher than 6.5Mb/Sec, 7Mb/sec tops on the video anyway, and using multipass VBR at a lower rate of say 5Mb/sec will still produce a perfectly good result. Plus it leaves you plenty of room for the all important audio.
You also say "only" 4dB. This is a lot of compression. Remember that 6dB will halve the level. So, by reducing peaks by 4dB, you are cutting off 2/3 of your transient levels. It's gonna sound compressed.
Again, as mudsmith says, if using a compressor across all 5 channels (we'll leave the LFE out of the picture for this discussion) you definitely want to be using a dedicated multichannel version or the phasing problems are the stuff of nightmares.
And also, although it is (sadly) common practise to put ambients and a bit of reverb to the rears, it is not a good idea. It makes for a boring mix, and the downmix will be swamped with reverb at this point. Better by far to do a proper, discrete mix.
Good luck, and please - keep posting! And keep us informed on how it is going.
|
|
Top
|
|
|
|
|
#1002936 - 06/14/05 09:39 PM
Re: Music DVD questions
|
Graham B
Member
Registered: 05/24/05
Posts: 4
Loc: East Anglia, England
|
|
Offline
|
|
Neil, Mudsmith,
Thank you so much again for your considered replies. The DVD in question is finished, authored and we are waiting for the first run to arrive from the Czech Republic! Thought you may be interested to hear how the mix went…
We discussed several issues here. Maybe I should have put them in separate threads.
Stereo mix
The stereo mix was already close to finished when I first posted. I am very aware ‘that the majority of DVD listener/watchers will not hear the surround mix’, and I don’t know anyone who has a surround system that approaches Hi-Fi! When Mudsmith said ‘the fold down of a Dolby Digital mix, to my ears, sounds greatly more degraded than a non-encoded fold down’, that was enough to persuade me to keep it separate. However, I have to say the fold down sounds pretty acceptable.
Surround mix
Neil said ‘sending ambients and a bit of reverb to the rears is not a good idea. It makes for a boring mix’. What else would you suggest Neil? This is a live show. I wanted it to feel like being in the audience so I felt it didn’t make sense when I pulled stage mics away from the front. I tried panning exactly as I mentioned in my last post, and it worked perfectly to my ears. I did add another separate stereo reverb for the rear and some ping pong delays across the back on the slide guitar.
As I expected, there were comments from the band like ‘shouldn’t the rear speakers be behind us’ and ‘I don’t hear much from behind me’. They did believe that I knew 110 degrees was correct. And could easily hear the contribution of the rears when I turned them off!
The surround mix for me has a wonderful sense of envelopment and I really can’t imagine that it would be improved if I’d used the centre.
Levels
The levels on our first test DVD seemed very low compared to other music DVDs. After some investigation, I found that the AC3 encoder we used (A-pack in Media100) had set dialogue normalisation in the files to -27dB, despite being set to -31dB during encoding! We found that saving and opening it’s settings put that right. I also decided to leave in the L2 on the mix bus for the stereo version.
Neil, I think we will have to agree to disagree about 4dB of L2 limiting being a lot. It’s virtually inaudible, just 4dB louder. It only hits reduction for a few moments in a 100min concert. Mastered CDs are way more limited (or often clipped) than that (not that I like this any more than you!).The apparent loudness of my mix is still well down from the Zep DVD stereo mix, which sounds like it’s had CD mastering treatment.
Whatever, as you suggested I used no limiting on the surround mix.
Encoding
To keep the data down, I decided to AC3 encode the 2.0 mix as well as the 4.0. Neil; thankfully your bitrate calculations were off by a factor of 8! You forgot to change from Megabits to Megabytes. I encoded the stereo at 224 and the surround at 448. Sounded great and just enough space left on the single layer DVD for 100min of video at decent rate.
BTW, I didn't realise that you can only put about 4.3Gb on a 4.7 DVD!
Thanks again for your help guys. All in all we are very pleased with this DVD. Hoping that the production run is delivered in time to sell to the dedicated fans at our gig this Saturday (18th June) at the Shepherds Bush Empire.
And Neil; I believe that your studio is nearby, if you fancy coming to the gig, I’ll put you on the guest list and I can give you a DVD to check out for yourself. Failing that, PM me with your studio address and I’ll put one through the door..
_________________________
Graham B
|
|
Top
|
|
|
|
|
#1002937 - 06/16/05 06:36 AM
Re: Music DVD questions
|
Neil Wilkes
Gold Member
Registered: 07/06/02
Posts: 514
Loc: London, UK
|
|
Offline
|
|
Graham.
I'd love to come along & check out the band.
My email is
neilwilkes@opusproductions.com if you want to let me know a time.
The reason the data on a DVD is 4.3 and not 4.7 is that DVD uses factors of 1,000 and not 1024 kb per Gb. It takes a little getting used to, but once you remember this it's all okay.
Encoding.
Whoops! Sorry about that. You are perfetly correct, I was thinking in Megabits, not megabytes.
Surround Mix.
It's a personal thing, I guess.
There are a couple of schools of though, and I usually use the centre as an anchor, and feed significant info to the rears, to give the listener the feeling they are onstage with the band.
It depends.
Using ambients & 'verb in the rears puts you about 20 rows back.
We can discuss comparative methods at the Empire!
|
|
Top
|
|
|
|
|
#1002938 - 06/17/05 10:37 AM
Re: Music DVD questions
|
mudsmith@earthlink.net
Gold Member
Registered: 10/30/00
Posts: 750
Loc: Martinsburg,WV,UNITED STATES
|
|
Offline
|
|
Great that things worked out so well for you.
Neil is right that there are lots of different philosophies in play for use of the 5.1 environment...including lots of folks who lobby logically for actually making it a 4.0 environment (no center channel, no LFE with expected bass management added later). I am not one of these, but understand the argument.
My own take on the ambients/rear discussion is that having ONLY ambience and reverb in the rears and ONLY music in the fronts gives a flat, unconvincing surround mix....kind of like those cheap surround receiver effects.
Creating a music stage with a little depth that comes into the room a bit, however, gives things a more 3D feel, and having the reverb and audience wrap around you to the sides tends to do the same, in my opinion.
There is an argument made against bringing the music elements fully panned to the rear speakers that says that sound localization perception is pretty poor to the rear(this is one of the places that thoughtful use of the center channel can help in image stabilization)....people who want to recreate a live experience also tend to find this hard rear panning unnatural. I think, however, that this can work well when the concept is carefully realized, and can give you the "sitting in the middle of the band" thing that Neil is talking about. It can also give you a totally new way of hearing the music.....
It's a wide-open format, but I just lean towards "sitting in the front row" with a little super-reality 3D overlay myself. ...Any creative use of the format needs to be thought through and listened to carefully. If it works, it works.
A big consideration for me is how it works with the picture. There is very little surround material out there that is not meant for listening with picture.
|
|
Top
|
|
|
|
|
#1002939 - 06/17/05 07:39 PM
Re: Music DVD questions
|
Graham B
Member
Registered: 05/24/05
Posts: 4
Loc: East Anglia, England
|
|
Offline
|
|
I hope you didn't get the impression that I only have reverb in the rears, I said:
In the stereo mix I'm working on at the moment, I have balanced my room mic (a Calrec Soundfield near the back of the room) quite high in the mix. I have pre-delayed it to align the PA sound with the multitrack sources, and it sounds great! A lot like being there in fact.
For the surround mix, I was thinking of simply panning my (seperate) stereo audience mic to the surrounds. This mic was actually on a lighting rail, pointing down and back.
Don't forget that both of these stereo sources contain masses of the PA mix as well as audience.
_________________________
Graham B
|
|
Top
|
|
|
|
|
#1002940 - 06/18/05 03:15 PM
Re: Music DVD questions
|
mudsmith@earthlink.net
Gold Member
Registered: 10/30/00
Posts: 750
Loc: Martinsburg,WV,UNITED STATES
|
|
Offline
|
|
That makes sense.
I have generally been creating a music stage with depth that takes up the first fourth of the room....I will then have 2 to 3 stereo pairs of audience mics (6 total mics) when possible, all facing away from the stage and staggered back in the room when I can. These tend to get mixed in progressively further back in the room, starting at about halfway, with panning across the room to taste for each pair..
Perhaps more importantly, I try to place all the reverbs that relate to the original stereo mix as effects in some sensible position of depth on the music stage, but then send them to another dedicated reverb that lives in the back of the room....this tends to give everything a motion and relationship with the whole room.
There are often some other reverb and delays I add on, but they are all based on this concept....Using the center channel to anchor things that should sensibly be in the center and up front seems to undo most back-to-front imaging problems....we'll see what else changes as time goes on...
Moving audience mics forward in time can really make the audience coherent and undo a lot of the hollow ambience effect, I know. My colleague has done that a bit, but I have not yet taken the time. On a bunch of projects, I was mixing from tape and could not have done it as easily as it can be pulled off in a DAW...
|
|
Top
|
|
|
|
Moderator: BobbyO, Neil Wilkes
|
|
|
