POST-WRITING this warning (LOL) : This may be a long read
I see what you are saying about the monster's "growl" and the fact that it does kinda seem like the same sound effect was used in a few places and how there is an issue with whether it was added later or not and it's a tough call. There is also a scene where as Frankenstein raises the monster up to the opening in the roof, it cuts to him and we only hear the crackling of the electricity; but, you can clearly see his lips moving and this happens in two shots (BD timecodes: 23min 59sec-24min 01.5sec right before shot of Fritz and 24min 07-08sec [this one is a shot of Victor and Elizabeth, where you can see her lips moving and on the French dub this is dialog here spoken by her] 24min 12.6sec-24min 14.7sec) and his lines are dubbed in the German and French versions, additionally the shot of where Victor and Elizabeth are onscreen and you can see her lips moving, the French dub has spoken dialog here too. Interestingly, subtitles for these parts do not exist for either language on any source, so I can't figure out what they are saying. Would be cool if we had any French and/or German speakers that could at least translate the lines (it's only like three or four lines) because maybe I could re-construct that too from dialog spoken elsewhere in the film by those voices (assuming that the dialog is not unique to those parts)
Yes, there will be multiple versions:
- "raw" LDPCM synced to BD
- "filtered and cleaned" LDPCM synced to BD
- "raw" (it's processed due to matching the audio tracks, but NO noise reduction yet - in fact, the noise plain is a little higher than on the "raw" LDPCM) hybrid-cut reconstruction track (multi-sourced)
- "filtered and cleaned" hybrid-cut reconstruction track
Ok, so I have been exhausted after the move, while still having to take care of a lot of stuff and I guess I ran myself into the ground because yesterday I woke up with a fever and stuffy nose and feel even worse today; my back is f***ing killing me now too
However, the project still moved forward, albeit just a little bit since I had to restart the reconstruction process from scratch...
I usually use Audacity when doing sync jobs, because I like the level of precision that the program has when it comes to being able to move stuff at increments of 1ms (yup, 1/40th of a frame at 23.976fps, LOL) and this makes stuff pretty easy. The times that I need a video reference, have been very few and in those cases I would throw the project into Sony Vegas. The details of the waveforms the way they are displayed in Vegas is sh*t and does not show enough detail (foley and background sfx are usually a flat line) and it's not a matter of incorrect settings on my part (been using Sony Vegas for the last decade) This makes it nearly impossible to insert snippets as short as 25ms (to fill in audio drop-out that I couldn't correct before, due to the nature of the English track) because I can't tell where they are exactly and the precision of moving the audio around to sync properly with what's on-screen becomes a freakin nightmare. Can it still be done? Sure. Am I doing it this way? HELL NO!
There's a paid audio editing app (although they offer a 60-day free trial and the app is only like seven bucks) called Reaper. I've had it on my computer for a while, but have only used it on a couple of occasions. It's kinda like Audacity on steroids (very much expanded version of it, with TONS of precision of every aspect involved in audio editing; but involves a much higher learning curve and takes some getting used to) AND an option to load ANY VIDEO track onto the timeline to have as video reference (done via a VLC plugin - which isn't mentioned anywhere, but I noticed it when I looked at the properties of a video dropped onto the timeline) which on this project is an ABSOLUTE necessity.
Ok, so the audio tracks that I have (although this doesn't mean I'm using all for reconstruction, but the more the better for reference - more on that in a moment) right now are as follows:
- the 1993 JPN LDPCM, of course
- 1986/87/91 VHS tape (English)
- 2004 R1 DVD (English) and the R5 (Russian) counter-part
- 2012 BD (English, French, German, Italian, Spanish)
- 20?? Polish TV broadcast (Polish voice-over, sounds like the 1991 VHS tape/LD audio)
- R2 DVD -unsure of year- (Italian, different master than BD, also censored version)
- R2 DVD -unsure of year- (Spanish, different disc than Italian, different master than BD, also censored version)
- 2007 German TV broadcast (German, different than BD, also censored version)
- the "OST" CD audio ripped "raw" directly into WAV
and the following TWO sources (originally was just one) that I'm now allowed to use for the small portions that may be and am allowed to mention now (came from private collectors and I am NOT ALLOWED to USE and/or SHARE THESE in their entirety, so PLEASE do NOT ask - thank you kindly)
- 8mm audio capture (German, this is the censored version, prior to restoration and it's missing a few bits, different master than the other German sources I have, very low fidelity and is more of a reference than anything else)
- 16mm audio capture (English, also the censored version prior to restoration like the German print. Different master than the other English sources I have, good chance this is the audio that was on the 1980 VHS tape, considering the fact that it's missing a few bits. However, it does contain a few sound effects that are missing on other English transfers [although just a few, but still it helps me verify more and more as I go] but are present on some of the foreign dubs. The quality is not the best and will only be used if can't be sourced from the Spanish dub [looks like three instances where there is dialog on the Spanish track and this will be easier to splice in then trying to clean the Spanish dub] which is the best quality of the "non-LD" sources, LOL)
I have been able to do a rough match (sync-wise) of all the tracks to each other and compared them all, so I know what needs to be used where. I have over 60 pages (double-sided) of handwritten notes that I'll be using to do the re-construction of the audio track. There is a TON of stuff missing on the English track. It becomes very obvious when comparing the various audio sources waveforms, seeing only one second at a time on-screen. It literally looks like the English audio track was put together from various sources and audio stems (some of them sound like very rough snippets and look obviously spliced in too late at beginning or too early at the end of an audio stem) that were found at random and I can tell which ones were inserted at a later time as they do NOT match the sound of what's on screen and are usually out-of-sync [i.e. footsteps on nothing but dirt, sounding like footsteps on wooden floor and vice-versa] I have had to spend a lot of time syncing various sfx to what's on-screen and then trying to accurately judge which ones are the correct ones. As far as I'm concerned, the two factors in my decision every time are whether it accurately represents what's onscreen AND if it syncs properly without me having to resync the sound effect once I found it's "starting point." It's how I have been able to determine that some of the sound effects are correct and incorrect on ALL audio transfers and I have been able to identify what I think are all the right ones between all the transfers [and it appears that they are actually nearly all there] and this really is a case of Frankenstein'ing the audio for Frankenstein, which I think is poetic justice
While working on matching the audio tracks to each other, I realized something I have to do when splicing them together... So, I have been able to match the fidelity and clarity, "warmth"/"depth" of all the sources being used pretty damn close to each other making for pretty seamless transitions EXCEPT for ONE HUGE thing: the background noise plain is very different between the transfers I'm using... SH*T! So I thought about it for a moment and decided that since I have everything mapped out, I can figure out where to insert noise.... "WHAT?!!" - I hear some of you saying...
So, to make this easy (and since most of the audio is coming from the LD, Spanish dub BD and Italian dub DVD) let's just use those three as the example...
So, for most of the audio, the LD is still the base, so on the "extra" audio track above it in my project timeline (where I'm inserting all the missing elements) I insert a track consisting of nothing but a few second loop, looped for entire film, of the noise plain from the Spanish dub except where I'm inserting audio from Spanish dub (believe it or not, the Spanish and Italian dubs are nearly identical in terms of the background noise, which makes this all easier) and on portions where I'm fixing the LD audio and have to splice it out, I have a different track consisting of nothing but a few second loop of the LD noise plain. What this does is keeps the noise plain of both tracks playing throughout the entire film, making for seamless transitions between sources. Yes, at times, the noise level is a little higher than desirable, (I will be the first to admit it) BUT there are is a benefit to this in a lot of the background stuff that's usually muffled being brought "back out" AND the noise can be cleaned up to a large extent in the post-processing part of this project. (I have already conducted some tests to see how well this can be done and while not perfect, can be done successfully)
I might have everything mapped out, but now I have to make all of the actual splices in all the sources and make sure I have everything synced down to the millisecond (I am also having to cover the dialog between the Spanish and Italian versions, with each other, when there are unique parts that are not present anywhere else and SHOULD be there) It is A LOT of work, but I'm feeling a lot more confident about it now that I have figured everything out that I need to know first.
With the amount of splices and everything involved in putting this together, I will NOT be keeping a changelog because I feel that it will take way too much time as it would contain hundreds of entries (I stopped counting after about 300 around the 20th minute of the film when I was making my handwritten notes) that would span for numerous pages. I'll be able to do this in about half the time if I'm not constantly stopping to log each change...
As far as status updates, I guess for now I will just mention how far in I am in regards to rough timecode. For now, I have just over three minutes finished. I know that doesn't seem like much, but I haven't had a lot of time to work on the actual editing yet... first had to figure out what I'm editing and now gotta find the time to do it all
:victory: