audio help please?

ushere schrieb am 09.05.2016 um 07:38 Uhr
ok, this is a plea for UNPAID audio help:

i have a commission from the local regional gallery to produce a video (to be displayed alongside an exhibition) of - '3 women went to china'. i took it on because one of the artists happens to be my wife, and the other two good friends. we are promised a budget from the local council from either anything left over from this financial year, or next years, if there is a budget next year. either way i'm not holding my breath given that our recent governments seem to put the arts in the same basket as climate change. ie. neither exists other than in the imagination of weirdo lefties, who might even be commies....

anyway. as i have often written, my hearing if fubar, and even more so than when i last wrote. the three interviews were shot on different mikes, and though intelligible are somewhat muffled. could well be over compressed. anyway, i'd like to try and sharpen them up a bit. unfortunately my idea of sharpening nowadays means everything sounds like it's coming trough a 2" speaker in a circa mid 60's transistor radio....

the audio is approx 12 minutes, with clear breaks between interviewees.

if anyone has time on their hands, and a good samaritan disposition please email me and i'll send a link to a .wav file via dropbox.

i know this is a chutzpah, but hey, in lieu of payment (for the moment) i'll give you a credit!!!

Kommentare

Grazie schrieb am 09.05.2016 um 07:52 Uhr
I'm in. You have my email addie Leslie.

G

Grazie

PC 10 64-bit 64gb * Intel Core i9 10900X s2066 * EVGA RTX 3080 XC3 Ultra 10GB - Studio Driver 551.23 * 4x16G CorsVengLPX DDR4 2666C16 * Asus TUF X299 MK 2


Cameras: Canon XF300 + PowerShot SX60HS Bridge

ushere schrieb am 09.05.2016 um 08:47 Uhr
ping - you have post...
Serena Steuart schrieb am 09.05.2016 um 08:52 Uhr
I'd offer too, but Grazie is probably enough to produce the best possible processing.
john_dennis schrieb am 09.05.2016 um 08:58 Uhr
This thought just occurred to me. How are you going to be able to tell if we did any good?
Serena Steuart schrieb am 09.05.2016 um 09:41 Uhr
Grazie, in one of the interviews the audio was recorded with two mono mikes, not one stereo. Only one audio track should be chosen and I'd suggest the R track. Using both will have some phase problems that won't help at all. Whoever recorded the tracks should have mentioned that to Leslie.
ushere schrieb am 09.05.2016 um 10:50 Uhr
This thought just occurred to me. How are you going to be able to tell if we did any good?

simple, i'll get my wife to listen ;-) she originally told me it sounded awful, but by that time she'd told me that i'd well and truly screwed it up...

i would dearly have liked to go back to the originals and rebuild the audio - i posted something to that effect a wile ago, re. pluraleyes. unfortunately the cut i have now is very heavily edited from the originals and of course, pluraleyes doesn't work that way - i just wish there was some software that would.

yes, i know i could do it 'manually', but i really don't have the time.
Grazie schrieb am 09.05.2016 um 11:31 Uhr
Serena, Leslie gave me just one channel. It would appear that either both channels have now been combined into one audio channel, or the best channel has been selected already. What there is is a range of echoing dolphin's/diver's ears type artifacts with sibilance and bass booming, folded in for good measure too.

Serena, you sound like you were in on this prior to Leslie posting this?

The only time I will choose to spend on it is to quickly determine if I could dig something out. From this there would need to be some very good autopsy determination going on.

So, Leslie any of the 2 Channel available?

G

Grazie

PC 10 64-bit 64gb * Intel Core i9 10900X s2066 * EVGA RTX 3080 XC3 Ultra 10GB - Studio Driver 551.23 * 4x16G CorsVengLPX DDR4 2666C16 * Asus TUF X299 MK 2


Cameras: Canon XF300 + PowerShot SX60HS Bridge

ushere schrieb am 09.05.2016 um 11:48 Uhr
serens kindly shot one of the interviews for me down in melbourne.

i've just emailed you a link to samples of 'clean', original single channel from all three interviews.

Serena Steuart schrieb am 09.05.2016 um 12:51 Uhr
Grazie, I haven't had a listen to the edited audio. I thought the problem might have been caused by my providing two mono tracks (separate lapel mikes) and these being mixed. Apparently Leslie was smart enough to see what he'd received. Just now checking the original and the audio seems satisfactory.
Grazie schrieb am 09.05.2016 um 12:56 Uhr
So, does that mean I should stand down? Or work with the latest Leslie has DropBoxed me?

Grazie

PC 10 64-bit 64gb * Intel Core i9 10900X s2066 * EVGA RTX 3080 XC3 Ultra 10GB - Studio Driver 551.23 * 4x16G CorsVengLPX DDR4 2666C16 * Asus TUF X299 MK 2


Cameras: Canon XF300 + PowerShot SX60HS Bridge

Serena Steuart schrieb am 09.05.2016 um 15:00 Uhr
Please continue. If the audio in the edited video can be rescued that will be good. If the original audio can be substituted without hand matching, that will be wonderful. I don't know of a means to automatically do the matching and replacement, but there's a lot I don't know.
riredale schrieb am 09.05.2016 um 17:38 Uhr
Not sure I understand exactly what you guys are working with and trying to do, but I do know that one can easily match raw audio with an edited soundtrack using Pluraleyes. A few years ago I did a bunch of 4-channel surround DVDs and the rear two channels were recorded onto Minidisc, with no timecode or metadata identifying the separate takes. Pluraleyes was easily able to find sync for each individual take on that continuous wav file, so I would think that if what you are trying to do here is to match and then replace audio by manually hunting for it then PluralEyes should be able to help. But I really don't know what it is you guys are doing, so apologies if my comments are even more useless than usual.
Grazie schrieb am 09.05.2016 um 18:47 Uhr
I agree rirdale, I've used PE to sync audio to audio. Not many people know this.

G

Grazie

PC 10 64-bit 64gb * Intel Core i9 10900X s2066 * EVGA RTX 3080 XC3 Ultra 10GB - Studio Driver 551.23 * 4x16G CorsVengLPX DDR4 2666C16 * Asus TUF X299 MK 2


Cameras: Canon XF300 + PowerShot SX60HS Bridge

Serena Steuart schrieb am 10.05.2016 um 02:04 Uhr
Sounds good. I haven't used PE and reading a review it looked like it would do this. Leslie says he didn't succeed so worth having another experienced worker on this.
Grazie schrieb am 10.05.2016 um 07:01 Uhr
OK, so the original Audio is good to go? Yes? But Leslie has the video edited down - yes? But the edited Video has this already "improved" audio - yes? So the way I see it is that what is needed to be done is:

1] Grazie not to spend any more time on attempting to further "improve" the "improved" Audio.

. . but rather . . .

2] To spend time to match the NOW edited/cut video BACK to the original Audio.

I suggest that either I or another gets hold of a low-res Vid stream of the edit with the "improved" Audio and the complete "non-improved" Audio channels and cross-matches to each other to substitute the original (Serena you say it is good) for the "improved" audio that has been sent to me by Leslie.

I'm happy to be part of that process and experiment with a 1 minute small section of edited Video and use this as a patch to get the better Audio to marry together. Yes? Do you and Leslie understand? Another way would be to set up Markers at the INs and OUTs of the events and transcribe those, using a SpreadSheet to the original dallies. I do this for clients wanting me to use their cuts against my shot dallies. What I would need from Leslie would be a lowres of the footage video+audio. Actually, maybe I could do this with just the audio? . . .. Hmmm...

BTW, the "improved" audio I got from Leslie has some hollowness and sibilance that can be ever so slightly reduced, however if the consensus is that the original Audio is more-or-less, good-to-go then I'd suggest that that be used as the way forward - yes? Far better than polishing the proverbial!

Until now I wasn't aware, Serena, you had been in on the capture. I wont do anything further till I hear directly from Leslie.

Grazie

Grazie

PC 10 64-bit 64gb * Intel Core i9 10900X s2066 * EVGA RTX 3080 XC3 Ultra 10GB - Studio Driver 551.23 * 4x16G CorsVengLPX DDR4 2666C16 * Asus TUF X299 MK 2


Cameras: Canon XF300 + PowerShot SX60HS Bridge

Serena Steuart schrieb am 10.05.2016 um 11:06 Uhr
Leslie has gone off to Sydney for a couple of days so indeed wait for his directions. Your 'yes' answers to your questions are correct.
I understand that Leslie would like to go back to square one with the audio (prior to processing it) but found he had overwritten the originals. I understand that he got into difficulties trying to improve the audio he recorded with his Senn 88 mike, which may be the hollowness you have observed (I haven't heard or seen the interviews he shot).
If you judge the processed audio to be fixable then that would be his preferred outcome.
My sole contribution has been to shoot one of the interviews here in Melbourne because the project budget (none) didn't enable the artist to travel up to Leslie in the Hunter Valley.
I can't comment on how processed audio sounds and assume that it isn't good because Leslie is quite unhappy with it. Leslie hoped that something could be done because the alternative of replacement with the original audio would be more work than he was could really afford to put in. I don't know if he had tried to use PE or only assumed that it wouldn't work (in reverse, as he put it).
The manual spreadsheet approach is about how I would go about replacing the audio (for lack of better knowledge) and in fact I'm happy to do that if Leslie wishes. However if you can demonstrate a more efficient approach I'm sure he'll be most grateful.
Grazie schrieb am 10.05.2016 um 11:10 Uhr
Thank you for your fuller explanation, as to what I had received and from where.

As you say, I await his further instructions.

G

Grazie

PC 10 64-bit 64gb * Intel Core i9 10900X s2066 * EVGA RTX 3080 XC3 Ultra 10GB - Studio Driver 551.23 * 4x16G CorsVengLPX DDR4 2666C16 * Asus TUF X299 MK 2


Cameras: Canon XF300 + PowerShot SX60HS Bridge

Serena Steuart schrieb am 10.05.2016 um 11:18 Uhr
Grazie, I edited my post and possibly added extra information. However Leslie is the man.
ushere schrieb am 11.05.2016 um 14:38 Uhr
hi grazie, serena...

first off thanks to both of you for your continued interest - obviously we're all past the age of 30 sec attention spans ;-)

serena, your audio is fine, i haven't messed around with it. sarah sounds crisp and clear (to me anyway). it's the first two interviewees that worry me - and they're the ones i played with...

grazie, i've just got back in and am knackered. however, i would love you to do what you can, as i've written off forum to you.

enough - bed calls...

farss schrieb am 12.05.2016 um 00:46 Uhr
Ushere said:

[I]it's the first two interviewees that worry me - and they're the ones i played with...
[/I]

What I'm hearing in the first one sounds like flanging.
That could have been caused by mixing the audio from two microphones or from aggressive noise reduction.

It'd be good to have a minute or so of the original camera audio.

Bob.
ushere schrieb am 12.05.2016 um 02:32 Uhr
hi bob - am sending email and samples...
farss schrieb am 12.05.2016 um 03:15 Uhr
Thanks,
had a listen. The "Sarah" audio has quite a bit of crackle in it. Aside from that, it and the other two are fine.

Using some gentle crackle removal with Rx I can get rid of most of the crackle, too much and I get too many artefacts. I would consider just that gentle treatment with Rx is enough.

Bob.
Grazie schrieb am 12.05.2016 um 07:02 Uhr
I agree with you Bob. You call it "crackle", is that the sibilance I'm hearing? Now, will you do it for Leslie? Leslie had been telling me of whoolliness. Do you also hear that?

G

Grazie

PC 10 64-bit 64gb * Intel Core i9 10900X s2066 * EVGA RTX 3080 XC3 Ultra 10GB - Studio Driver 551.23 * 4x16G CorsVengLPX DDR4 2666C16 * Asus TUF X299 MK 2


Cameras: Canon XF300 + PowerShot SX60HS Bridge

Serena Steuart schrieb am 12.05.2016 um 07:46 Uhr
sibilance, I think.