#1  
05-15-2021, 11:39 AM
mirkorm mirkorm is offline
Free Member
 
Join Date: Apr 2021
Posts: 15
Thanked 0 Times in 0 Posts
Hi everybody, some days ago I was struggling with video capture levels and after some tweaking I managed to get a good image from my VHS-C thanks to you kind people on digitalfaq but now it's a couple of days I'm struggling with audio:

I'm still working on this copy that I have both on the original VHS-C from the camcorder and the VHS copy made back in 97 connecting the camcorder to my old VCR because I'm trying to make sure I'll get the right video and audio for all my VHS-Cs made with the same camcorder, since this is the only recording I have on both media.

Now: I'm noticing the sound captured from the VHS-C sound very muffled, it is missing some frequencies, while the VHS one has a way cleaner sound.

This is the information I've managed to gather so far:

- since I've captured both audios through my sound card's line-in input, I've excluded that it is a capture device problem, because being it the same that captured the audio from the VHS, if that was the case it wouldn't have captured those frequencies from that tape either.

- and I've excluded it's actually the problem of the camcorder or VHC-S on which it was recorded or those frequencies wouldn't be in the VHS transfer either.

I add that the only thing I know of that can be set for the audio input is though the windows mixer, I've tried different volume levels but nothing changed, and that I use a VHS adapter to capture VHS-Cs using the same VCR, audio and video capture devices for both VHS and VHS-C.

I attached small samples of the same audio from the VHS and VHS-C captures, I'm trying to mix the audio in audacity but those frequencies (especially the high ones) seem completely missing from the capture.

I was wondering if there was a way to set some kind of equalizer on capture (my Realtek driver doesn't natively have one) or if you know of some way to help capture a better audio.

Thanks


Attached Files
File Type: wav vhs extract.wav (1.51 MB, 26 downloads)
File Type: wav vhs-c extract.wav (1.61 MB, 13 downloads)
Reply With Quote
Someday, 12:01 PM
admin's Avatar
Ads / Sponsors
 
Join Date: ∞
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #2  
05-15-2021, 04:51 PM
timtape timtape is offline
Free Member
 
Join Date: Sep 2020
Location: Perth, Western Australia
Posts: 536
Thanked 99 Times in 89 Posts
Possibly there is a slight misalignment of the VHS deck's A/C head azimuth re the VHS-C recorded audio azimuth, leading to "comb filtering". In your VHS-C sample I hear what is probably comb filtering centred around 5 kHz which reduces speech clarity and "presence". If so the only way to fix this is to align the head azimuth manually, using a suitable small demagnetised screwdriver while listening to the playback. Sometimes the adjustment needed is less than 1/8th of a screw revolution. Back in the day some people adjusted head azimuth with their audio cassettes. It's harder with VHS or Beta because we have to remove the VCR's cover, find the correct screw and carefully adjust it without doing damage to ourself, the deck or the tape. Some of us adjust azimuth on every tape. Most people dont bother, if they even know about the phenomenon.

The A/C head may also be dirty. It takes only a small amount of dirt on the head face to muffle the audio. The A/C head may also be worn, complicating azimuth adjustment.

You can try and equalize after the fact but it will bring up the hiss. It will not lift the wanted audio above the hiss.

If you Google "tape head azimuth", most references and photos are to audio cassette or open reel audio tape head azimuth. Almost never to VCR A/C head azimuth misalignment even though it's a very common problem, contributing to less than ideal playback of the linear audio track.

Having said that I've heard a lot worse muffling problems than in your sample, even on commercial/professional video transfers.

Last edited by timtape; 05-15-2021 at 05:11 PM.
Reply With Quote
The following users thank timtape for this useful post: mirkorm (05-16-2021)
  #3  
05-16-2021, 05:37 AM
mirkorm mirkorm is offline
Free Member
 
Join Date: Apr 2021
Posts: 15
Thanked 0 Times in 0 Posts
Tanks timtape, I'm looking deeper into what you suggested. Is the vhs-c a peculiar case for heads alignment? I'm asking because the vhs was captured without the combing effect on the same vcr, wouldn't head misalignment reflect badly on regular vhs capture too?

Thanks for your kindness and patience
Reply With Quote
  #4  
05-16-2021, 07:57 AM
timtape timtape is offline
Free Member
 
Join Date: Sep 2020
Location: Perth, Western Australia
Posts: 536
Thanked 99 Times in 89 Posts
Quote:
Originally Posted by mirkorm View Post
Tanks timtape, I'm looking deeper into what you suggested. Is the vhs-c a peculiar case for heads alignment? I'm asking because the vhs was captured without the combing effect on the same vcr, wouldn't head misalignment reflect badly on regular vhs capture too?

Thanks for your kindness and patience
If the VHS capture was sourced from the same camcorder which shot the footage then there wouldnt normally be an error as the camcorder head should normally have not changed its azimuth setting, especially if the transfer was made close to the time of shooting.
The problem usually comes when playing back the tape on another camcorder or VCR, perhaps after the original camera or deck has failed or been discarded. Audio azimuth misalignment is not specific to VHS or VHS-C. The effect is worse with LP or ELP recordings. Hope this helps.
Reply With Quote
  #5  
05-16-2021, 08:57 AM
mirkorm mirkorm is offline
Free Member
 
Join Date: Apr 2021
Posts: 15
Thanked 0 Times in 0 Posts
- The video was originally shot with a camcorder on a VHS-C in 95 and copied to a regular-sized VHS connecting the camcorder to a VCR in 97

- Now the samples I uploaded where part of digital captures of both those previously mentioned VHS-C and VHS, done connecting a VCR to a PC (I used the same VCR for all my digital transfers).

Since I captured to PC some VHS-C both before and after capturing the regular VHS I noticed that all the VHS-C have that problem so I don't think anything happened between captures.

What I actually think, and I made this guess based on what happened to the video levels, is that the camcorder to VHS copy process applied some equalization and filters by default, resulting in better audio and video.

With the video I noticed that with the default settings, the VHS-C capture to PC crushed the brights or "video high frequencies" for the sake of our logical thread, and I managed to solve that adjusting the proc amp video levels (replicating what the camc. to VCR copy did to capture a good video), that allowed me to adjust those video frequencies on capture that I couldn't possibly get by fixing the video later in post production once the video was already captured with bad video.

So: since the camc to VCR "equalized the video levels" it's possible that it did the same thing for the audio to capture the best audio possible and all the available frequencies from the source, and, as it was the case for video, if the VCR back in 97 managed to capture all those frequencies, I guess they are present in the original VHS-C in order to show in the VHS copy, but the default capture equalization or some other setting on the PC system kind of messes it up and doesn't allow to capture every and each usable frequencies resulting in a flatter or less ample sound frequency-wise.

A great thing would be to find a sort of "proc amp" but for audio.

If we'll manage to find a solution I think it would actually be a great help for everybody doing quality VHS transfer/restoration, or that just would like to be able to capture the best possible audio to start with.
Reply With Quote
  #6  
05-16-2021, 03:31 PM
timtape timtape is offline
Free Member
 
Join Date: Sep 2020
Location: Perth, Western Australia
Posts: 536
Thanked 99 Times in 89 Posts
Quote:
Originally Posted by mirkorm View Post
- The video was originally shot with a camcorder on a VHS-C in 95 and copied to a regular-sized VHS connecting the camcorder to a VCR in 97...
Yes and that you used for playback the same camcorder as the tape was recorded in, and while the camera was still in the same state as when it shot the wedding, is probably the reason for the relatively faithful audio copy. The playback was good.

Quote:
Originally Posted by mirkorm View Post
What I actually think, and I made this guess based on what happened to the video levels, is that the camcorder to VHS copy process applied some equalization and filters by default, resulting in better audio and video...
The only way to confirm that would be to eliminate a more likely explanation of a sub optimal playback of the VHS-C tape on your VHS deck today. But that involves actually doing it, playing the tape back optimally. I mentioned some pointers as to how to achieve that, but I understand that for most people this may be too difficult an exercise.

So much of the key to good video and audio digitisation is using the best original (first generation) source and playing back that best original source optimally. We only go to a copy if the first generation is either damaged (such that it now plays back worse than the best copy) or is unavailable.

If the copy really sounded better than the original only because of EQ tweaking at time of transfer, you could use the original VHS-C as source, do the same tweaking and achieve an even better result - due to bypassing the generational losses in the analog copy. Why not test this?

Last edited by timtape; 05-16-2021 at 04:06 PM.
Reply With Quote
  #7  
05-18-2021, 06:28 AM
dpalomaki dpalomaki is offline
Free Member
 
Join Date: Feb 2014
Location: VA
Posts: 1,694
Thanked 369 Times in 325 Posts
The VHS-C clip is about 4 dB higher average level than the VHS clip.

Frequency spectrum of both clips is attached for your consideration. Note the significant bleed of horizontal scan frequency in the capture at around 15750 Hz. The VHS has some strange stuff going on in it - not a very faithful copy. But no evidence of clipping. Above around 4500 Hz all you see is probably noise.

A "proc amp for audio" would be a simple two channel audio mixer with tone controls. Something like a Mackie 802 for example.

FWIW: Video high frequencies in a composite or s-video signal generally relate to fine detail, not brightness. The video signal amplitude (voltage) generally relates to brightness (color saturation for the "C" portion) and the voltage swing (range of that voltage) to contrast.


Attached Images
File Type: jpg VHS-C.jpg (57.5 KB, 7 downloads)
File Type: jpg VHS.jpg (62.1 KB, 6 downloads)
Reply With Quote
  #8  
05-18-2021, 07:40 PM
timtape timtape is offline
Free Member
 
Join Date: Sep 2020
Location: Perth, Western Australia
Posts: 536
Thanked 99 Times in 89 Posts
Quote:
Originally Posted by dpalomaki View Post
The VHS-C clip is about 4 dB higher average level than the VHS clip.
Perhaps but such an average level difference is relatively unimportant here. Levels of that small magnitude can easily be changed in post without penalty. The VHS-C capture is louder but lacks the clarity of the VHS dub. That muffled effect is the real problem.

Quote:
Originally Posted by dpalomaki View Post
Frequency spectrum of both clips is attached for your consideration. Note the significant bleed of horizontal scan frequency in the capture at around 15750 Hz. The VHS has some strange stuff going on in it - not a very faithful copy. But no evidence of clipping. Above around 4500 Hz all you see is probably noise.
Yes the 15 kHz tone is almost identical in both captures. But then it's easy to display because throughout the clip its level remains constant. Apart from that tone, why do the two plots look so very different? I suspect they display different segments in time of the one original track. What else could account for such widely different looking displays? They do seem different around 5 kHz but apart from that, when listening to the audio, why do they not sound that different.

It's important not to be guided merely by the visual display. Especially with audio, a static display can be very misleading. We cant "still frame" audio like we can the visual. It would look like a single dot. FFT displays try to get around this by averaging short or long slices of time to create at least some idea of what is going on. But with rapidly changing program like speech, an integration of the entire soundtrack, or at least here of the clip provided would be more meaningful. One of the early digital tools to display long term averages in terms of spectral energy was, I believe, Har Bal.

Dialogue especially jumps all over the place - from moment to moment - and so "grabs" of very short segments tend to be meaningless for our purposes here. I felt there was a "hole" in the spectrum centred around 5 kHz not by looking at a static display but by listening to the same audio while at the same time looking at the moving display of the same audio. Ideally we would have a series of reference test tones (or pink noise) recorded on the VHS-C tape and would measure the playback response to see how closely it matched what was on the tape. Since in this example we dont have that luxury we need to use all the tools and clues available to us from just the speech audio on the tape, including any background noise.

Last edited by timtape; 05-18-2021 at 07:53 PM.
Reply With Quote
  #9  
05-18-2021, 09:07 PM
dpalomaki dpalomaki is offline
Free Member
 
Join Date: Feb 2014
Location: VA
Posts: 1,694
Thanked 369 Times in 325 Posts
No doubt your hearing is much better than mine - I have significant hearing loss, so I have to rely on tools to help understand things I have difficulty hearing.

I took another look, this time matching the clips to the same time window so as to look at the same audio segment.
I also amplified the VHS capture so the average level, energy, and peak matched the VHS-C capture within a half dB or so. It took a fair amount of gain added to the VHS clip to match them. (This may reflect the result of ALC in the VCR used to make the VHS copy.)

The source was mono, so for display purposes I put the VHS-C audio on one chanel and the VHS on the other. This allowed ready analysis of the two sources in a common display and bringing them into sync (withing a few frames)

The AvgLevelMatched frequency display shows they track closely except for somewhat earlier roll off in the VHS copy around starting ~3300 Hz vs. ~4800 Hz for the VHS-C. (The hole.) This high frequency loss is no surprise in a copy of an old linear track recording. That could account for a dull sound. This also shows up in the spectral display. I suspect there is little if any useful audio above 6-7 KHz.

As to a fix, the "hole" probably could be equalized out with judicious application of a graphics equalizer (which may boost the noise floor as well). Matching the amplitude can help in A/B comparisons of the two captures by bringing them to a common baseline on subjective hearing response curves.


Attached Images
File Type: jpg AvgLevelMatched.jpg (77.5 KB, 5 downloads)
File Type: jpg Spectral.jpg (73.6 KB, 4 downloads)
Reply With Quote
  #10  
05-19-2021, 06:09 PM
mirkorm mirkorm is offline
Free Member
 
Join Date: Apr 2021
Posts: 15
Thanked 0 Times in 0 Posts
Thank you very much for your answers, they've all been very informative, unfortunately this week has been particularly hard and haven't managed to fix the vcr head azimuth yet, but this weekend I'll study the situation further because I wouldn't want to misalign it even further, not having ever done it before.

As for the samples I checked but the closest I've found to a reference tone is some room noise (that on the files I called room) and some kind of static noise at the beginning of the captures (called beginning). I tried to compared them on my DAW and realized how impossible it is to try to visually compare them, if you have any idea on any other part of a tape that could be useful I'll upload it and try to analyze it as best as I can.
I attached the short audio clip for the room noise and the noise at the beginning of the capture with some of the visual representations of their frequencies compared side by side that looked the closest, but as timtape said the saound always changes, they're just approximations.

The biggest questions that remained for me are:

- If the head azimut of this VCR I'm using for digital capture is skewed how did it manage to capture a bad audio only on the VHS-C but captured it right on the regular VHS (other VHS-Cs captured before and after a couple of days this VHS, all have the same issue) and

- How did that old VCR got that clean audio from that same VHS-C I've digitally captured, I've tried a bit of EQuing but they seem not to be there at all in the digital VHS-C capture

Thanks


Attached Images
File Type: jpg room.jpg (31.8 KB, 4 downloads)
File Type: jpg beginning.jpg (31.5 KB, 2 downloads)
File Type: jpg beginning2.jpg (30.3 KB, 2 downloads)
Attached Files
File Type: wav room vhs.wav (8.18 MB, 1 downloads)
File Type: wav room vhs-c.wav (8.18 MB, 1 downloads)
File Type: wav beginning vhs.wav (5.52 MB, 0 downloads)
File Type: wav beginning vhs-c.wav (5.52 MB, 0 downloads)
Reply With Quote
  #11  
05-19-2021, 07:34 PM
timtape timtape is offline
Free Member
 
Join Date: Sep 2020
Location: Perth, Western Australia
Posts: 536
Thanked 99 Times in 89 Posts
Quote:
Originally Posted by mirkorm View Post

- If the head azimut of this VCR I'm using for digital capture is skewed how did it manage to capture a bad audio only on the VHS-C but captured it right on the regular VHS (other VHS-Cs captured before and after a couple of days this VHS, all have the same issue)...
As already said, we assume there is a misalignment between the VHS-C recorded audio and that of your current VHS playback deck. It's not a huge misalignment but enough for you to have noticed the difference in speech clarity.

Quote:
Originally Posted by mirkorm View Post
- How did that old VCR got that clean audio from that same VHS-C I've digitally captured, I've tried a bit of EQuing but they seem not to be there at all in the digital VHS-C capture
The key was that 1. The camera was playing its own recording so probably no misalignment and so a clean audio capture. 2. Your current VCR fortunately happens to be aligned to the old VCR (or the original dub was in HiFi which doesnt have this azimuth problem) so a clean playback of the dubbed audio.

Often people arent as lucky and the misaligned audio playback of the VHS-C tape is "baked in" to the taped copy. If they have thrown away the original and only kept the VHS copy it's impossible to retrieve the lost audio detail. You're fortunate to have kept the original and the earlier copy. If you had only kept the original you would probably have assumed that the somewhat muffled audio on it was just the way it always was from new. Most people make this assumption when they play back originals with an azimuth error. "Oh well it's an old recording..." But you now know better. You can now make a new transfer direct from the camera tapes, avoiding the picture degradation that's inevitable in a second generation tape copy. You can also avoid audio degradation, but probably only if your current VCR's audio/control head is tuned or focussed for maximum clarity.

Again, if you're not confident to dive in and adjust the head's azimuth (it's easy for a novice to make a mess of things) why not use the picture from your new dub but use the audio from the old dub and sync them on your timeline?
Reply With Quote
  #12  
05-22-2021, 10:35 PM
timtape timtape is offline
Free Member
 
Join Date: Sep 2020
Location: Perth, Western Australia
Posts: 536
Thanked 99 Times in 89 Posts
Quote:
Originally Posted by dpalomaki View Post
No doubt your hearing is much better than mine - I have significant hearing loss, so I have to rely on tools to help understand things I have difficulty hearing.
Me also. My upper hearing limit is now about 10 kHz and increasingly I find the measuring tools a great help in diagnosing audio problems.

Quote:
Originally Posted by dpalomaki View Post
I took another look, this time matching the clips to the same time window so as to look at the same audio segment.
I also amplified the VHS capture so the average level, energy, and peak matched the VHS-C capture within a half dB or so. It took a fair amount of gain added to the VHS clip to match them. (This may reflect the result of ALC in the VCR used to make the VHS copy.)

The source was mono, so for display purposes I put the VHS-C audio on one chanel and the VHS on the other. This allowed ready analysis of the two sources in a common display and bringing them into sync (withing a few frames)
Thanks for the work on this. Yes I noticed some compression or limiting on the VHS C to VHS copy, but mainly in peaks so the general frequency distribution comparison in red and blue is still very clearly displayed.

Quote:
Originally Posted by dpalomaki View Post
The AvgLevelMatched frequency display shows they track closely except for somewhat earlier roll off in the VHS copy around starting ~3300 Hz vs. ~4800 Hz for the VHS-C. (The hole.) This high frequency loss is no surprise in a copy of an old linear track recording. That could account for a dull sound. This also shows up in the spectral display. I suspect there is little if any useful audio above 6-7 KHz.
I think you have it around the wrong way here. The "hole" in the audio is apparently in the playback of the original VHS-C audio, not the VHS copy. All things being equal, copies have higher noise and poorer response. I suggested as a possible explanation, azimuth misalignment in the playback of the original.
But without optimised playback of the VHS-C audio it's difficult to judge for certain the losses and their causes.

Quote:
Originally Posted by dpalomaki View Post
As to a fix, the "hole" probably could be equalized out with judicious application of a graphics equalizer (which may boost the noise floor as well). Matching the amplitude can help in A/B comparisons of the two captures by bringing them to a common baseline on subjective hearing response curves.
Yes boosting that "hole" area would probably help and if I had no other choice I'd do it. But again adjusting azimuth on the original tape would be my first priority if I had access to the tape. Likely, an azimuth screw adjustment taking all of 5 seconds would remain valid for all of the OP's VHS-C tape transfers so long as it was know the same camera recorded all the footage, but I understand that most people are not confident to do this.

I may upload a motion spectrogram of what changing the playback azimuth does to the audio in terms of comb filtering, from a very mild loss to severe, and everything in between. Ideally, a full screen shot of the spectrogram, with a smaller inset shot of the azimuth screw as it is adjusted. But just the audio degradation itself should be enough to make the point.

In the meantime here's a basic presentation on head alignment. All angles are important. Azimuth comes at the end and he shows how azimuth misalignment causes comb filtering or cancellation. When it's full cancellation, the signal is gone. No amount of EQ can compensate and we end up just boosting the unwanted background noise.

https://youtu.be/qHeKes8maKU

Last edited by timtape; 05-22-2021 at 11:05 PM.
Reply With Quote
  #13  
05-23-2021, 12:15 AM
lordsmurf's Avatar
lordsmurf lordsmurf is online now
Site Staff | Video
 
Join Date: Dec 2002
Posts: 13,501
Thanked 2,448 Times in 2,080 Posts
Quote:
Originally Posted by timtape View Post
My upper hearing limit is now about 10 kHz
All this talk of hearing limits got me curious. So I dug out my last tests from about 5 years ago. At the time, that test showed I could perceive at least as low as 4db, but only went to 8kHz. In other more recent tests, I can hear at least 18khz (possibly higher, must re-test), but db not tested. I don't like hearing tests, as it aggravates my audio-induced anxiety. Some Google results would tell you this is "normal range", but BS. Most people can't here that well. If that were the case, we'd have a lot less noise in the world. Realize that "good hearing" can become a burden at a certain thresholds. This is more accurate: "People of all ages without a hearing impairment should be able to hear 8kHz. People in 40s should be able to hear 12kHz, 30s 15khz, late 20s 16kHz, early 20s and teens 17kHz."

Reference-grade speakers have a good frequency response that dips in the 10kHz to 20kHz range. My Monsoons have at least 10kHz, and are near-reference. Most speakers suck, with falloff dipping badly in the 3kHz to 8kHz range. FYI, "reference" refers to flat audio curve, meaning the speakers faithfully output the input sound. So consumer speakers lose quality, and you're not hearing the true audio. Professional speakers mean you hear everything, or most everything. At higher ranges, most of the audio is air and noise, but lower quality speakers may distort and amplify the air/noise. Hence why you want quality speakers to master.

While the is slightly off-topic to the thread, sometimes speakers are to blame for bad audio quality more than the VCR. Or at least share some of the fault.

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
Reply




Tags
audio capture, audio quality, digital capture, sound card, vhs-c

Similar Threads
Thread Thread Starter Forum Replies Last Post
Audio problem: muffled, quiet VHS conversions? waloshin Capture, Record, Transfer 11 02-05-2023 04:43 PM
Panasonic AG-1970P audio muddy/muffled? uofmrapper Restore, Filter, Improve Quality 20 11-12-2020 04:44 PM
JVC SR-MV45 distorted audio when playing linear audio on SLP tapes? ChunkDaMan Capture, Record, Transfer 17 08-12-2019 04:55 PM
Capturing VHS: Audio to AIW 9800 Pro vs. audio to sound card? ChunkDaMan Capture, Record, Transfer 4 07-31-2019 04:18 AM
Consumer JVC VCRs: Muffled Linear EP Audio? How to fix? Belmont Capture, Record, Transfer 5 08-06-2012 02:45 PM

Thread Tools



 
All times are GMT -5. The time now is 08:05 AM