VHS transfered to DVD, sound has hisss

[rockonvidtrade1]

Hello , i have old footage of a band i enjoy , the picture is great but it has alot
of VHS hisss noise in the background , any way to get rid of ?
cheeers

[Baldrick]

Post a short audio sample.

[rockonvidtrade1]

its on dvd , how do i do that , i have the ripped file on my hdd

[videobruger]

Use PgcDemux. Select a single cell. Use "Upload files/Manage attachment"

[lordsmurf]

Certain audio issues can be trivial to fix, with the right software and experience.
Post a sample.

[johnmeyer]

Two thoughts:

1. Hiss is common on the linear VHS audio track, but not on the Hi-Fi track. If you still have the VHS tape and can transfer it again, you can check to make sure you are transferring from the Hi-Fi track.

2. Removing hiss from a non-Dolby source (which is known as single-ended noise reduction, unlike Dolby which is double-ended) is difficult to do well. The tools that do a good job are very expensive. In my experience, by far the best tool is iZotope RX4 (I own RX3 Advanced). It can perform minor (and major) miracles with audio. You can download the trial version and test it yourself. If you post some audio, I'll apply the noise reduction and you can judge for yourself.

[lordsmurf]

Nah, hiss is usually on the HiFi, too.
Hiss can be removed really well by Sound Forge, Audacity, or a combo on the too.
Sometimes Goldwave works well, too.

[johnmeyer]

If there is hiss on the Hi-Fi channel, then something is wrong. By contrast, there is always hiss on the linear track.

The hiss on the linear channel is due to the fact that that audio is recorded on an extremely thin strip of oxide along the outer edge of the tape. However, the main reason for the sometimes overwhelming hiss is tape speed. A reel-to-reel audio tape recorder uses one of three tape speeds: 1 7/8 ips; 3 3/4 ips; and 7 1/2 ips. Only 7 1/2 ips is considered "Hi-Fi." By contrast, the forward tape speed on a VHS tape is 1 1/3 ips in SP mode (the fastest) going down to only 2/3 ips in EP mode. At this speed, the frequency response is limited to about 4 kHz. This is about the same as AM radio, and the S/N ratio is horrible, with tape noise ("hiss") becoming prominent, and at times, overwhelming.

I don't know about Goldwave, but I've owned Sound Forge for at least a dozen years. The NR plugin used to be a separate purchase, but is now included. In its day (around 2003) the Sound Forge noise reduction plugin was considered pretty good, but Sonic Foundry (now owned by Sony) has not upgraded that noise reduction in a dozen years. It is now hopelessly out of date, and completely and totally outclassed by iZotope RX. The difference is not subtle or minor: you can achieve far, far better results with iZotope, and without getting those horrible flanging artifacts.

Finally, my business is restoring all manner of media: photos, audio, movie film, and video. I deal with these issues every day. While I don't have a before/after immediately at hand to demonstrate linear track hiss reduction, I do have something which shows the kinds of advanced problems I have to deal with. So, even though this is somewhat OT, here is an example of the kind of thing I do. AFIK, no one has ever figured out how to do this particular restoration. I was able to figure out how to remove the noise bars from the only known recording of a 1950s NFL TV broadcast. Yes, there is plenty of "NFL Films" footage, but because videotape hadn't been invented, and Kinescopes of football games weren't made, there is no commercial record of the actual TV broadcast. This footage was taken by someone at home who simply pointed a movie camera at the TV screen, resulting in the predictable beat pattern noise bars. I figured out how to remove them.

.

[lordsmurf]

If your speakers are not reference grade, you probably won't hear the hiss.
But it's there.

[johnmeyer]

If your speakers are not reference grade, you probably won't hear the hiss.
But it's there.

Maybe we're talking about different things. The hiss on a VHS linear track can be heard with a speaker made out of a tin can and a string: it is not subtle.

There was a reason that JVC (and all their licensees) added Hi-Fi in 1984: VHS linear track audio stinks.

Even after Hi-Fi was introduced, it was a long time (early 1990s) before camcorders incorporated this better audio technology. And, even after Hi-Fi was available , most VCRs still recorded the same audio to the more-or-less-useless linear track, something that has occasionally proven useful to me when, for some strange reason, the Hi-Fi audio is borked.

What I have found, when people send things to me for restoration that they have already transferred themselves, is that once in awhile they will accidentally enable the linear track, thus resulting in lousy audio. And, for video taken on most pre-1990 camcorders, the linear track is all they have.

[JVRaines]

If the hiss really bothers you, put the tape away and wait 30 years before you listen to it again. You will find the hiss is much quieter.

[johnmeyer]

If the hiss really bothers you, put the tape away and wait 30 years before you listen to it again. You will find the hiss is much quieter.

As someone who is north of 60, I can tell you that this is true.

[bendixG15]

If the hiss really bothers you, put the tape away and wait 30 years before you listen to it again. You will find the hiss is much quieter.

As someone who is north of 60, I can tell you that this is true.

As someone north of 70, I say this, listen to the hiss and enjoy it, while you can....

[johnmeyer]

As someone north of 70, I say this, listen to the hiss and enjoy it, while you can....

That's too good.

I just got back to my editing computer. I still have about 20 hours of very typical "home movie" VHS transfers I did for someone, all with linear-only audio. Here's a short clip. The first half of the clip is the original audio. Those under 30 can tell us if they hear any hiss.

The second half of the clip is the same audio after processing with RX. I not only reduced the hiss, but also the horrible camera motor noise, another issue for home video recordings. While I was at it, I removed a few dozen "thumps" where the camera operator bumped the camera with his hand.

Before-After Hiss Reduction

There is some very slight artifacting you'll hear at the beginning of the restored version, but you shouldn't hear any flanging or other unpleasantness with the audio. The audio at the very end of the clip, with the adult speaking softely in the background a long distance from the camera, is now intelligible.

[JVRaines]

Nice job, though of course, it dulls the voices just a tad.

The flyback transformer in TVs used to bother me when I was a kid. I could tell when a TV was on even with the speaker muted. Now that I can't hear it anymore, they go and get rid of all the CRTs!

[lordsmurf]

Valiant attempt, johmmeyer. But all that really did was muffle the entire clip. And then at best, it maybe removes a third of the hiss/hum noise. It still has lots of noise.

Listen to my restore work:

[johnmeyer]

I had the same problems with both flyback transformers as well as the ultrasonic burglar alarms that first started to appear in stores in the late 1960s.

As for the dulling of voices, maybe yes, and maybe no. You have to A/B the two while looping, which I didn't do in my sample, but which I did do on the timeline. There is a very interesting effect of noise making audio appear "brighter." I think the high frequency of the noise makes your ears think that the actual program audio also contains higher frequencies.

Oddly, the same thing also can happen with video or film, where noise or grain can provide the appearance of a picture that is sharper than it really is.

Here is a repeated loop of the end of the clip I posted, before-after-before-after, etc. It's annoying to listen to, but that's what you do when working on noise reduction. Concentrate on the voice, and I think you'll find the upper frequencies, such as they are, are still there.

Before/After (looped)

[lordsmurf]

There's a very definite muffling going on in your before/after.

Sometimes, yes, you must muffle to make it overall better. But not this time. Not this sample clip. The full tonal value can still exist, sans hiss/hum noise. It was really easy to isolate and attack here.

Mine could have probably been made more perfect, but I don't want to spend lots of time on it. That was good enough. It probably would have been a bit better, had I created a custom filter.

[johnmeyer]

Valiant attempt, johmmeyer. But all that really did was muffle the entire clip. And then at best, it maybe removes a third of the hiss/hum noise. It still has lots of noise.

Listen to my restore work:

Well, as the French like to say, à chacun son goût: each to his own taste.

As I mentioned, I do restoration work for a living. I've been doing this for over a decade. When I first started doing it, I was thrilled with various tools I bought because they could totally eliminate whatever artifact I was trying to get rid of. So, I cranked up the settings and enjoyed the total absence of noise ...

... until I realized that the result was artificial, annoying, and not at all pleasing.

What I found is that the real goal of restoration, IMHO, should be to make the resulting media more enjoyable to view or listen to. So, if my restoration got rid of ALL the noise, but introduced all sorts of "breathing" and "thumping," which your noise gate approach does**, and if it cut off the beginning and ending of words, as noise gates always do, then that approach had to be abandoned.

As a result of thousands of hours doing this work, I developed this mantra:

Noise reduction is better than noise elimination.

In the video realm, I've seen lots of results from people who want to get rid of every last dot of film grain, or every last speck of video sensor noise. They succeed, but at a great cost to how the video looks. Their approach adds new artifacts, just as your noise reduction effort adds new artifacts.

I do a lot of work for people like the sports film collector who supplied that NFL film video sample I posted earlier in this thread. People like him are very concerned about fidelity to the original intent. While they permit me to do some restoration, they often reject my work when the result looks or sounds artificial to them.

So, I come back to my initial statement: to each his own. In the end, I usually send samples of my efforts to the client before I undertake the entire restoration, and let him or her decide if they like the results.

I am, however, glad that you pulled out your reference grade speakers so you could hear noise in the linear audio track.

-------------
** Even if you didn't use a noise gate, many NR tools use gating as part of the algorithm. A noise gate cuts off absolutely all audio when the average audio falls below a certain dB level. Unfortunately real life doesn't take room noise to zero when someone stops speaking, and there are often barely-audible sounds in the background that we would still like to hear.

[JVRaines]

Le goût de l'Audition:

johnmeyer


lordsmurf

[johnmeyer]

P.S. I just A/B/C'd the original, my restoration, and yours. Listen to the last few seconds of your restored audio: the speech is almost unintelligible. I'm not trying to pick a fight, but instead just want to point out that your stronger noise reduction does do damage to the audio that results in it being more difficult to understand. This gets to the points in my previous post.

The famous architect, Mies van der Rohe, was fond of saying "less is more." It is true in architecture, but even more true when doing restoration work.

[edit]I too was going to post the spectrum analyzer shots, but wasn't sure anyone could interpret them. You have to specify what portion of the clip you feed into the spectrum analyzer.[end edit]

[JVRaines]

Kinda messed it up; sorry, I'm rusty on the image upload tools. Anyway, I used the "after" portion of john's file and the entire smurf file.

What I see and hear: john's file slopes from 6 to 15 KHz then abruptly cuts off. smurf's file dives down from 6 to 9 Khz, then jumps back up to 15 KHz and gently arcs down. So smurf's file has a high frequency emphasis that john's is lacking, but dumps high vocal frequencies, which makes the women's speech at the end of the clip unintelligible.

[johnmeyer]

Kinda messed it up; sorry, I'm rusty on the image upload tools. Anyway, I used the "after" portion of john's file and the entire smurf file.

What I see and hear: john's file slopes from 6 to 15 KHz then abruptly cuts off. smurf's file dives down from 6 to 9 Khz, then jumps back up to 15 KHz and gently arcs down. So smurf's file has a high frequency emphasis that john's is lacking, but dumps high vocal frequencies, which makes the women's speech at the end of the clip unintelligible.

Good catch. My restoration tool lets me specify a curve that can force the algorithm to be more aggressive at some frequencies than others. I did tell it to be pretty aggressive at frequecies > 5 KHz. Perhaps I should have been less aggressive at those higher frequencies.

I really welcome the feedback, because my goal is always to satisfy the client, and everyone has different perceptions and tastes. Thanks, JVRaines, for taking the time to give me your analysis.

[lingyi]

As a former Betaphile, just want to add to and clarify a few points in your post. I'm sure lordsmurf and others will correct any incorrect information I may post.

If your speakers are not reference grade, you probably won't hear the hiss.
But it's there.

Maybe we're talking about different things. The hiss on a VHS linear track can be heard with a speaker made out of a tin can and a string: it is not subtle.

There was a reason that JVC (and all their licensees) added Hi-Fi in 1984: VHS linear track audio stinks.

JVC added Hi-Fi capability to VHS as a response to Beta Hi-Fi which was introduced in 1983. VHS had previously upped the ante by splitting the already thin linear track into two for stereo sound. With the execption of a few non-Sony Betamaxes and Sony professional desk, linear stereo was skipped for Beta Hi-Fi.

Even after Hi-Fi was introduced, it was a long time (early 1990s) before camcorders incorporated this better audio technology. And, even after Hi-Fi was available , most VCRs still recorded the same audio to the more-or-less-useless linear track, something that has occasionally proven useful to me when, for some strange reason, the Hi-Fi audio is borked.

VHS Hi-Fi (and PAL Beta Hi-Fi) is recorded using depth multiplexing, which records the Hi-Fi track first below 1.6Mhz, then places the higher frequency video track over that.

Because the VHS Hi-Fi information is "below" the video information and therefore weaker, there were some issues (especially early on) particularly with playback. As I recall, some early VHS Hi-Fi recordings would even cause herringbone interfence on non-Hi Fi machines.

The linear audio track is on the bottom of the 1/2" tape and is subject to damage when loading and unloading. The control track is wisely on the top of the tape where damage is less likely. This is why you can sometimes salvage one of the stereo channels (don't recall whether it's left or right that's on the bottom) from a damaged linear track.

What I have found, when people send things to me for restoration that they have already transferred themselves, is that once in awhile they will accidentally enable the linear track, thus resulting in lousy audio. And, for video taken on most pre-1990 camcorders, the linear track is all they have.

AFAIK, with the exception of professional decks, ALL consumer VCRs (both VHS and Beta) always recorded the linear track simultaneously with the Hi-Fi.

[lordsmurf]

As I mentioned, I do restoration work for a living. I've been doing this for over a decade. When I first started doing it, I was thrilled with various tools I bought because they could totally eliminate whatever artifact I was trying to get rid of.

Hey, good for you! I've been doing it professionally for about 10-15 years as well, and it was a hobby for many before that. Before my medical issues, I had done work for studios. My entry to video was NOT digital, but had begun in the S-VHS analog era. BTW, you and I had email conversation more than a decade ago. I actually still have several emails from you from 2004.

So, I cranked up the settings and enjoyed the total absence of noise ...

Yeah, everybody has to learn this the hard way. I learned this about digital audio in the 90s.

the speech is almost unintelligible.

She was unintelligible period. She was clearly a good 10 feet or more away from the mic. And again, it could be improved with custom filter work. This was just a quick example of how you don't need to muffle, and how you can better remove noise. I just ran it through a few pre-existing customs, and didn't make a new one. (That's what paid work is for!)

What I found is that the real goal of restoration,

Restoring is about making it better, not making it perfect. I've said this for years. (However, don't half-ass it either. "Good enough" is not good enough.)

In the video realm, I've seen lots of results from people who want to get rid of every last dot of film grain, or every last speck of video sensor noise. They succeed, but at a great cost to how the video looks. Their approach adds new artifacts, just as your noise reduction effort adds new artifacts.

Sometimes you have to trade a better error for a worse one. For example, the Panasonic ES10 helps remove tearing, but often leave posterization instead. Which is better? (Posterization for sure.)

I am, however, glad that you pulled out your reference grade speakers so you could hear noise in the linear audio track.

There was so much present, that I could hear it on normal speakers. Sorry to disappoint. (Admittedly, my hearing is far better than the average person. It's one reason I do what I do. And my "normal" speakers are probably better than most as well.)

[johnmeyer]

Ha! You remember my emails. I am flattered. I too still have them and remember having sent them. Your site was one of the first really helpful places to get information you could actually use, rather than containing just a lot of opinion. You had already begun posting great information about media quality (I think Ritek was still considered the best stuff back then, and Taiyo-Yuden and Verbatim hadn't yet emerged). You were just starting to add some pages about restoration, and I wanted to share some things I thought I had learned. I think this was in the earliest days of my own restoration explorations, and I had just discovered that you can capture the same videotape multiple times, average the resulting captures, and reduce the noise. It worked, but yikes, was that ever a lot of effort!

I bring all that up, not only as a way of saying hello, but also to acknowledge my own failings in this field. In particular, every time I think back to what I thought I knew about this when I first started, and now realize most of it was wrong, I wonder what I'll think when I look back in another ten years (if I make it that far), and remember what I'm doing now. Will I think I was clueless?

I always want to go back and re-do every restoration project I've ever done. It is never quite good enough.

Thanks for all your good work over the years.

[davexnet]

Even remastered CD's, using the same analog source tapes, worked on by different, experienced engineers,
often have obvious differences in the sound of the finished product .

I know it's apples and oranges, but it's still an example of how one experienced guy (or girl) thinks it should be.
Whether it be EQ choices, NR usage (or not) or overall loudness.

I'm a Hobbyist myself, particularly interested in vinyl restoration. As a learning aid, I downloaded a few needle drops
from youtube; this is an interesting challenge, the kinds of noises can vary so much. It's so easy to lose the
energy of a song if the NR is wrong.

I agree about Sound Forge NR, it's poor compared to other products. I haven't invested in any expensive tools
(like Izotope), but I have the Sony NR, Bias Soundsoap and an older version of the Waves restoration bundle.

[lordsmurf]

I never use the SF presets. They are pretty much useless. SF is only good if you use it as a tool to create your own filter presets.

I have older versions of iZotope and DiamondCut. But neither is all that useful for movie/video audio. (DC is quite useful for forensic work, however. iZotope is more for processing audio tracks for music.)

[johnmeyer]

I have older versions of iZotope and DiamondCut. But neither is all that useful for movie/video audio. (DC is quite useful for forensic work, however. iZotope is more for processing audio tracks for music.)

The company, iZotope, makes quite a few different audio products:

iZotope Products

When you say that iZotope is more for processing audio tracks for music, that is most definitely true of many of their products. However, the RX series is specifically for audio restoration, and I don't think would normally be found in a music studio, whereas the others would be.

If you haven't tried iZotope RX (which is now at version 4), it is an absolute revelation. I have not found anything under $10,000 that even comes close.

John

[davexnet]

Regarding the RX series. If I were making some money from audio work, I might have invested in it.
I did trial it once. Seemed very impressive. Eventually I found some instructional videos devoted to it.
Even without the product, the videos were useful for general audio tips and tricks.

At the other end of the scale there's a VST plugin called Elevayta Klean which seems promising -
claims to be able to remove the noise without harming the music. $29.99.
I trialled this also. It left gaps in the audio every 20 seconds in trial mode. Difficult to evaluate properly.