join 2 clips of different quality

[mathmax]

Hello

I have two audio clips in different quality. Unfortunately the best one is not full..whereas the second one is full but in low quality. Is there any enhancement that can be made to make the lower quality clip match the other one as much as possible? For example is it possible to adjust the frequency spectrum?

http://www.mediafire.com/?j95mj67s8y492xv
http://www.mediafire.com/?7jalswsfp93wyaw

Here are 2 samples. Any advices will be greatly appreciated

thanks

[mike20021969]

The difference in lengths between the 2 files is 0.123s. One is in stereo and one is in mono.

What is the name of the music? - It may be easier to get hold of a full quality version.

[Cornucopia]

There are complex filters that can match the EQ of one file with the EQ from another file, but you probably won't find those filters on any freeware app, only occasionally on payware apps. One can use a general "convolution filter" to do this, but then it gets very convoluted (pun intended ).

You've got a more basic problem: you can't join a Mono to a Stereo file. You either have to mix down the stereo to mono first, or duplicate the mono track, so you at least have the same # of channels. NO app will/can join dissimilar channel files.

Scott

[mathmax]

Thank you for your answers

The difference in lengths between the 2 files is 0.123s. One is in stereo and one is in mono.

I didn't cut so accurately so there might be a difference of length + the tempo is slightly different..

There are complex filters that can match the EQ of one file with the EQ from another file, but you probably won't find those filters on any freeware app, only occasionally on payware apps. One can use a general "convolution filter" to do this, but then it gets very convoluted (pun intended ).

You'd have to explain a bit more.. I'm very novice in this field.. What is EQ? Equalizer? Which software can offers such filter to match the EQ? Could you also explain briefly what is the convolution filter and how to use in my case?

You've got a more basic problem: you can't join a Mono to a Stereo file. You either have to mix down the stereo to mono first, or duplicate the mono track, so you at least have the same # of channels. NO app will/can join dissimilar channel files.

Scott

the mono file is already duplicated in two channels... so sample2 is a fake stereo.

But I wonder why the two files sound drastically different. On the sample1 the sound seems way more present than on sample2. Is it only caused by compression that cut high frequencies?

Thank you again for your advices

[jagabo]

I think what you're really looking for is a bass and trebble boost on sample 1 (always start with the higher qulaity source). I assume you don't want to duplicate the high frequency cuttoff. The computer I'm on right now has crappy speakers but try this in audacity. Then use the equalizer with a U shaped curve to boost frequencies below about 400 Hz, decrease the mid range, and boost above about 4000 Hz:



and adjust the bass boost to your liking. You can use the Analyze -> Plot Spectrum tool to compare spectrums.

[sum_guy]

I used to use EditPro for audio and split a mono into a stereo and then reverse the channels and got pretty decent results, you might be able to do the same with Audacity although I've never tried it.

[mathmax]

I think what you're really looking for is a bass and trebble boost on sample 1 (always start with the higher qulaity source). I assume you don't want to duplicate the high frequency cuttoff. The computer I'm on right now has crappy speakers but try this in audacity. Then use the equalizer with a U shaped curve to boost frequencies below about 400 Hz, decrease the mid range, and boost above about 4000 Hz:

sample 1 is the high quality source. Did you equalize the high quality source or the low quality source?
I wanted to edit only sample 2 to make it match the sample 1..

I used to use EditPro for audio and split a mono into a stereo and then reverse the channels and got pretty decent results, you might be able to do the same with Audacity although I've never tried it.

what do you mean by "reverse the channels"?

[jagabo]

I think what you're really looking for is a bass and trebble boost on sample 1 (always start with the higher qulaity source). I assume you don't want to duplicate the high frequency cuttoff. The computer I'm on right now has crappy speakers but try this in audacity. Then use the equalizer with a U shaped curve to boost frequencies below about 400 Hz, decrease the mid range, and boost above about 4000 Hz:

sample 1 is the high quality source. Did you equalize the high quality source or the low quality source?
I wanted to edit only sample 2 to make it match the sample 1..

When you said the first sampe was "not full" I thought you meant the sound wasn't full -- lacking in bass. I guess what you meant was it is not complete.

You will never make sample 2 match sample 1. It rolls off at 7 KHz so it will never have the presence and clarity of the better track.

what do you mean by "reverse the channels"?

I think he means "invert" one of the channels. In Audacity split the dual mono track, invert either the left or the right channel, then merge them together again. It gives the mono track a diffuse sound. Personally, I hate that sound. If someone plays it on a mono system they will get no sound at all. On a Dolby Matrix system they'll get mono audio from the rear speakers.

[mathmax]

When you said the first sampe was "not full" I thought you meant the sound wasn't full -- lacking in bass. I guess what you meant was it is not complete.

exact. Sorry for confusion.. sample1 comes from a HQ source but not complete, sample2 comes from a LQ complete source. I want to complete the first source using the second one for the missing parts.

You will never make sample 2 match sample 1. It rolls off at 7 KHz so it will never have the presence and clarity of the better track.

of course... but if I join the 2 files without adjusting, the transitions are so bad. I thought I could maybe adjust to make the 2 sources match, even a little...

I think he means "invert" one of the channels. In Audacity split the dual mono track, invert either the left or the right channel, then merge them together again. It gives the mono track a diffuse sound. Personally, I hate that sound. If someone plays it on a mono system they will get no sound at all. On a Dolby Matrix system they'll get mono audio from the rear speakers.

Thank you, I thought about that, but I didn't think such operation could enhance in any way...

I tried something. Please tell me if that makes sens:
- for each sample I plot the spectrum and then exported in .txt files. Each file contain a column for frequencies and a column for the levels.
- I then pasted the content of these columns in Excel and make the differences between the levels of file1 and the levels of file2.
- I took again the frequencies and this new calculated column to create an xml file, which I imported in the equalizer.

The LQ sample is now corrected and its spectrum looks almost identical to the spectrum of sample1. Of course, I'm aware this doesn't recover the lost frequencies and doesn't make anything HQ, but I thought it could balance the two audios in the same way... what do you think?

I maybe need to leave the high frequencies untouched in the LQ clip. No need to amplify something that is definitely lost..

[jagabo]

I tried something. Please tell me if that makes sens:
- for each sample I plot the spectrum and then exported in .txt files. Each file contain a column for frequencies and a column for the levels.
- I then pasted the content of these columns in Excel and make the differences between the levels of file1 and the levels of file2.
- I took again the frequency and this new calculated column to create an xml file, which I imported in the equalizer.

The LQ sample is now corrected and its spectrum looks almost identical to the spectrum of sample1. Of course, I'm aware this doesn't recover the lost frequencies and doesn't make anything HQ, but I thought it could balance the two audios in the same way... what do you think?

Yes, I think that's your best bet.

I maybe need to leave the high frequencies untouched in the LQ clip. No need to amplify something that is definitely lost..

I agree with that. You'll just accentuate high frequency hiss.

If the transitions are still too jarring you might consider degrading the better sample a bit. And cross fading to spread out the transition.

[mathmax]

thank you

how can I cross fade the transition in audacity? Should I use the envelop tool? If I decrease linearly the first clip and increase linearly the second at the same time, would I get a smooth transition?

[mathmax]

I don't think it sounds any better after all..

I uploaded my result here (before and after frequency correction)
I included the xlm file I used for equalization

http://www.mediafire.com/?a49hn8ae97f1bvy

[mike20021969]

Did we ever find out the name of the music?

[mathmax]

yes, the song is called Slapstick. It's an unreleased song of Michael Jackson.

[jagabo]

Here's what I meant by crossfading between the two. I made no attempt to make the two sound similar or to perfectly align the two tracks, just applied the crossfade, one faded out, the other faded in, with a one second overlap.

[mike20021969]

yes, the song is called Slapstick. It's an unreleased song of Michael Jackson.

There is a 4m14s copy of this on 4Shared, in stereo but lesser quality.

http://www.google.co.uk/search?q=michael+jackson+slapstick&ie=utf-8&oe=utf-8&aq=t&rls=...ient=firefox-a

(It should be the 7th result).