CEDAR wins retouch patent fight

[Hugh Robjohns]

This is an interesting story that could have very far reaching implications for other software manufacturers...

http://www.cedaraudio.com/news/dcp19may14.html

Intriguing.

H

[Jonesd90]

Fortunately I use CEDAR and not one of the other restoration suites. Could this have an impact on the likes of iZotope RX3 and such? Could be a massive gain for Cedar to knock out a whole load of it's main competitors.

Dave

[Wease]

...but only I the US??

Wouldn't it be wiser to now licence that touch system to izotope et al?? Rather than recall??

[Jeraldo]

Fortunately I use CEDAR and not one of the other restoration suites. Could this have an impact on the likes of iZotope RX3 and such? Could be a massive gain for Cedar to knock out a whole load of it's main competitors.

Dave

How did the folks at Cedar come up with this method? I wonder if they were editing photo's....

[Mixedup]

The text of that article implies that this is an out of court settlement with one of the smaller companies to use this technology. In which case we're yet to see any precedent established and deep pockets would still be required to take on Sony, Steinberg (Yamaha) etc in the courts. Not saying it won't happen but I'd be surprised.

[Hugh Robjohns]

I believe CEDAR has actually already negotiated several agreements with companies who have released products that infringe the Retouch patents -- arrangements that CEDAR describes as 'positive for all parties'. I think we can take that to mean some kind of licensing arrangement.

CEDAR came to an arrangement with iZotope for RX2 a couple of years ago:

http://www.cedaraudio.com/news/izotopelicence_9may12.html

H

[muzines]

How did the folks at Cedar come up with this method? I wonder if they were editing photo's....

I think it's more likely that they are, well, smart people with maths and DSP algorithms oozing out of every pore...

[James Perrett]

How did the folks at Cedar come up with this method? I wonder if they were editing photo's....

This kind of spectral display is a fairly common way of displaying all kinds of data so it wouldn't have been a huge step to say "if we can display it we can change it". The work is in making that happen in a sensible way.

[Hugh Robjohns]

How did the folks at Cedar come up with this method? I wonder if they were editing photo's....

Audio spectrographs of one form or another have been in use for a very long time as a means of displaying and analysing audio content. The challenge -- and the clever stuff that the CEDAR boffins created and patented -- is finding ways of manipulating selected portions of the audio, in the time, frequency and amplitude domains. It was a revolutionary development and their innovation deserves recognition and respect -- which it seems to be getting from many manufacturers, if not all.

H

[Jeraldo]

Thanks for both links, Hugh. Actually, I had long been wondering if there had been any sort of agreement or more between Cedar and iZotope. Interesting reading. And now I know.

[Goddard]

Interesting. Kinda odd though (or perhaps merely the cheaper way out?) that Diamond Cut would recall their wares from the US market rather than take a license from Cedar covering them.

Btw, Algorithmix' reNOVAtor and easyreNOVAtor, Magix Sequoia's Spectral Cleaning, Stillwell's spectro plug-in and the "In The Spectral Wash" section in this SOS review of Samplitude may be of some interest.

[Kwackman]

Interesting. Kinda odd though (or perhaps merely the cheaper way out?) that Diamond Cut would recall their wares from the US market rather than take a license from Cedar covering them.

Purely hypothetical, but if you were a Diamond Cut customer using their now naughty version of Retouch, would you "update" your license?

[Happyandbored]

This is completely and totally insane.

Why should the concept of drawing a box around a visualisation of a set of frequencies and hitting delete be protected, while drawing a box round a set of notes in a piano roll and hitting delete not be? Hey, what if it were an extremely high resolution piano roll with thousands-on-thousands of keys each controlling an individual sine wave? And what if the colour of the pixels representing each note as it's played changed colour depending on it's velocity? You know, a little like the way the notes change colour in Logic when you use the velocity tool.

Cedar did not invent the mathematics behind notch filtering and spectral analysis. Neither did they invent the concept of drawing a box round something and hitting delete. Plenty of prior art, this is patent trolling, call it for the bulls*** it is FFS.

[Hugh Robjohns]

Why should the concept of drawing a box around a visualisation of a set of frequencies and hitting delete be protected...

The granted patent covers a little more than that.

Abstract: A sampled digital audio signal is displayed on a spectrogram, in terms of frequency vs. time. An unwanted noise in the signal is visible in the spectrogram and the portion of the signal containing the unwanted noise can be selected using time and frequency constraints. An estimate for the signal within the selected portion is then interpolated on the basis of desired portions of the signal outside the time constraints defining the selected portion. The interpolated estimate can then be used to attenuate or remove the unwanted sound.

The patent is not about the spectrogram display or drawing pretty boxes. It's about defining and enacting time and frequency-based interpolation of specific parts of the audio signal.

Cedar did not invent the mathematics behind notch filtering and spectral analysis.

True... but that's not what they were granted a patent for either.

Plenty of prior art

Patents don't get awarded if there is relevant 'prior art'.

The simple fact is that spectral editing didn't exist before CEDAR's Retouch. It was an entirely unique, genuinely innovative, and completely revolutionary concept. There was no 'prior art' -- no part of spectral editing in the form Retouch allows existed before it's invention. The granting of the patent protects the company's investment in world-class R&D and ensures that other manufacturers respect that investment. Without that protection gaining investment in innovation is extremely difficult.

I don't know why Diamond Cut chose to recall its products, but I don't think you can legitimately call it 'patent trolling'. CEDAR has established mutual licensing agreements with a number of other manufacturers to develop products using the patented technology and techniques. SO it's not like their are stifling the industry.

H

[Happyandbored]

Re: The patent is not about the spectrogram display or drawing pretty boxes. It's about defining and enacting time and frequency-based interpolation of specific parts of the audio signal.

You missed my point...

What my absurd example showed, is that conceptually at least this is little different than deleting some notes in an piano roll, assuming of course those notes were sine waves. The underlying idea of graphing pitch to the vertical axis of a graph and colour to volume and being able to directly edit the display existed long before. Conceptually, the only real difference is the resolution of each event and the sampling rate. So at least conceptually, if not specifically to the word, there is plenty of prior art!

You could conceivably dress up the specific processes involved in MIDI sequencing in flowery language too and then whoever invented the first midi sequencer could claim they are due licensing fees from all other sequencer developers... Assuming of course they had the wonderful foresight and lack of any reasonable moral code to consider patenting it... To which I'm sure someone who believes in software patents would say sure why not, but hell, if a Cambridge-based magazine wants to defend the arbitrary legal structures which happen to benefit a major Cambridge-based audio tech developer, then who am I to argue?

(last bit intended slightly tongue-in-cheek)

[muzines]

What my absurd example showed, is that conceptually at least this is little different than deleting some notes in an piano roll, assuming of course those notes were sine waves. The underlying idea of graphing pitch to the vertical axis of a graph and colour to volume and being able to directly edit the display existed long before. Conceptually, the only real difference is the resolution of each event and the sampling rate. So at least conceptually, if not specifically to the word, there is plenty of prior art!

Sure. But it's not really "concepts" that you patent, it's the actual technology to do it. Which is a lot more involved than moving around single note events, because you have to identify and extrapolate those frequency domain parts of the signal and they have impact on the areas around them.

In any case, they were the first to invent and patent the tech, so they have a right to exploit it, regardless of what the current state of software patents is (which is a whole different discussion...)

[Hugh Robjohns]

You missed my point...

Actually... I think you missed mine!

conceptually at least this is little different than deleting some notes in an piano roll, assuming of course those notes were sine waves

It's entirely different. Retouch doesn't (just) delete unwanted noises. It replaces those unwanted noises with background sound derived from the surrounding timeline and spectrum. That's what makes it such a powerful and unique process, and that's fundamentally what the patent was granted for.

The underlying idea of graphing pitch to the vertical axis of a graph and colour to volume and being able to directly edit the display existed long before.

Yes it did. That's why CEDAR employed it as a familiar and inutitive tool for controlling their unique signal processing technique. CEDAR even acknowledged the history of the spectrograph display in the patent application FFS.

The patent isn't about the display, but rather it describes how CEDAR's unique and innovative interpolation process is manipulated using an existing display format. And that's why the patent was granted

Conceptually, the only real difference is the resolution of each event and the sampling rate.

Nope. Sorry. You've completely misunderstood what spectral editing is.

...at least conceptually, if not specifically to the word, there is plenty of prior art!

Only in the form of the colourful spectrograph display -- but that's NOT what the patent is about. There is no prior art in the process of interpolating audio on a temporal and spectral basis to replace unwanted sounds. That was a unique CEDAR invention.

if a Cambridge-based magazine wants to defend the arbitrary legal structures which happen to benefit a major Cambridge-based audio tech developer, then who am I to argue?

I'd say you are someone who doesn't understand the specific patent we're discussing, or the general patent process. (tongue slightly in cheek here too... )

H

[Happyandbored]

Re: "...it's the actual technology to do it"

Really? And what technology is that? Oh, it's Fourier transforms isn't it? Thanks Joseph Fourier! Too bad you're dead.

Spectral editing has been possible for far longer than Cedar's patent has existed. It's perfectly possible to achieve with some automated notch filters and cross-referenced to a spectral analysis tool. Maybe not as elegant but that's the technology being used behind the scenes.

Yes, Cedar were the first to patent and therefore have a dubious legal right to exploit it, but they were not the first to actually invent it.

[Hugh Robjohns]

Really? And what technology is that? Oh, it's Fourier transforms isn't it?

Impressive... but not appropriate.

It's perfectly possible to achieve with some automated notch filters and cross-referenced to a spectral analysis tool.

No. That's not even close to what CEDAR's Retouch does, or how it works... or what the patent is all about.

Maybe not as elegant but that's the technology being used behind the scenes.

Sorry, but you are completely and utterly wrong.

Yes, Cedar were the first to patent and therefore have a dubious legal right to exploit it, but they were not the first to actually invent it.

So wrong...

H

[Happyandbored]

Re: "It replaces those unwanted noises with background sound derived from the surrounding timeline and spectrum. That what makes it such a powerful and unique process, and that's fundamentally what the patent was granted for."

Oh wow, how amazingly powerful and unique... Filtering surrounding noise and editing it in to fit the hole, no one using an audio editor and relevant processing has ever done that before!!!

Re: "Nope. Sorry. You've completely misunderstood what spectral editing is."

Oh please, get your head out of your arse! You're deliberately misrepresenting my self-declared absurd example to avoid engaging with the actual argument!

Re: "There is no prior art in the process of interpolating audio on a temporal and spectral basis to replace unwanted sounds"

Yes there is! Use audio editor, look at spectrum analyser, find unwanted frequencies, use notch filter, automate to ensure correct frequencies are filtered at correct time, find bit of audio nearby without unwanted frequencies, copy, band-pass filter, paste! Again, not so elegant and a lot more time consuming doing it manually, but the idea that one company should be able to patent software implementation of such a blatantly obvious process is insane.

I'd say you are someone who doesn't understand the difference between legality and morality. You are supporting an arbitrary legal structure which rewards those who are already established to the detriment of any newcomers. It rewards what is first, regardless of what is best.

It seems you would prefer a world where one person or company is rewarded over and over again for any given invention than a competitive one which allows improvements to be made to existing inventions. If this were a whole method or specification which had actually been designed and invented that is one thing, but when it is merely the fairly obvious recombining of existing technology it's a little hard to take seriously.

Your friends over at Cedar are patent trolls attempting to profit off the hard-work of others.

[Hugh Robjohns]

H

[Dave Blackman]

Have you actually used ReTouch?

[John Willett]

Re: "It replaces those unwanted noises with background sound derived from the surrounding timeline and spectrum. That what makes it such a powerful and unique process, and that's fundamentally what the patent was granted for."

Oh wow, how amazingly powerful and unique... Filtering surrounding noise and editing it in to fit the hole, no one using an audio editor and relevant processing has ever done that before!!!

Re: "Nope. Sorry. You've completely misunderstood what spectral editing is."

Oh please, get your head out of your arse! You're deliberately misrepresenting my self-declared absurd example to avoid engaging with the actual argument!

Re: "There is no prior art in the process of interpolating audio on a temporal and spectral basis to replace unwanted sounds"

Yes there is! Use audio editor, look at spectrum analyser, find unwanted frequencies, use notch filter, automate to ensure correct frequencies are filtered at correct time, find bit of audio nearby without unwanted frequencies, copy, band-pass filter, paste! Again, not so elegant and a lot more time consuming doing it manually, but the idea that one company should be able to patent software implementation of such a blatantly obvious process is insane.

I'd say you are someone who doesn't understand the difference between legality and morality. You are supporting an arbitrary legal structure which rewards those who are already established to the detriment of any newcomers. It rewards what is first, regardless of what is best.

It seems you would prefer a world where one person or company is rewarded over and over again for any given invention than a competitive one which allows improvements to be made to existing inventions. If this were a whole method or specification which had actually been designed and invented that is one thing, but when it is merely the fairly obvious recombining of existing technology it's a little hard to take seriously.

Your friends over at Cedar are patent trolls attempting to profit off the hard-work of others.

This looks like a post from a troll who cannot listen to reason.

[Happyandbored]

So when you say:

"The patent is not about the spectrogram display or drawing pretty boxes. It's about defining and enacting time and frequency-based interpolation of specific parts of the audio signal."

Do you mean that somehow Cedar have stumbled across an entirely new form of spectral editing which does not somehow rely on filtering DSP along with spectral analysis that works on some entirely new mathematics in no way related to Fourier transforms? Or some utterly new unique way of copying and pasting noise from one part of a recording and replacing offending frequencies in another?

[Happyandbored]

"Re: This looks like a post from a troll who cannot listen to reason."

Better a vanilla troll than a patent troll. And definitely much better than a lizard.