According to this chart, DP's SRC quality has diminished...

[FMiguelez]

http://src.infinitewave.ca

If the chart is to be trusted, then:

- DP no longer scores respectable results compared to other DAWs in the SRC department (compare it to Ableton's Live 9. It blows DP out of the water. Even PT LE (!) seems to do much better than DP 9.).
A few years ago DP shined in that chart. But a few years later, most DAWs seem to have left DP behind in the dust.

- And, more worrying, DP9's SRC algorithm seems to be worse than DP 7's. Should it not have improved instead of worsened?


... If the chart is to be trusted.

Hmmmm....
During the week, if I have time, I will compare DP to Live 9 in my studio by actually listening.

[mikehalloran]

Where is the key that explains this chart? Without that, it’s meaningless.

I have a hard time believing that DP 6 was the best sounding version ever—or that 5.1 was horrible.

If you believe those charts...

[HCMarkus]

It is interesting data... but note that there is more to conversion than the first graph presented.

Although I am not an EE, I believe that there are tradeoffs in conversion algorithms. Note the superior impulse response of DP compared to, say Live 9, which exhibits significantly greater and longer pre- and post-ringing, likely due to steeper filter response.

Which is worse:

Artifacts at -170dB (Live) vs -150dB (DP8-9)

or

Ringing at -17dB decaying for 0.0020 sec (Live) vs -20dB decaying for 0.0012 sec (DP8-9)?

Also note that the "better" DP7 graphs reveal ringing similar to the Live algorithms. Old saying:

There is no such thing as a free lunch.

All of the above aside, I'd be extremely surprised if anyone could reliably tell the difference on musical material in an ABX Test.

[FMiguelez]

Where is the key that explains this chart? Without that, it’s meaningless.

I have a hard time believing that DP 6 was the best sounding version ever—or that 5.1 was horrible.

If you believe those charts...

According to one of the engineers at SOS magazine, an ideal graph would be with a total black background. The artefacts are shown in faint red. I'll dig through my iPad's history to find the article where I found it.

It just seemed curious how, apparently, DP had an algorithm that did better in earlier versions in those same charts, and its competition does better now. No wonder they looked familiar, BTW... A few years ago, when DP did much better, among the best, it was praised and raved about in this very place, with little to no skepticism

For the charts to be accurate, the MOTU engineers would've had to mess with the code and do a bad job with it. We'll probably never know if that ever happened, so I do share your skepticism.

I've done my share of SRC with DP, and I've never heard anything unacceptable. I actually think it is quite clean and transparent, FWIW. I'll have to test it and compare it to Live 9 (which apparently is even better now).

All of the above aside, I'd be extremely surprised if anyone could reliably tell the difference on musical material in an ABX Test.

Agreed on the better ones. But there are some that look really bad and those are probably audible. When I get time, I'll also test the QT Pro 7 one, which doesn't look particularly good at all.

[HCMarkus]

It just seemed curious how, apparently, DP had an algorithm that did better in earlier versions in those same charts, and its competition does better now.

FM, with all respect, I'd ask that you consider the entirety of my prior post, as I feel that referring to DP7 as "better" is a generalization that fails to take into account all of the factors in play here. It is certainly different than DP8-9, but, in my estimation, which algorithm's artifacts are more audible/objectionable is the real question. It is entirely possible that ringing has a greater impact on perceived quality than noise.

So why does ringing occur? All (nontrivial) filters ring, regardless of whether they are brick-wall or not, regardless of the shape of the input signal, and regardless of whether the input is continuous or has sharp transitions. The reason is that if the input has energy in the frequency bands that are stopped (whether wholly or in substantial part), that energy is effectively stored internally in the filter and released slowly as in-band energy as time progresses. Most of the time this release is not noticed very much because it is drowned out by the response to the in-band signal that is present. However, if the in-band signal changes (or ceases) relatively suddenly, that energy stored from previous times still has to be released, and this is the ringing that is observed after the in-band signal has disappeared. In DSP terms, the FIR filter buffer continues to empty out even after the signal has ended, and so the output continues even after the signal ends. Since sharp-cutoff filters have long buffers (many biquad sections if you will), this emptying takes a long time and is much more noticeable than with a more easy-going filter (emphasis added) which empties out quite quickly.

Source: https://dsp.stackexchange.com/questions ... transition

Here's some more info:

https://en.wikipedia.org/wiki/Ringing_artifacts

If one reviews the filter curves of DP8-9 vs DP7, it is clear that the DP8-9 algorithm uses a less-steep filter which, as a result, causes less ringing. In this sense, the DP8-9 algorithm could be considered superior. I refer us again to the "free lunch" quote...

[bayswater]

FM, my recollection from the older charts is that DP was not that bad at SRC, but not one of the best either. The other thing I think I remember —someone suggesting that DP results tracked pretty closely to those of CoreAudio itself suggesting that DP was using whatever OS X offered up for the purpose. Somewhere back there I saw that DSP-Q performed better than DP and I used that for a while to convert samples but I cant say I could ever hear the difference.

[FMiguelez]

FM, with all respect, I'd ask that you consider the entirety of my prior post, as I feel that referring to DP7 as "better" is a generalization that fails to take into account all of the factors in play here.

I didn't mean to ignore the "uncomfortable" part of your post. I actually found it very interesting, and I wanted to do some further reading, but I fell asleep in the couch

Thank you very much for the info you provided. I'm reading it right now, and you are correct in your assessment. They aren't telling the full story.

FM, my recollection from the older charts is that DP was not that bad at SRC, but not one of the best either. The other thing I think I remember —someone suggesting that DP results tracked pretty closely to those of CoreAudio itself suggesting that DP was using whatever OS X offered up for the purpose. Somewhere back there I saw that DSP-Q performed better than DP and I used that for a while to convert samples but I cant say I could ever hear the difference.

[/quote]
I remember DP did very good. We were quite proud about it, but my recollection might definitely be skewed. It was long ago...

So, it looks like I over-reacted a bit with the meaning of those charts, didn't I? Good thing I can always count on my MOTUNation friends to set me straight

[HCMarkus]

I wanted to do some further reading, but I fell asleep in the couch

Ah, the pitfalls of the studio musician's late night lifestyle! (Says the guy who posted at 1:30AM)