CD Ripping - SQ Differences across Rippers

Has anyone else noticed differences in SQ between same CD files ripped on different rippers?

My set-up : I rip on a standalone, networked machine and drag the files across my wireless network to live on a Melco unit serving a USB DAC.

My original ripper recently died. I’ve replaced it with latest model. I just compared a few rips made on the mk1 with same rips from same cd’s on new model. There is a clear difference. Those from the mk1 are more natural and deeper detailed than those on the new model; files from latter have less mid range body, high frequencies which are harsher than mk1 at extremes, less sense of recording space. They sound more compressed and are not as relaxing to listen to.

(I notice the kbps on the rips differs between the two units but sometimes by only 1. In any event, the difference is still clear).

Everything in my system makes a sonic difference but I really didn’t expect a change in ripper to have any effect. It’s worth noting but can anyone explain how it can be?


What file format are you ripping to?

I’ve tried FLAC and WAV. Neither overcomes the problem. Both suffer similar attributes. (But if you were thinking p’raps I’m ripping as MP3 - well, that’s not a million miles off how they sound…)

Also, whilst I can plainly hear the positive difference between the two (in favour of WAV) when I download WAV from Qubuz) I’m less convinced of any positive difference from the new model’ WAV rips.

It’s very tedious…

My thoughts exactly…
Are both rips to a LPCM format like WAVE?

Run a checksum program on both ripped files. If the checksums are the same, the files contain the exact same bits.

Thanks Suedkiez, will do.

In fact, I’d thought to compare file sizes in the first instance - and find that the original rips from mk1 are either 0.1 or 0.2 mb smaller than the same rips from the newer (poorer SQ) new model. Makes no sense.

Also downloaded an excellent cheksum app - which obviously says the two files don’t match…

So, not all rippers rip the same - even between models by the same manufacturer… Shame and PIA given how good was the mk1.

Meantime I’ve downloaded a FLAC track from Qubuz at 28.8mb compared to 32.1mb off mk1 and 32.2mb off mk3. I’ll compare SQ out of interest…

There are so many versions of the same track, I’m not sure what that will tell you. Do you have a PC that you can rip you CD to, then compare size and listening

Interesting. I forgot of course that even different metadata (if embedded) would make the checksums not match even if the music data was the same. So not very conclusive unfortunately but at least you know that the files are definitely not the same and sound differences are not impossible.
The whole thing is at best only applicable to WAV anyway, when using flac it is certainly possible that different encoder versions might create different encodings (that would still decompress to the same PCM). Can you try ripping something to WAV without metadata?

It won’t tell me anything about the rips; just whether the Qubuz d/load sounds worse or better than either of them. It muddies the waters, really, so I guess should be ignored for purposes of the thread’s topic.

My takeaway is that the rips of the two machines are different according to checksum and, more obviously, visible file size; and one sounds a) noticeably inferior to the other and b) inferior full stop. So anyone thinking that all reputable CD rippers will produce the same result, either technically or sonically, needs to think again - and try before you buy. Like I didn’t!

Have you checked the rips against AccurateRip?

I wonder if the issue is with the CD transport in the MK3. Possibly having problems reading the CD, jitter etc. Presumably you get the same with other CD’s, and is not just that the CD you are eating has a scratch or something. If you are comparing WAV, then they wont have metadata, so should be the same. I wouldn’t have thought different CD rippers would vary for the same CD. Most rippers will read multiple times to ensure they aren’t reading errors, so perhaps your settings are set differently on the MK3 (sorry dont know the device, so dont know what settings exist), but its default settings may be a more dirty read? Comparing to a PC ripped CD may help the diagnostics.

Thanks Guinnless. I hadn’t. Have d/loaded PerfectTunes and there are errors but not throughout. One album gets a complete thumbs-up - yet the SQ is per above.

Thanks GadgetMan - there aren’t any ripper settings on the machine which would affect SQ in this way. Yes, the issue appears on both CD’s I’ve tried so far. So I’m baffled. Something is causing the new machine’s rips to be inferior to the old machine. Tomorrow I’ll check 2 or 3 more albums and see what comes out, including finding old album rips I can check against AccurateRip, as well as the new rip.

Thanks Suedkiez, I hadn’t realised encoding might be done differently across the two machines; though I suppose the difference in kbps rates supports that. But doesn’t AccurateRip do its checking against CD FLACS? Also the m/data is identical - I’m ripping the same CD I used to rip from the original machine.

Have you spoken to Innuos about this? The UK sales guys I’ve spoken to in the past have always been very approachable and I dare say they would be keen to know why you’re getting poor performance from one of their products.

Yes indeed C - I wrote to them yesterday afternoon. But I’ve learned some very useful stuff here meantime; about AccurateRip e.g.

I would qualify all the above by saying the new unit is an excellent product in terms of user interface, wide spec’, VFM, aesthetics and build; and I’m not using any of that wide spec’ so make no comment on its performance e.g. as streamer, CD player etc. My usage is solely as CD ripper - perhaps 15% of what it’s about.

I’ll feedback whatever I find out but if anyone has any other ideas I’d be happy to hear them.

1 Like

Not wishing to sound flippant, but there are about 100 threads on this topic on this and the archived forum. Some thousands of comments long. There’s some very detailed discussion and test results buried in them and I suspuct most of the people who contributed aren’t keen to recount their findings for the Nth time.

To be clear, it’s just a possibility in principle - a WAV will always be the PCM data ripped from the disc (if ripped correctly). But a FLAC is then compressed, and in principle, a FLAC software version x.y might have slight differences to x.z - let’s say if they fixed a bug in x.z. You could of course uncompress both FLACS and the resulting WAV should be the same, I think


I did look briefly before posting and couldn’t find a precise fit. However, I’ve now looked deeper and find a thread about SQ differences between almost identical rips from two different machines. It doesn’t categorically answer the question but it suggests a categorical answer isn’t going to be forthcoming. So I’ll await advice from manufacturer and leave it here if it adds to the current level of understanding of the issue.