www.makemkv.com

Hi folks.

I'm building a new PC dedicated to ripping Blurays and 4K movies.

Been using MakeMKV and handbrake on a Win10 Pro VM on my QNAP, but it's as slow as molasses in Quebec in January.

My plan is to rip them on the PC, then transfer them to My QNAP via 10G DAC, through my Microtik 10G switch. Playback from QNAP to AppleTV.

I'd like to go AMD x570 for the Proc & Mobo, M.2 4X4 for fast R/WR.

Any contructive input would be appreciated.

Cheers!

The ripping part can be done on a Raspberry Pi. Encoding, not so much.

Encoding wants either dedicated hardware (QSV on Intel, or the equivalent AMD or nVidia hardware) if speed is all you need (quality suffers with hardware encoders), or lots of cores (8-12 "real" cores) and clock speed. Ryzen seems to be the current cost/performance leader.

If absolute encoding speed isn't needed, you can put together a previous-generation system with good specs for a few hundred dollars.

Thanks Woodstock.

I should have used the correct terminology. I was combining the whole enterprise of "ripping" and "encoding" under the "ripping" label.

Yes, the encoding is taking a very long time.

The system I was designing from NewEgg had a newer Mobo and Ryzen 5700. Probably serious overkill.

You mentioned "previous gen". Would you mind sending some more detail on what you had in mind? Again, the goal is to speed up the entire process with emphasis on the encoding to .MP4

Thanks!

If you can afford a Ryzen, it has advantages over the AM3+ FX-series processors I've been buying. It used to be that AM3+ motherboards were as low as $50 with onboard network/video/USB3, but those days are over - if you want one, you're buying "new old stock" that's 2x or more costly.

A disadvantage the Ryzens have is that you have to be careful which processor you buy, UNLESS you're buying separate video. Some do not come with the video stuff on-processor, so even a motherboard that says it "includes video" doesn't really.

Probably the biggest recommendation - don't skimp on CPU cooling. If I'm going to encode on a machine, the CPU cooler is HUGE. There isn't a need to overclock, but there is a need to cool well. And getting the hot air out of the case is a big consideration, too.

I had to build a new system to encode 4K UHD and gaming (was long overdue, just trying to stretch it out as far as I could). Was previously on an Intel i5 3570K that was gonna take 3 days (!) to encode a UHD at my settings - said it's finally time. Was originally gonna only go with a 5900x, decided to bump up to the 5950x to help with encoding - glad I did. Encodes take about 6 hours now per ~1 hour 4k HEVC video, so around 12-15 hours a movie. Much better than 3 days (3570k used to do 1080p Blu-rays in 6 hours at my settings, so total encoding time doubled for me in the end, but for 2160p encodes now, not 1080p - not bad IMO). For HEVC material that's around 18-20mbps or lower I can also do simultaneous encodes, and it only increases encode time around ~33% versus encoding a single UHD file. 1080p AVC Blu-ray encodes went from previously mentioned 6 hours per movie to now 1 hour per movie on the new system.


Here are my specs: Ryzen 9 5950x | Noctua NH-D15 | ASUS TUF Gaming x570-Pro WiFi | 64GB DDR4 3600 | EVGA 1080 FTW | Super Flower Leadex Platinum SE 1000W 80+ Platinum | ASUS BW-16D1HT | Samsung 980 Pro 250GB PCIe 4.0 NVMe (OS) | 40TB HDD (Plex Server Backup) | 12TB HDD + 250GB WD Black SN750 NVMe PCIe 3.0 + Samsung 850 EVO 250GB (Games and VMs)

It's in a Phanteks Enthoo Pro case, with 2x Noctua 200mm and 2x Noctua 140mm fans replacing the default case fans

Are you using Handbrake to encode? Are you keeping 1080p blu-rays as h.264 files?

Lately, I've been encoding 1080p blu-rays into 10-bit h.265 files. The 10-bit color depth helps reduce color banding and the h.265 encoding helps minimize file sizes. Using hardware-based encoders a typical 2-hour blu-ray takes less than 30 minutes. Encoding with modern Nvidia GPUs or Intel Quick Sync Video can produce decent results.

I've been using this for encoding lately: https://github.com/donmelton/other_video_transcoding

As Woodstock has suggested, you can probably come up with the 'best' encode using the software encoder… There is, however, a philosophical question about what does 'best' mean? If you really want the best possible, just put the rip on your QNAP and watch that. It is the 'best' you can watch and it saves lots of compute time. If you are going to transcode, then does 'best' mean visually good enough at a sufficiently small file size? And how much compute time is worth it to get there?

Where ever you fall in the spectrum, I suggest giving a hardware encoder a try on a few movies you like and know well. The output might be good enough.

Folks, thank you very much for your thoughtful replies and the time you spent writing them!
I'm going to soak this all in and reply with a few thoughts and questions.
Cheers!

I thought I'd throw my $0.02 in to this conversation...

At this point, you may actually find it cheaper to buy more storage space as opposed to encoding. It sounds somewhat crazy but if you run the numbers using average electricity costs, the $/kwh power usage over 4 - 12 hours of encoding is greater than the $/Gb to store the movie. Hardware encoding can reduce the time but even then, you're looking at hardware cost plus power.

Look for a sale on large hard drives and just rip then use original quality. You'll have the best audio / visual option plus save money and time.

I didn't believe it until I broke out the Excel sheet and ran the numbers myself...

Lost Dog wrote: ↑

Thu Oct 21, 2021 4:31 pm

I thought I'd throw my $0.02 in to this conversation...

At this point, you may actually find it cheaper to buy more storage space as opposed to encoding. It sounds somewhat crazy but if you run the numbers using average electricity costs, the $/kwh power usage over 4 - 12 hours of encoding is greater than the $/Gb to store the movie. Hardware encoding can reduce the time but even then, you're looking at hardware cost plus power.

Look for a sale on large hard drives and just rip then use original quality. You'll have the best audio / visual option plus save money and time.

I didn't believe it until I broke out the Excel sheet and ran the numbers myself...

+1

Not to hurt anyone's feelings, but I don't how their could be the desire to encode at all nowadays.

I'd rather encode than buy 'xx TB' HDDs all the time, simple as that really (well, and it takes care of subs for me and having to worry about 'are there forced subs for this movie' - 'are the subs timings right with the video file' - 'what format are the subs' - 'what format subs does the device I'm watching on support' - etc etc etc). More than just the upfront dollar cost to it, it's having to always buy more drives, and then the space/noise/etc they take up in an enclosure/area as the amount of drives grows.

I'd have to have around 80TB currently instead of the 40TB I'm working with ATM. No thanks. I'm sure I'll get there one day, but no where near as quickly as others will need to if I encode first, and by then it's cheaper for those same drives or around the same amount for now even larger capacities - reducing how many drives I need again, rinse and repeat.

For me, compression is important. Even with 70tb available, that would only take a third of my collection uncompressed, and I'm not even dealing with 4k content.

One of the reasons I choose to transcode is it makes the files more easily portable. Putting a few movies on a phone, tablet or laptop is fairly easy when they're about 5GB. It can be a lot harder when they're 30-ish GB.